You are on page 1of 141

How to make Two-Lane

Rural Roads safer


Scientific Background and
Guide for Practical Application

WITPRESS
WIT Press publishes leading books in Science and Technology.
Visit our website for the current list of titles.
www.witpress.com

WITeLibrary
Home of the Transactions of the Wessex Institute, the WIT electronic-library provides the
international scientific community with immediate and permanent access to individual
papers presented at WIT conferences. Visit the WIT eLibrary at
http://library.witpress.com
This page intentionally left blank
How to make Two-Lane
Rural Roads safer
Scientific Background and
Guide for Practical Application

Authors:
R. Lamm
University of Karlsruhe (TH), Germany
A. Beck
University of Karlsruhe (TH), Germany
T. Ruscher
University of Karlsruhe (TH), Germany
T. Mailaender
Mailaender Ingenieur Consult GmbH, Karlsruhe, Germany

Co-Authors:
S. Cafiso
University of Catania, Italy
G. La Cava
University of Catania, Italy
Published by

WIT Press
Ashurst Lodge, Ashurst, Southampton, SO40 7AA, UK
Tel: 44 (0) 238 029 3223; Fax: 44 (0) 238 029 2853
E-Mail: witpress@witpress.com
http://www.witpress.com

For USA, Canada and Mexico

WIT Press
25 Bridge Street, Billerica, MA 01821, USA
Tel: 978 667 5841; Fax: 978 667 7582
E-Mail: infousa@witpress.com
http://www.witpress.com

British Library Cataloguing-in-Publication Data

A Catalogue record for this book is available


from the British Library

ISBN-10: 1-84564-1566
ISBN-13: 978-1-84564-156-6

Library of Congress Catalog Card Number: 2005928179

No responsibility is assumed by the Publisher, the Editors and Authors for any injury and/or
damage to persons or property as a matter of products liability, negligence or otherwise, or
from any use or operation of any methods, products, instructions or ideas contained in the
material herein.

WIT Press 2007

Printed in Great Britain by Athenaeum Press Ltd., Gateshead

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted in any form or by any means, electronic, mechanical, photocopying,
recording, or otherwise, without the prior written permission of the Publisher.
Dedications

Ruediger Lamm
1937-2005
The Authors dedicate this book to the memory of Professor Ruediger Lamm, an
excellent man, a great scientist and a leading figure at an international level in the
field of Road Safety, who suddenly and unexpectedly passed away before
publication.

Christa Lamm
Before his death, Professor Ruediger Lamm dedicated this book to his wife
Christa Lamm for four decades of selfless support.
This page intentionally left blank
Co-workers

Co-workers:

A. Beck
beck-consult.de Berghausen, Germany
R. Heger
Dresden University of Technology, Germany
B. Psarianos
National Technical University of Athens, Greece

Supported by:

AKG Software Consulting GmbH


A. K. Guenther
President, Ballrechten-Dottingen, Germany
J. C. Hayward
Robert Morris University, USA
M. Eugen Rapp
Bureau of Engineering, Max Eugen Rapp & Partners, Germany
K. Wolhuter
Council for Scientific and Industrial Research (CSIR), South Africa
This page intentionally left blank
Contents

Preface..................................................................................................................xi

Introduction ...................................................................................................... xiii

Background.........................................................................................................xv

List of acronyms .............................................................................................. xvii

Chapter 1 ..............................................................................................................1
Highway safety evaluation terminology, framework and overview ......................1
1 Curvature Change Rate of the singular circular curve
with transition curves.....................................................................................1
2 Design classification based on accident and operating
speed research................................................................................................6
2.1 Relative accident numbers ....................................................................6
2.2 Design vs. safety ...................................................................................8
2.3 Design vs. speed .................................................................................10
2.3.1 Design speed (new alignments)...............................................11
2.3.2 85th-percentile speed................................................................12
2.3.3 Speed data collection and reduction........................................17
2.3.4 New speed developments ........................................................18

Chapter 2 ............................................................................................................23
Three quantitative safety criteria for highway geometric design.........................23
1 Classification of Safety Criterion I..............................................................23
2 Classification of Safety Criterion II ............................................................23
2.1 Evaluation of tangents in the design process ......................................25
2.2 Relation design ...................................................................................30
3 Classification of Safety Criterion III ...........................................................35
4 Safety Criteria vs. alignment design............................................................39
5 Safety module..............................................................................................40
Chapter 3 ............................................................................................................45
Comparative analyses of the actual accident situation with the results
of the Safety Criteria ............................................................................................45
1 Database: Schneider ....................................................................................46
2 Database: Ruscher .......................................................................................47

Chapter 4 ............................................................................................................49
Case studies..........................................................................................................49
1 Example I ....................................................................................................49
1.1 Results of the Safety Criteria ..............................................................49
1.2 Results of the safety module...............................................................59
2 Example II ...................................................................................................60
2.1 Results of the Safety Criteria ..............................................................60
2.2 Results of the safety module...............................................................70
3 Example III..................................................................................................70
3.1 Results of the Safety Criteria ..............................................................70
3.2 Results of the safety module...............................................................80
4 Example IV..................................................................................................83
4.1 Results of the Safety Criteria ..............................................................83
4.2 Results of the safety module...............................................................92

Chapter 5 ............................................................................................................95
Influence of road equipment on traffic safety ......................................................95
1 Pavement width ...........................................................................................96
2 Radius of curve............................................................................................98
3 Curvature Change Rate of the single curve.................................................98
4 Road equipment and design (Curvature Change Rate) classes ...................98
5 Road equipment and Safety Criteria..........................................................102

Conclusion and Outlook ..................................................................................103

References .........................................................................................................107

Index ..................................................................................................................111

Personal information .......................................................................................113

Conversion factors ...........................................................................................119


Preface

It is interesting to note that all analyzed highway geometric design guidelines include
at the beginning in one way or another the following sentence: The Guidelines are
the basis for the design of safe and functionally justified roads. If the guidelines
guarantee the safety of the road, then no or only a few accidents should occur on
that road. When accidents happen, drivers are always the ones who take the blame for
the mishap. When drivers fail a number of times at certain locations, then it becomes
obvious that the problem lies not with the drivers, but mainly with the geometry of the
road itself. Since accidents are not uniformly distributed on the road network, high
accident locations are a clear indication that, besides drivers error, there exist other
influencing parameters that are characterized by the road itself.
With respect to the development of guidelines and standards for highway geometric
design in many countries, it can be noticed that from 1940 to 1960 especially driving-
geometric and driving-dynamic models have been relevant, which were directed to
constant design speeds, however, traffic safety was only indirectly if at all regarded.
Since the mid 1960s questions about the actual speed behavior were emphasized
for the assessment of design parameters, however, traffic safety was again only indirectly
considered. Nevertheless, many experts recognize the fact that abrupt changes in
operating speeds lead to accidents, particularly on two-lane rural roads, and that these
speed inconsistencies may be largely attributed to abrupt alignment changes. Thus, to
help ensure design consistency between design elements and to coordinate design
speed and operating speed became major research issues. So far, however, any
evaluation of a roads safety had been conducted more or less qualitatively. In this
connection it was safe to say from a traffic safety point of view that no one could
predict with great certainty, or prove by measure or number, where traffic accidents
might occur or where accident black spots might develop.
Keeping this in mind, a practical procedure, which considers safety rules and criteria
for the safety evaluation of new designs, redesigns, and Restoration, Rehabilitation,
and Resurfacing (RRR) projects, became of major international concern. This book,
entitled How to Make Two-Lane Rural Roads Safer, has been prepared in response
to the expressed need.
The new book presents a through practical and scientific approach to designing
highways for maximum safety. Based on original research plus scrupulously collected
data amassed over more than two decades by the main author, this important book
originates vital criteria for safe design and shows how best to achieve the lowest
possible accident risk.
The book incorporates a methodology for evaluating planned or existing highway
alignment designs with respect to their expected impact on traffic safety. The designer
is able to evaluate alternative designs in terms of the relative danger they will impose on
the traveling public. The operations engineer is able to prioritize highway improvement
strategies based on the expected improvement to traffic accident patterns. Engineers
are able to quantitatively predict the accident consequences of their proposed or
existing alignments by using this process and employing these criteria.
Application of the described methodology will support the achievement of quantified
measures of
design consistency,
operating speed consistency, and
driving dynamic consistency.
All three criteria are evaluated in terms of three ranges, described as good, fair
(tolerable), and poor, with cut-off values between the ranges. It has been proved
that the results of the safety criteria coincide with the actual accident situation prevailing
on two-lane rural roads. By using the good ranges for the three safety criteria, sound
alignments in plan and profile, which match the expected driving behavior of the
motorists, can be achieved. These may significantly reduce accident risk and severity.
Finally, for a simplified general overview of the safety evaluation process, for
example, for network investigations, the three safety criteria were combined in an overall
safety module.
It is known that signs and markings can improve the safety record of a road section.
However, the improvement is seldom substantial and certainly not to the level of
transforming a poor design to a good design. On the other hand, the developed
concept does improve safety and does not rely on signage to achieve this improvement.
The developed safety evaluation process has been accepted by the professional
highway engineering community as illustrated by the fact that numerous publications
and research reports deal with it and that several Road Agencies internationally have
adopted or referenced it in their geometric design guidelines.
In general, the book is an invaluable source of information for educators, students,
consultants, highway engineers, and scientists in the field of highway design and
traffic safety engineering on new and existing (old) two-lane rural roads, which
encompass in most countries about 90 per cent or more of the rural road network. The
authors give essential information on:
Design cases to avoid,
Examples of good and poor solutions,
Redesign of existing roads.
In addition, this valuable and necessary resource gives guidance in coordinating
safety concerns with important economic, environmental, and aesthetic considerations.

The Authors,
November 2007
Introduction

What has to be considered in establishing modern highway geometric design


recommendations remains an exciting, thought-provoking question in the field of
highway engineering. While several important goals in highway geometric design,
such as function, traffic quality (capacity), economy, and aesthetics, are reasonably
well understood today, deficiencies still exist in the proper analysis and evaluation
of the impact of highway geometric design on traffic safety.
Unfortunately, most people are unaware of how large a problem unsafe traffic
operations represents on a worldwide basis. The tragic consequence of traffic
accidents puts unsafe traffic operations on a par with war or drug use, as an
example of irresponsible social behavior that must change. This lack of awareness
and responsibility may be an important reason why more than 500,000 people are
killed per year or about one life every minute and over 15 million suffer injuries
as a result of road accidents every year worldwide. Of the millions who are injured,
tens of thousands are maimed for life.
The above numbers seldom appear in a newspaper or in a TV bulletin but they
actually summarize what happens in one year worldwide. A single airline crash, or
a maritime disaster, is front-page news and prompts a federal investigation. But,
death in a traffic accident remains, for the most part, an invisible slaughter.
It is estimated that nearly 60 per cent of highway fatalities occur on two-lane
rural roads outside of cities or towns. About half of these fatalities occur on curved
roadway sections. Accident research has consistently found that accident rates
on horizontal curves are 3 to 5 times higher than the accident rates on tangent
sections of rural two-lane highways. Generally speaking, curved roadway sections
and the associated transition sections present a great opportunity for reducing
accident frequency and severity. Therefore, it was the life-task of the main author
to reduce accidents brought on by excessive speeds inconsistent with roadway
conditions or geometry.
Many of these speed errors may be related to inconsistencies in the horizontal
alignment, which cause the driver to be surprised by sudden changes in the roads
characteristics leading him or her to exceed the critical speed of a curve hence
losing control of the vehicle. These inconsistencies can and should be controlled
by the engineer. Since two-lane rural roads exhibit higher accident rates and severity
than multilane highways, special emphasis should be placed on this portion of the
road network when designing, redesigning, or conducting restoration, rehabilitation,
or resurfacing projects.
To improve the highway engineers ability to analyze rural roads and to provide
safer designs, three quantitative safety criteria have been developed [1], which,
when properly applied, are intended to provide rural two-lane highways with:

Design consistency,
Operating speed consistency, and
Driving dynamic consistency.
Background

Geometric design guidelines have long been the subject of dispute in the literature.
Some argue that the guidelines do not present a clear measure for evaluating the
safety level of roadways. Many authors have expressed concern over the lack of
quantitative safety considerations in the highway geometric design guidelines of
the last several decades. Some of them, for example, state:

Feuchtinger and Christoffers [2]


When a road goes into operation, the accident experience afterwards is the only
indicator of the safety performance of the road. During the planning stage, there is
no way to tell what level there is for traffic safety?

Bitzl [3]
Unlike other engineering fields, in road design it is almost impossible to determine
the safety level of a road. In other words, the guidelines provide no basic values to
describe the safety level of a road in relation to design parameters and traffic
conditions; whereas in other engineering fields, such as structural, there exist safety
criteria for constructing, for example, bridges or buildings.

Krebs and Kloeckner [4]


- If the guidelines guarantee the safety of a road, then no or only a few
accidents should occur on that road. When accidents happen, drivers are always
the ones who take the blame for the mishap.
- Accidents are not uniformly distributed on the road network. High accident
locations are clear indication that, besides drivers error, other influencing parameters,
which are characterized by the road itself, exist.

Mackenroth, cited in [4]


- No-one is in a position to state whether or not a drivers discipline was in
order in advance of a high accident location, but then failed at that location. When
a driver fails at a high accident location, it is often aid that it was his way of driving
that caused the accident.
- When drivers fail a number of times at certain locations, it then becomes
obvious that the problem lies not with the drivers, but mainly with the geometry of
The preceding statements indicate that no one is in a position to state whether
a road section of considerable length is safe or not, nor can anyone guarantee that
a road section will provide a minimum level of safety or a maximum level of
endangerment, that is, unsafety.
The safe and efficient movement of traffic is greatly influenced by the geometric
features of the highway. A review of accident spot maps normally shows that accidents
tend to cluster on curves, particularly on very sharp curves. In this connection it is
a fact, that two-lane rural roads pose the highest accident risks and severities.
Therefore, this portion of the road network should be given special emphasis, when
designing, redesigning, as well as conducting restoration, rehabilitation or
resurfacing (RRR) projects for these roads.
Many experts believe that abrupt changes in operating speed lead to accidents
on two-lane rural roads and that these speed inconsistencies may be largely brought
about by abrupt changes in road characteristics. Thus, providing longer roadway
sections with relatively consistent alignment, and thereby achieving a more consistent
driving behavior, is an important step toward reducing critical driving maneuvers,
leading to less hazardous road sections, and enhanced traffic safety on two-lane
rural highways. The same is true for the transitions between individual design
elements, like, for example, from tangent to curve or from curve to curve.
Therefore, it appeared necessary to develop a practical procedure that considers
driving behavioral and safety rules for the evaluation of new designs, redesigns,
and Rehabilitation Restoration, Resurfacing projects. On the basis of large operating
speed and accident databases in Europe and the USA, design classes were developed
to classify, from a traffic safety point of view, roadway sections as good, fair or poor
designs. These design classes are evaluated in terms of three safety criteria that
examine design, operating speed, and driving dynamic consistency for new designs
and existing alignments with respect to two-lane rural roads [1].
List of acronyms

a acceleration / deceleration [m/sec2]


AADT average annual daily traffic [veh. per day]
ACR accident cost rate [monetary unit per 100 veh.-km]
AR accident rate [acc. per 106 veh.-km]
CCRS curvature change rate of the single circular curve with transition
curves [gon/km]
CCRS average curvature change rate of the single circular curve with
transition curves
for the observed roadway section [gon/km]
DC degree of curve [deg./100 ft] or [deg./100 m]
e superelevation rate [%]
fRA side friction assumed [-]
fRD side friction demanded [-]
fT friction factor in tangential direction [-]
G longitudinal grade [%]
L length of curve or section [m] or [km]
Lcl length of clothoid [m]
Lcr length of circular curve [m]
n utilization ratio of side friction [%]
R radius of circular curve [m]
R2 coefficient of determination [-]
S sight distance [m]
S sum of all property and personal damages in the time period
observed
[monetary unit of the country under study]
T length of the investigated time period [years]
TL tangent (transition) length between two successive curves [m]
TLmax necessary acceleration/deceleration length to reach V85Tmax between
curves 1 and 2 [m]
TLmin necessary acceleration/deceleration length between curves 1 and 2
[m]
Vd design speed [km/h]
V85 85th-percentile speed of passenger cars under free flow conditions
on clean,
wet road surfaces [km/h]
V85T 85th-percentile speed in tangent [km/h]
V85Tmax maximum 85th-percentile speed in tangent [km/h]
V85 average 85th-percentile speed for the observed roadway section
without considering tangents. V85 represents a good estimate for
the design speed (Vd) of existing (old) alignments [km/h]
Chapter 1

Highway safety evaluation terminology,


framework and overview

On the basis of numerous research works, fundamental relationships could be


analyzed and evaluated between highway geometric design, driving behavior,
driving dynamics and accident situation. These form the basis for the
development of three quantitative safety criteria for the evaluation of the
endangerment of two-lane rural roads with respect to new designs, redesigns,
RRR-projects, and existing alignments. The criteria enable the examination of:
- Consistency in the alignment (Safety Criterion I), which corresponds to a
reasonable agreement between the design speed and the actual driving
behavior, expressed by the 85th-percentile speeds of passenger cars under
free flow conditions, thereby achieving a consistent road characteristic;
- Consistency in the operating speed (Safety Criterion II), which corresponds
to a limited variation between the 85th-percentile speeds of successive
design elements, thereby achieving a more consistent driving behavior;
- Consistency in the driving dynamics (Safety Criterion III), which represents
a favorable correspondence between side friction assumed with respect to
the design speed and side friction demanded with respect to the 85th-
percentile speed.
These safety criteria support the design engineer with respect to the
classification of roadway sections according to good (sound), fair (tolerable,
endangered), and poor (dangerous) design practices [1].

1 Curvature Change Rate of the Single Curve with Transition


Curves

Research, which evaluated the impact of design parameters (curvature change


rate of the single curve, length of curve, superelevation rate, lane width, shoulder
width, sight distance, longitudinal grades) and traffic volume between 1000 and
12,000 vehicles per day on two-lane rural highway sections in the United States,
2 How to make two-lane rural roads safer

Germany, Greece and Italy demonstrated, that the most successful parameter in
explaining much of the variability in operating speeds and accident rates was the
design parameter Curvature Change Rate of the Single Curve. All the other
design parameters tested were not significant in the regression models at the
95%-level of confidence.

The formula for determining the curvature change rate of the single curve with
transition curves is given by the following equation [1, 5, 6]:

L L L L L L
( Cl1 + Cr + Cl2 ) ( Cl1 + Cr + Cl2 )
2R R 2R 200 3 2R R 2R
CCR S = 10 = 63, 700
L L

conversion factor (1)

where:
CCRS = curvature change rate of the single circular curve with
transition
curves [gon/km],
L = LCl1 + LCr + LCl2 = overall length of unidirectional curved
section [m],
LCr = length of circular curve [m],
R = radius of circular curve [m],
LCl1, LCl2 = lengths of clothoids (preceding and following the circular
curve), [m]
63,700 = 200/ 103.

(The dimension gon corresponds to 400 degrees in a circle instead of


360 degrees according to the new European definition.)
The general graphical definition of CCRS is given under sketch a), whereas
equations for determining CCRS for different design cases are given under
sketches b) to e), exemplarily.
After the calculation of the new design parameter CCRS, for a unidirectional
curved site according to fig. 1 the 85th-percentile speed V85, can be directly
estimated from the operating speed background of fig. 3, with respect to an
individual country or according to the equations given in Section 2.3.2 of this
book.
Furthermore, illustrative sketches for determining CCRS along the course of a
roadway are given in fig. 2, cases a) to e), with respect to the combination of
curved sites and independent or non-independent tangents. (An independent
tangent is classified as one that is long enough to be considered in the curve-
tangent-curve safety evaluation design process, as an independent design
element. While a short tangent is called non-independent and can be neglected.
For defining and classifying independent tangents, see Section 2.1.)
Highway safety evaluation terminology, framework and overview 3
4 How to make two-lane rural roads safer

Figure 1: Sketches and equations for determining CCRs for design cases a) to e)
[1]

Note that individual elements (radii of curve and clothoids) of unidirectional


curvature are referred to as element sequences or curved sites according to
fig. 1. Correspondingly, directional changes separate those element sequences
into right-handed (+) and left-handed (-) element sequences (curved sites or
tangents) according to fig. 2.
Thus, independent tangents separate element sequences and form their own
element sequence, as shown in fig. 2. It is to be noted that compound curves
represent one element sequence if Rmax 3Rmin. However, if this condition is not
met, then two or even more element sequences with unidirectional curvature
have to be formed.
Highway safety evaluation terminology, framework and overview 5
6 How to make two-lane rural roads safer

Figure 2: Systematic sketches for determining CCRS for design cases a) to e).
(Note that design cases a) to e) in fig. 2 are different from design cases
a) to e) in fig. 1.) [1]

2 Design classification based on accident and operating speed


research

The design features of a roadway that are relevant for the driving behavior and
the accident situation along a continuous roadway section, represent the road
characteristics. Relevant design features are alignment, cross section, design of
intersections and accesses to adjacent properties. With respect to alignment
design, the new parameter Curvature Change Rate of the Single Curve has
been found to be the best predictor of driving behavior and accident potential.
Many investigations have shown a significant correlation between the parameter,
CCRS, and observed operating speeds, as well as accident rates and accident cost
rates [1].

2.1 Relative accident numbers

In evaluating the safety of alternative highway design concepts, it is the accident


rate or accident severity that is more relevant than the absolute number of
accidents. While it may be possible to predict numbers of accidents for a planned
highway section, it makes more sense to express the safety performance in terms
of accidents per vehicle-km traveled or average cost per accident.
For this reason, relative accident numbers, such as accident rate and accident
cost rate, which consider the length of a roadway section and the traffic volume,
allow a direct comparison of different roadway sections with respect to traffic
safety.
Highway safety evaluation terminology, framework and overview 7

Accident rate
The accident rate relates the number of accidents that occur on an investigated
section in a given time period to vehicle-kilometers traveled on the section. The
accident rate is determined by the following equation:
6
accidents 10 6
AR = accidents per 10 vehicle kilometers, (2)
AADT 365 T L

where
AR = accident rate
AADT = average annual daily traffic, (vehicles/24 h)
L = length of the investigated section, km
T = length of the investigated time period, yr
365 = number of days/yr.

Accident cost rates


The accident rate evaluates all accidents equally and does not differentiate them
by accident severity. Therefore, influences according to accident severity and,
especially, accident costs cannot be described by the accident rate. While the
accident rate represents the individual risk of being involved in an accident, the
accident cost rate additionally quantifies and compares the accident severity
using cost estimates. Therefore, the accident cost rate provides a measure that
quantifies the accident danger (risk) in terms of monetary units.
The accident cost rate, which represents the sum of property and personal
injury, is calculated from the following formula:
S 100
ACR = monetary unit per 100 vehicle kilometers (3)
AADT 365 T L

where
ACR = accident cost rate
S = sum of all property and personal injury in the time period T
observed, (monetary unit of the country under study)
AADT = average annual daily traffic, (vehicles/24 h)
L = length of the investigated section, km
T = length of the investigated time period, yr
365 = number of days/yr.

For determining accident costs, the property and personal damages have to be
calculated separately. The property damages are estimated by the police at the
accident scene and are typically listed in the accident report. Based on economic
losses, the costs for quantifying personal injury were evaluated by different
authors and are based on different economic costs with different amounts. For
example, the costs for personal damages were estimated by the German Federal
Highway Research Institute (BASt) in 1998 for:
8 How to make two-lane rural roads safer

- Fatalities at 2,358,000 German Marks (1 German Mark 0.51 EURO)


- Serious injuries at 161,000 German Marks
- Slight injuries at 7300 German Marks.

These numbers are actualized once a year.


In the relationship between road design and accident characteristics, the
speed is of major importance. Therefore, to evaluate accident rates and accident
cost rates specifically those accident types that are caused by false speed
estimations or excessive speed errors have to be regarded in this study. The
accident types Run-off the road (ROR) or Head on/rear end are examples,
although the latter may be of minor importance. According to Tables 1 and 2 the
accident types ROR and Deer were mainly used to detect possible
correlations between highway road geometric design and accident situation.

2.2 Design vs. safety

To get a better overview of the real accident situation the Curvature Change
Rate of the Single Curve was arranged into different design, respectively CCRS-
classes, for 5 databases, one from the USA., and four from Germany, which
fundamentally all reveal similar results. The results of three databases, which
reveal the highest number of Test Sites, are listed in Table 1 for the accident rate.
For every design (CCRS)-class a mean accident rate was calculated.
The selected ranges of the CCRS-classes from 180 to 360 gon/km and from
360 to 550 gon/km go back to the original investigations in the United States,
which were related to the US design parameter degree of curve. The
conversion of the original ranges of the DC classes (DC = 5 to 10 deg/100 ft
and DC = 10 to 15 deg/100 ft) leads to the selected CCRS-classes in Table 1.
As shown in Table 1, the t-Test results indicate significant increases (at the
95% level of confidence) in the mean accident rates between the different CCRS-
classes compared; which means higher accident rates can be expected with
higher CCRS-classes, despite the stringent traffic warning devices often installed
at curve sites.
The significant results of Table 1 indicate for three databases and different
accident types:
1) gentle curvilinear horizontal alignments consisting of tangents or transition
curves, combined with curves up to CCRS-values of 180 gon/km (that
corresponds roughly to radii of curve of greater or equal than 350 m)
experienced the lowest average accident risk, classified here as good
design;
2) the accident rate on sections with CCRS-values between 180 and
360 gon/km (which means roughly radii of curve between 175 and 350 m)
was at least twice to three times that on sections with CCRS-values of up to
180 gon/km, classified here as fair (tolerable) design;
3) the accident rate on sections with CCRS-values between 360 and
550 gon/km (databases 1 and 2) was about four to five times higher than that
on sections with CCRS-values of up to 180 gon/km, classified here as poor
design;
Highway safety evaluation terminology, framework and overview 9

Table 1: t-Test results of mean accident rates for different CCRS-classes for
Germany (West) and for the USA [1, 7]
Design/
CCRS-classes Mean AR tcalc. tcrit. Significance; Remarks
[gon/km]
Database 1 : United States of America (261 two-lane rural test sites,
1987 including all accidents
Considered as
tangent (0) 1.17 --Good Design
4.00 > 1.96 Yes
35 180 2.29 --Good design
7.03 > 1.96 Yes
> 180 360 5.03 --Fair design
6.06 > 1.99 Yes
> 360 550 10.97 --Poor design
3.44 > 1.99 Yes
> 550 990 16.51 --Poor design
Database 2: Germany (657 two-lane rural test sites), 1994 including
run-off-the-road, and deer accidents
Considered as
tangent (0) 0.35 --Good Design
5.20 > 1.99 Yes
35 180 0.51 --Good design
10.70 > 1.96 Yes
> 180 360 1.72 --Fair design
2.64 > 1.98 Yes
> 360 550 2.78 --Poor design
Database 3: Germany (2726 two-lane rural test sites), 2001 including
run-off-the-road, and deer accidents
Considered as
0 - 180 0.22 --Good design
27.92 > 1.65 Yes
> 180 - 360 0.87 --Fair design
15.69 > 1.65 Yes
> 360 2.27 --Poor design
Database 3: Germany (2726 two-lane rural test sites), 2001 including
run-off-the-road, head-on, and deer accidents
Considered as
0 180 0.33 --Good design
28.04 > 1.65 Yes
> 180 - 360 1.12 --Fair design
14.09 > 1.65 Yes
> 360 2.52 --Poor design
Legend:
AR = accident rate (acc. per 106 veh.-km).
Deer = animal
10 How to make two-lane rural roads safer

4) for CCRS-values greater than 550 gon/km (corresponding roughly to radii of


curve of less than 115 m), the average accident rate was even higher,
particularly, as demonstrated by database 1 (Table 1).
With respect to the differences in the accident rates for the different databases
note that databases 1 and 2 contain only those elements, respectively, element
sequences where at least one accident has occurred, whereas database 3 also
considers those elements, where no accident has occurred. In this way, the
differences in the accident rates between databases 1 or 2 and database 3 become
understandable.
The relationships between design (CCRS)-classes and accident rates, shown
in Table 1, are confirmed by the results of Table 2 with respect to the accident
cost rate. As can be seen from this table, the t-Test results also revealed
significant increases in the accident cost rates between the different CCRS-
classes compared. Thus, mean accident rates and accident cost rates of large
databases prove that, with increasing CCRS-ranges or classes, accident risk and
severity increase significantly.
Based on the presented results, it can be assumed that the proposed design
(CCRS)-ranges of Tables 1 and 2 represent a sound classification system for the
arrangement of good, fair and poor design practices in modern highway
geometric design. This finding was the basis for the development of the three
quantitative safety criteria in Chapter 4 for evaluating alignment design with
respect to
1) achieving design consistency (Safety Criterion I),
2) achieving operating speed consistency (Safety Criterion II), and
3) achieving driving dynamic consistency (Safety Criterion III).

2.3 Design vs. speed

Based on the literature review in [1], which covers the highway geometric design
guidelines and research of many countries, it can be concluded that two different
speeds are used to govern highway geometric design: design speed and operating
speed. The design speed is the basis for determining design-related parameters
such as horizontal curves, grades, and vertical curves. Operating speed considers
the expected actual speed behavior, and is usually expressed as the 85th-
percentile speed. In general, this speed is used to determine operating speed-
related design parameters such as sight distances, crest vertical curves,
superelevation rates, etc.
Even though design speed has been used for several decades to determine
sound horizontal alignments, it is possible to induce certain inconsistencies into
the highway alignment. At low and intermediate design speeds, sections of
relatively flat alignment, interspersed between the controlling curvilinear
portions, may produce operating speed profiles that exceed the design speed in
the curvilinear sections by substantial amounts. This is especially true on two-
lane rural roads.
Highway safety evaluation terminology, framework and overview 11

Table 2: t-Test results of mean accident cost rates for different CCRSclasses
for Germany (West), [7]
Design/ CCRS-
classes
Mean ACR tcalc. tcrit. Significance; Remarks
[gon/km]

Database 3: Germany (2726 two-lane rural test sites), 2001


including run-off-the-road and deer accidents
Considered as

0 - 180 4.31 --Good design


17.15 > 1.65 Yes
180 - 360 13.61 --Fair design
8.68 > 1.65 Yes
> 360 31.02 --Poor design

Database 3: Germany (2726 two-lane rural test sites), 2001


including run-off-the-road, head-on and deer
accidents
Considered as

0 - 180 6.04 --Good design


17.75 > 1.65 Yes
180 - 360 17.37 --Fair design
7.44 > 1.65 Yes
> 360 33.55 --Poor design

Legend:
ACR = accident cost rate (German Marks).

2.3.1 Design speed (new alignments)

The design speed Vd reflects environmental and economic conditions based on


the assumed network function of the road and the desired quality of traffic flow,
and should be assessed for new designs according to the classification in Table 3
of road-category functions. Limiting and standard values for most of the design
elements are defined according to the design speed. The design speed
determines, for example:
- minimum radii of curve and minimum parameters of clothoids,
- maximum longitudinal grades,
- required parameters for vertical curves,
within a specific roadway section. The design speed decisively influences the
road characteristic, traffic safety, the quality of traffic flow, as well as the costs
of rural roads of Category A I to A V (Table 3).
12 How to make two-lane rural roads safer

Therefore, for new designs a constant design speed should be applied


consistently on longer sections, or at least on longer connected roadway sections,
and can be selected from Table 3 depending on the function of the road in the
network. Furthermore, the design speed is used in the safety-evaluation process
of highway geometric design according to Safety Criterion I (see Chapter 4,
Section 1).
The design speed is classified from Statewide or Interstate functions (Road
Category A I), via regional functions (Road Category A II), and functions
between municipalities (Road Category A III), to large-area accessibility
functions (Road Category A IV), and, finally, to subordinate connections (Road
Category A V), [1].
However, for existing (old) alignments the design speed is often not known
and certainly was not selected according to the assumptions of Table 3 with
respect to a specific network function of the road and a desired traffic quality.
Therefore, for the selection of a sound design speed, especially for redesigns and
RRR-strategies in cases of existing or old alignments, a new procedure was
developed. This procedure selects the design speed on the basis of the actual
driving behavior on the observed roadway section. The procedure will be
explained in detail in Section 2.3.4.

2.3.2 85th-Percentile Speed

The operating speed considers the expected actual speed behavior, and is usually
expressed by the 85th-percentile speed. For the Road Categories AI to AV
V85 corresponds to the speed, below which 85 per cent of passenger cars
operate under free-flow conditions on clean, wet road surfaces.
Based on V85:
- the superelevation rates in circular curves, and
- the necessary stopping and passing sight distances
are determined and evaluated, for example.

The 85th-percentile speed depends on the road geometry, and is also used for
the safety evaluation process according to Safety Criteria I to III, developed in
Chapter 4. With respect to the term wet road surfaces discussions, analyses
and evaluations are given in [1]. Based on these investigations, it is concluded
that the 85th-percentile speed on dry road surfaces does not differ significantly
from the 85th-percentile speed on wet road surfaces with rainfall varying
between a sprinkle and moderately heavy rain, and when visibility is not
appreciably affected. This assumption corresponds to a sight distance of about
150 m. This means, for the critical design case, that high speed levels are still
observed on wet road surfaces. Thus, the established 85th-percentile speeds are
valid for dry and wet road surfaces considering the above assumptions.
Table 3: Classification of Roads by Categories (Applicable for Most Countries)

Design and operational characteristics


Road Category Kind of Cross Intersection Design Speed
traffic section access Vd [km/h]
AI Statewide or Vehicles Multiple Controlled 130 120 110

Highway safety evaluation terminology, framework and overview 13


interstate Vehicles lane (Controlled) 100 90 80
functions 2+1 lane free
A II Regional Vehicles Multiple (Controlled) 120 110 100 90 80
functions All* lane free 100 90 80 70
2 or 2+1 Free
lane
A Functions Vehicles Multiple (Controlled) 90 80 70
III between All* lane free 90 80 70 60
municipalitie 2 lane Free
s
A Large area All* 2 lane Free 80 70 60 50
IV accessibility
functions
AV Subordinate All* 2 lane Free 70 60 50
connection

Legend:
* Indicates All Types of Road User Groups Combined
Exceptional Values (italic)
14 How to make two-lane rural roads safer

For determining operating speeds, as defined by the 85th-percentile speeds


with respect to the previously discussed design parameter Curvature Change
Rate of the Single Curve (CCRS), the operating speed backgrounds for a
number of countries are illustrated in fig. 3. As can be seen, the 85th-percentile
speeds decrease with increasing CCRS-values. The regression equations for the
curves of fig. 3 are also listed in Table 4.
When searching for reliable safety ranges with respect to operating speeds
(V85i) it should be noted according to fig. 3, that changes in the CCRS-values of
between 180 and 200 gon/km between successive curves or between independent
tangents (CCRS = 0 gon/km) and curves correspond to differences in the 85th-
percentile speeds of, or greater than, 10 km/h.
For changes in CCRS-values of between 360 and 400 gon/km, differences in
the 85th-percentile speeds equal to or greater than 20 km/h can be expected.
Since the CCRS-changes for the above-observed speed differences
respectively of 10 and 20 km/h approximately coincide to the design (CCRS)-
classes of Tables 1 and 2, which are based on accident research, the classification
system of Table 5 could be developed for evaluating operating speed
consistency.
This suggests that, if a speed difference of V85 10 km/h between two
successive design elements (CCRS 180 gon/km) is observed, a good design
level with a low accident risk can be expected as shown in Tables 1 and 2. For
such road sections, consistency in horizontal alignment exists between
successive design elements, and inconsistencies are not created in vehicle
operating speeds. No adaptations or corrections are necessary.
In Table 5 speed differences between 10 and 20 km/h represent a fair
(tolerable) design level with significantly higher accident risks in comparison to
the good design level, as Tables 1 and 2 reveal. These road sections may
represent at least minor inconsistencies in geometric design between successive
design elements. Normally, they would warrant speed regulations and/or traffic
warning devices, but not necessarily a redesign, unless a documented safety
problem exists.
For speed differences of V85 > 20 km/h poor design practice definitely
exists with high accident rates as the databases, compiled in Tables 1 and 2,
clarify. These road sections reveal strong inconsistencies in horizontal geometric
design between successive design elements. These inconsistencies, when
combined with breaks in the speed profile, lead to critical accident frequencies
and severities, and thereby to an uneconomic and unsafe operation. Under these
circumstances, redesigns would normally be recommended.
By knowing the CCRS-values for the curved roadway sections and
independent tangents (CCRS = 0 gon/km), the 85th-percentile speeds can be
determined from fig. 3 or Table 4 for the respective country under study. For
example, at a curved site with a CCRS = 200 gon/km, the expected 85th-
percentile speed for the Greek operating speed background is about
V85 = 85 km/h; or, for independent tangents, operating speeds of about
100 km/h can be expected on two-lane rural roads in Greece.
130
120
110
Germany, ISE
Average Australia
100 Germany, ISE
France
85th-Percentile Speed V85 [km/h] Greece

Highway safety evaluation terminology, framework and overview 15


Italy
90
U.S.A.
80 Lebanon France
70 Australia
Canada
60 USA Lebanon
Greece Canada
50
Italy
40
Average
30
20
10
0
0 50 100 150 200 250 300 350 400 450 500 550 600
CCRs [gon/km]

Figure 3: Operating speed backgrounds for two-lane rural roads in different countries for longitudinal grades G 6%
16 How to make two-lane rural roads safer

Table 4: Regression models for operating speed backgrounds for two-lane rural
roads in different countries [1]

Germany, ISE
V85 = 106 /(8270 + 8.01 CCRS) R2 = 0.73 (4a)
Speed Limit: 100 km/h
Greece
V85 = 106 /(10150.1 + 8.529 CCRS) R2 = 0.81 (4b)
Speed Limit: 90 (100) km/h
U.S.A.
V85 = 103.04 0.053 CCRS R2 = 0.80 (4c)
Speed Limit: 90 km/h
France
V85 = 102 /[1 + 346 (CCRS / 63,700)1.5] (4d)
Speed Limit: 90 km/h
Australia
V85 = 101.2 0.043 CCRS R2 = 0.87 (4e)
Speed Limit: 90 km/h
Lebanon
V85 = 91.03 0.056 CCRS R2 = 0.81 (4f)
Speed Limit: 80 km/h
Canada
(4.5615.27 104 CCRS )
V85 = e R2 = 0.63 (4g)
Speed Limit: 90 km/h
Italy [8]
V85 = 118.9 0.062 CCRS R2 = 0.94 (4h)
Speed Limit: 90 km/h
Highway safety evaluation terminology, framework and overview 17

Table 5: Classification System for the Safety Evaluation Process: Operating


Speed Consistency [1, 7]

Safety evaluation process for operating speed consistency


Case 1: Good
design level Case 2: Fair design level Case 3: Poor design level

Permissible differences: Tolerated differences: Non-permissible differences:


CCRSi - CCRSi+1 180 180 < CCRSi - CCRSi+1 360 CCRSi - CCRSi+1 > 360
V85i - V85i+1 10 10 < V85i - V85i+1 20 V85i - V85i+1 > 20

Legend:

CCRSi = CCRS-value of design element i [gon/km],


CCRSi+1 = CCRS-value of design element i+1 [gon/km],
V85i = 85th-percentile speed of design element i [km/h],
V85i+1 = 85th-percentile speed of design element i+1 [km/h].

Operating speed backgrounds (like those in fig. 3 or Table 4) should be part


of every modern highway design guideline when striving for a good curvilinear
alignment, and for a more consistent and safer road characteristic. In this way,
operating-speed consistency can be achieved through the good design level
according to Table 5. Minor inconsistencies correspond to the fair design level,
whereas major design inconsistencies correspond to the poor design level.

2.3.3 Speed data collection and reduction

Since operating speed backgrounds are not yet available for many countries,
some proposals about data collection of operating speeds and design parameters
are offered here.
Data collection is broken down into three phases:
- the selection of road sections appropriate for the study;
- the collection of as much field data as possible about the road sections; and
- the measurement of operating free speeds at each section.
Selection of appropriate road sections. Site selection should be limited to
sections with the following characteristics:
1. A tangent to curve or curve to curve section.
2. Removed from the influence of intersections.
3. Without physical features adjacent to or in the roadway that may create
abnormal hazards.
4. Average annual daily traffic between 1000 and 12,000 vehicles/day.
It is desirable to maintain a regional distribution and, at the same time, retain
the longest road segments in the selection process. The road sections selected
18 How to make two-lane rural roads safer

should provide the widest range of changes in alignment that can be found by
observation.

Field data collection. This phase involves obtaining as much data in the field as
possible about the road sections specifically about the curves or curved
sections within the observed road section. Information recorded should include
radius of curve, length of curve, superelevation rate, gradient, lane width,
AADT, etc.

Speed data collection and reduction. In order to ensure that the speeds measured
represent the free speeds desired by the driver under a set of roadway conditions
and are not affected by other traffic on the road, only the speeds of isolated
vehicles with a minimum time gap of about 6 seconds should be measured.
Speed measurements should be made during daytime hours on weekdays under
both dry and wet pavement conditions.

The basic method used for speed data collection requires the measurement of
the time required for a vehicle to traverse a measured course laid out in the
center of a curve. Speed measurements would also be taken on preceding and
succeeding tangents to the curved site. The length of the course is 50 m. The
method used for measuring time over the measured distance involves the use of
transverse pavement markings at each end of the course, and an observer who
starts and stops an electronic stop watch as a vehicle passes the markings. The
observer should be located at least 5 meters from the pavement edge of the road
to ensure that his or her presence would not influence the speeds of passing
vehicles, but not so far away that inaccurate measurements occur. Normally, the
observer would be located at a position, where she or he could not directly be
seen by the drivers of traversing vehicles, alternatively be disguised as
construction or forest workers, etc.
Satisfactory speed data can be obtained for both directions of travel by
applying this procedure. Because of cost, time, and personal constraints, only
about 80 to 100 passenger cars under free-flow conditions need to be sampled at
each site for both directions of traffic. Speed data can then be analyzed to obtain
the operating speed, expressed herein by the 85th-percentile speed.

2.3.4 New speed developments

85th-percentile speed
Based on additional research, in completion to the Highway Design and Traffic
Safety Engineering Handbook [1], sound relationships between the curvature
change rate of the single curve (CCRS) and the 85th-percentile speed (V85) can
be developed for worldwide application with respect to the longitudinal grade, if
a specific operating speed background for the country under study is not
available. The future equations read, as follows:
Highway safety evaluation terminology, framework and overview 19

V85 for longitudinal grades G 6 % and CCRS-values between 0 and


600 gon/km
N. B. For practical design tasks it is recommended that the upper limit of
CCRS 1600 gon/km should not be exceeded. This value corresponds to a radius
of curvature of R = 40 m. Such radii or smaller should not be regarded in the
safety-evaluation process, since for these radii of curvature (for example:
R 30 m) the geometric influences prevail over the driving dynamics. As such
typical operating speeds, which are important assumptions for Safety Criteria I to
III, are not represented.

(Proposal for worldwide application, if the operating speed background of the


country under study is not available [7, 9]).
5 2
V85 = 105.31 + 2 10 CCR S 0.071 CCR
S
R 2 = 0.98 (5)

V85 for longitudinal grades G > 6 % and CCRS-values between 0 and


1600 gon/km
N. B. For practical design tasks it is recommended that the upper limit of
CCRS 1600 gon/km should not be exceeded. This value corresponds to a radius
of curvature of R = 40 m. Such radii or smaller should not be regarded in the
safety-evaluation process, since for these radii of curvature (for example:
R 30 m) the geometric influences prevail over the driving dynamics. As such
typical operating speeds, which are important assumptions for Safety Criteria I to
III, are not represented.
(Proposal for worldwide application, if the operating speed background of the
country under study is not available [10]).

9 3 5 2 2
V85 = 86 3.24 10 CCR S + 1.61 10 CCR 4.26 10 CCR S
S
2
R = 0.88 (6)
Based on the presented results of accident- and operating speed research in
Tables 1, 2, 4 and 5 it can be assumed that the proposed CCRS-ranges and 85th-
percentile speed ranges represent a sound classification system for the
arrangement of good, fair and poor design practices in modern highway
geometric design.

Design speed

As already mentioned in Section 2.3.1, the design speed often is not known for
existing alignments. However, it was certainly not selected according to the
assumptions of Table 3 with respect to a specific network function of the road.
Many case studies, conducted at the Institute for Highway and Railroad
Engineering of the University of Karlsruhe, Germany, revealed that existing
20 How to make two-lane rural roads safer

(old) alignments were normally not constructed for an exactly defined design
speed according to the above-mentioned section.
Thus, in the case of redesigns and RRR strategies, achieving concord
between network function, quality of traffic flow, design speed and 85th-
percentile speed often leads to unfavorable results in respect of economic,
environmental and safety-related issues. Therefore, a new procedure for
determining sound design speeds for existing (old) alignments has been
developed.
For existing alignments it is possible to estimate an average curvature change
rate of the single curve CCRS for the observed roadway section, along which
the design speed should be relevant. The method is based on a length-related
calculation of the average of the CCRSi-values of the individual curves along the
observed roadway section disregarding the intervening tangents:
i= n
(CCR L )
Si i
CCR S = i=1 (7)
i=n
Li
i=1

CCRS = average curvature change rate of the single curve for the
observed roadway section without regarding tangents [gon/km],
CCRSi = curvature change rate of the single curve or unidirectional curved
site i [gon/km],
Li = length of curve or unidirectional curved site i [m].

The value of CCRS can be used for determining an average 85th-percentile


speed V85, based on the regression model for the operating speed background
of the country under study (fig. 3 or Table 4), respectively, the new eqns. (5) and
(6).
This so-called average V85 will be considerably exceeded in the case of
large radii of curvature or independent tangents, whereas, in the case of small
radii of curvature it will not be reached. However, since the design speed (Vd)
should be constant on longer sections, or at least on longer connected roadway
sections, it is reasonable to regard the estimated average 85th-percentile speed
(V85) of the observed existing alignment as the basis for the selection of a
meaningful design speed, Vd. This could be applied, for example, to future
redesigns or RRR-strategies.
For example, with respect to eqn. (5) the formula for the design speed would
be [7, 9]:
V85 Vd = 105.31 + 2 10-5 CCRS2 0.071 CCRS (8)
It can be assumed that, for such a design speed, over- and under-
dimensioning of existing elements can be avoided, that they even can be adapted
Highway safety evaluation terminology, framework and overview 21

to each other and can be optimized to a certain extent in terms of economic,


environmental and safety-related issues.
The above findings build the basis for the development of the three
previously cited quantitative safety criteria for evaluating the design of existing
alignments.
22 How to make two-lane rural roads safer
Chapter 2
Three quantitative safety criteria for
highway geometric design

1 Classification of safety criterion I

Numerous investigations have revealed that the design speed (Vd) is often
exceeded by the actual operating speed (V85). Especially on roadway sections
for which low design speeds were selected, this difference can become very
dangerous. These critical speed differences would be detected by Safety
Criterion I. For every element sequence (curved or independent tangent
section), Criterion I determines the speed difference between V85 and Vd
according to Table 6.
Achieving design consistency is of special interest in modern highway
geometric design. This means that the design speed (Vd) should remain constant
on longer roadway sections, and simultaneously should match the actual driving
behavior, expressed by the 85th-percentile speed (V85) of passenger cars under
free-flow conditions. To assess meaningful variations between design and
operating speed for Safety Criterion I, a broad literature review was conducted
with respect to existing guidelines and research, and the results are arranged in
Table 6. Furthermore, the outcome of Table 5 is additionally incorporated in
Table 6 to tune the speed ranges of Safety Criteria I and II to each other.
Additionally, the corresponding design (CCRS)-classes are added in Table 6.
As can be seen for the design level of Safety Criterion I, considered as
good, it is required that the difference between the 85th-percentile speed and
the design speed should not exceed 10 km/h along the whole observed roadway
section. Consequently, the road characteristic is well balanced for the motorist
along the entire length of the road section.
The design procedure according to Safety Criterion I is illustrated in the flow
chart of Figure 4 [5].

2 Classification of safety criterion II


A well-balanced operating speed sequence between successive design elements
(curved sites and/or independent tangents) promotes a consistent, economic
24 How to make two-lane rural roads safer

Table 6: Classification of safety criterion I [1, 7]

Classification
Design Symbol Speed difference [km/h] Design (CCRS)-class [gon/km]
Good + |V85i Vd| 10 |CCRSi CCRS| 180
Fair o 10 < |V85i Vd| 20 180 < |CCRSi CCRS| 360
Poor - |V85i Vd| > 20 |CCRSi CCRS| > 360

pattern. While Safety Criterion I compare the 85th-percentile speed (V85) with
the design speed (Vd) for each individual curved or tangent site, Safety
Criterion II evaluates the speed differenceV85i V85i+1between the 85th-
percentile speeds of successive design elements.
Based on the preceding analyses, the evaluations of driving behavior
(expressed by the 85th-percentile speeds), and the accident situation (expressed
by the accident rates, and cost rates), the ranges of differences in operating
speeds, shown in Table 7, are considered reasonable for the classification of
good, fair, and poor design levels for Safety Criterion II. This criterion was
developed for achieving operating speed consistency in horizontal alignment,
and is related to the transition between two successive design elements. The
differences in the 85th-percentile speeds in Table 7 match those of Tables 5 and 6.

Table 7: Classification of safety criterion II [1, 7] (see also Table 5)

Classification
Design Symbol Speed difference [km/h] Design (CCRS)-class [gon/km]
Good + |V85i V85i+1 | 10 |CCRSi CCRSi+1| 180
Fair o 10 < |V85i V85i+1 | 20 180 < |CCRSi CCRSi+1| 360
Poor |V85i V85i+1 | > 20 |CCRSi CCRSi+1| > 360

With respect to the CCRSi-values, the V85i values can be determined for the
operating speed background of the country under study (see fig. 3 or Table 4). If
no operating speed background is available, the designer should select one
appropriate to a comparable country. Alternatively, a new operating speed
background could be developed, or, as a last resort the normally worldwide valid
eqns. (5) or (6) could be used.
In Tables 6 and 7 the ranges for good, fair, and poor design levels for the
design parameter, Curvature Change Rate of the Single Curve, are also
incorporated. This is not necessary for practical design work, since the CCRS-
ranges and the V85-ranges correspond to one another within relatively narrow
limits, as previously discussed in respect to Table 5.
Three quantitative safety criteria for highway geometric design 25

Therefore, in order to simplify matters, only the ranges of the 85th-percentile


speeds between successive design elements are referred to in Table 11 for
distinguishing good, fair, and poor design practices. Case studies for the
evaluation process of Safety Criteria I to III are presented in Chapter 4.

2.1 Evaluation of tangents in the design process

As stated earlier, abrupt changes in operating speeds, created by the horizontal


alignment, are among the leading causes of accidents on two-lane rural roads.
Particularly at lower levels of design speed, the changing alignment may cause
variations in operating speeds that, in turn, may substantially increase the
accident risk. For this reason, the transition from a tangent to a curve (especially
an isolated curve) has to be considered as one of the most critical design cases.
Therefore, one of the important tasks in modern highway geometric design and
rehabilitation strategies for two-lane rural roads in many countries is to ensure
operating speed consistency and to detect critical inconsistencies in the
horizontal alignment.
According to Refs. 11 to 14 the tangent is considered as a dynamic design
element, by taking into account the longitudinal acceleration and deceleration
movements observed on tangents. In contrast to the tangent, the circular curve
has been considered since the 1920s as a dynamic design element with respect to
the lateral (centrifugal) acceleration as a driving dynamic input. Safety
Criterion II, achieving operating speed consistency is significant for the safety
evaluation of tangents and curves, in order to distinguish good, fair, and poor
design levels for a tangent-to-curve transition, especially on two-lane rural roads.
For the following considerations two types of tangents are relevant:
An independent tangent is classified as one that is long enough to be
considered as an independent design element in the curve-tangent-curve safety
evaluation design process, while a short tangent is called non-independent and
can be neglected.
The procedure for evaluating tangent speeds and lengths in the safety-
evaluation process is presented in fig. 5, [10]. For the classification of tangents in
the design process eqns. (9) to (11) were developed for the following three
design cases according to fig. (5):

Case 1: TL TLmin
(non-independent tangent, not considered in the safety evaluation
process, the sequence curve-to-curve is relevant).
Case 2: TL TLmax
(independent tangent, considered in the safety evaluation process,
the sequence tangent-to-curve or vice versa is relevant).

Case 3: TLmin < TL < TLmax


(independent tangent, considered in the safety evaluation process,
the sequence tangent-to-curve or vice versa is relevant).
26 How to make two-lane rural roads safer

CCRSi Li
n
(
CCRSi Li )
CCRS = i=1 n
CCRSi=0 L
i=1 i

yes no without considering tangent sections

tangent curved site

5 2
V85i = 105.31 + 2 10 CCR S 0.071 CCR S
V85i
according
to SC II Proposal for worldwide application, if the operating speed
background of the country under study is not available
(CCRS 1600 gon/km; G 6%)

V85i Vd10 km/h

yes no

V85i Vd20 km/h

yes
no

+ o V85i
good fair poor
design SC I design SC I design SC I

Selection of Vd
if Vd is unknown, tune Vd with
V85 by bringing up or down
V85 to the next round figure
in accordance with the road
characteristics
Figure 4: Design procedure according to safety criterion I
Three quantitative safety criteria for highway geometric design 27

The following formulae employ the fundamental equation for the


evaluation of the transition lengths between two successive curves according to
eqn. (9), see also fig. 5.

Case 1: For TL TLmin non-independent tangent, NIT (figs. 5 and 6):

2 2
(V851 ) (V852 )
TL min = (9)
2
2 3.6 a

2 2
(V851 ) (V852 )
TL min = (9a)
22.03

Based on car-following techniques, an average acceleration or deceleration


rate of a = 0.85 m/s2 was established.
In eqns. (9) and (9a), TL TLmin means that the existing tangent segment
can, at most, be of sufficient length for accelerating from the operating speed of
Curve 1 to that of Curve 2. In this case, the element sequence curve-to-curve,
and not the interim (non-independent) tangent controls the evaluation process
according to Safety Criterion II (see Table 7).

Case 2: For TL TLmax independent tangent, IT (figs. 5 and 6):


2 2 2 2
(V85 ) (V85 ) (V85 ) (V85 )
TL = Tmax 1 + Tmax 2 (10)
max 2 2
2 3.6 a 2 3.6 a

2 2 2
2 (V85Tmax ) (V851 ) (V852 )
TL max = (10a)
22.03

For the definition of the symbols, see fig. 5.

In eqns. (10) and (10a) TL TLmax means that the existing tangent segment is
long enough to allow an acceleration and deceleration maneuver up to the
maximum operating speed (V85Tmax) on tangents (see fig. 5).
In this case, the element sequences independent tangent-to-curve or curve-to-
independent tangent become relevant for the evaluation of Safety Criterion II in
Table 7.
28 How to make two-lane rural roads safer

V85Tmax

V85T (example)

V852

V851

TLmin

Case 1 Case 3 Case 2


TL

TLmax

Legend:

V851, 2 = 85th-percentile speeds in curves 1 and 2 [km/h],


V85Tmax = Maximum operating speed in tangents [km/h] for
CCRS = 0 gon/km
(depending on the regression equations in Table 4 for different
countries or for worldwide application according to Eqs. 5 and
6),
V85T = Operating speed in tangents [km/h]
(V85T can maximum reach V85Tmax),
TL = Existing tangent length between two successive curves [m],
TLmin = Necessary acceleration/deceleration length between curves 1
and 2 [m],
TLmax = Necessary acceleration/deceleration length to reach V85Tmax
between curves 1 and 2 [m].

Figure 5: Systematic sketches for determining tangent speeds and lengths in the
safety evaluation process [10]
Three quantitative safety criteria for highway geometric design 29

Case 3: For TLmin < TL < TLmax independent tangent, NIT (figs. 5 and 6):

2 2
TL TL min (V85T ) (V851 )
= for V85 > V85 (11)
2 22.03 1 2

2
V85T = 11.016 (TL TL min ) + (V85 ) (11a)
1

For the definition of the symbols, see fig. 5. Always use the larger value of
V851,2.
The existing tangent length lies somewhere between TLmin and TLmax.
Although the tangent segment does not allow accelerations up to the highest
operating speed (V85Tmax), additional acceleration and deceleration maneuvers
are possible (see fig. 5). In this case, the realizable tangent speed (V85T) has to
be calculated according to eqn. (11a) for the evaluation of Safety Criterion II.
Accordingly, Safety Criterion I has to be controlled in tangents with respect
to Table 6 for the differences between the design speed, Vd, and the 85th-
percentile speeds in independent tangents, V85T or V85Tmax, for design Cases 2
and 3. Corresponding to Safety Criterion II, non-independent tangents (Case 1)
are too short to be considered in connection with Safety Criterion I. Thus, the
tangent length also represents an important issue in highway geometric design.
For a better understanding with respect to the previously discussed three design
cases, the following examples are given in Table 8.

Table 8: Example applications for evaluating tangents in the design process [10]

Eg. V851 V852 V85Tmax TL TLmin TLmax V85T


No. km/h km/h km/h m m m Case km/h

1) 70* 45* 86* 100 130 357 1

2) 95* 85* 100* 200 82 170 2 100

* Calculated based on the corresponding regression equations in Table 4 or eqns. (5) and (6) for
the individual CCRS value of the observed curved or tangent site.
30 How to make two-lane rural roads safer

Safety evaluation for relevant transitions according to Criterion II in Table 7:

1) Worldwide, mountainous, operating speed background according to Eq. 6


TL TLmin case 1 non-independent tangent
Curve 1 curve 2:V851 V852 = 70 45 = 25 km/h poor design

2) Greece, operating speed background according to Table 4 and Eq. 4b


TL TLmax case 2 independent tangent
Curve 1 tangent:V851 V85Tmax = 95 100 = 5 km/h good design
Tangent curve 2:V85Tmax V852 = 100 85 = 15 km/h fair design

3) Italy, operating speed background according to Table 4, and Eq. 4h


TLmin < TL < TLmax case 3 independent tangent
Curve 1 tangent:V851 V85T = 85 95 = 10 km/h good design
Tangent curve 2:V85T V852 = 95 60 = 35 km/h poor design

For a better illustration, the design procedure according to Safety Criterion II, is
presented for curved and tangent sites in the flow chart of fig. 6 [7].

2.2 Relation design

Relation design is another important aspect of modern highway geometric


design. Relation design means that sequences of design elements are formed,
such that the successive design elements are subject to specific relations or
relation ranges. This concept is the opposite of the practice in which single
design elements are put together more or less arbitrarily.
In order to achieve operating speed consistency between two circular curves
in the same or in the opposite direction the radii of these curves should have a
well-balanced relationship (known as relation design). The same is true for a
sequence independent tangent to curve.
Relation design is intended to ensure that the 85th-percentile speed (V85)
between successive design elements does not change abruptly, and coincides
with the assumptions of Safety Criterion II in Table 7.
Therefore, the corresponding ranges of change in the 85th-percentile speeds
of Safety Criterion II were taken as the basis for calculating relation design
backgrounds, based upon the respective operating speed background of the
country under study according to fig. 3 or Table 4, or eqn. (5).
The development of relation design diagrams will be described in the
following, using Germany as an example (fig. 7).
According to Table 4, eqn. (4a) of the operating speed background for
Germany (ISE) is:
6
10
V85 =
8, 270 + 8.01 CCR S

Step 1. Set, for example, R = 1000 m.


Three quantitative safety criteria for highway geometric design 31

CCRSi

CCRSi = 0
yes no

tangent curved site

2 2 2 2 2
V851 V85 2 2 (V85Tmax ) (V851 ) (V852 )
TL max =
TL min = 2
22.03
2 3.6 0.85 fig .5

no
TL TL > TLmin NIT Case 1

yes

TL < TLmax
yes no

IT Case 3
IT Case 2

2 V85i = V85Tmax V85i


V 85 i = V 85T = 11.016 (TL TL min ) + (V 851 )

fig .5 according
to SCI
with V851 > V852
|V85i V85i+1|

|V85i V85i+1| 10 km/h


no

yes |V85i V85i+1| 20 km/h


yes no

good design fair design poor design


SC II SC II SC II
Figure 6: Design procedure according to safety criterion II
32 How to make two-lane rural roads safer

Step 2. Calculate CCRS with respect to R from eqn. (1c) in fig. 1 without
regarding transition curves:
63, 700
CCR S =
R
63, 700
CCR S = = 63.7 gon/km
1000

Step 3. Determine V85 from eqn. (4a) or graphically from fig. 3 (Curve:
Germany, ISE):
V85 = 114 km/h.
Step 4. Subtract 10 km/h from V85 in Step 3 to reflect good design or
15 km/h1 to reflect fair design:
V85 = 104 km/h for good design
V85 = 99 km/h for fair design.
N. B. Regarding Tables 1 and 2, the expected accident rates for the
fair range are at least 2 to 3 times higher, than those of the good
range. Therefore, it may be beneficial to reduce the speed range
of the fair design level from 20 km/h to 15 km/h with respect to
the development of relation design backgrounds in order to limit
the accident risk and accident severity, thus making relation design
between successive design elements much safer.

Step 5. For V85 in Step 4, determine CCRS from eqn. (4a) and R from
eqn. (1c). For good design:
CCRS = 169 gon/km
or
R = 377 m,
and for fair design:
CCRS = 230 gon/km
or
R = 277 m.

Step 6. The intersections of the lines drawn horizontally or vertically from


radii of curve of 1000 m and 377 m, and from 1000 m and 277 m,
respectively, indicate the points that should fall on the relation
design curves for good and fair design (see fig. 7).

Step 7. Repeat Steps 1 through 6 for radii of curve of less than and greater
than 1000 m with increments of 100 m.

Fig. 7 shows the relation design background for Germany, as an example.


Relation design backgrounds were also developed in [1] for Australia, Canada,
France, Greece, Lebanon and the United States.
Based on the relation design curves, the designer in the country under study
could immediately decide whether certain radii of succeeding curves fall into the
Three quantitative safety criteria for highway geometric design 33

ranges of good, fair, or poor design. For instance, from the relation design
background in fig. 7 a radius of curve of 500 m, combined with the following
radii of curve, leads to:
R = 150 m poor design,
R = 220 m fair design,
R = 300 m good design,
R = 1500 m also good design.

The radii of curve sequences should, for new designs, always fall into the
good range. In cases of redesigns and of RRR-strategies for existing roads
the good relation design range may lead to conflicts with goals related to the
protection of the landscape or related to local demands. Therefore, in
substantiated individual cases, the fair relation design range is acceptable,
especially if adverse conditions can be avoided. In cases where relation design
ranges are at the poor level, a redesign should be considered for existing
roadway sections. If not undertaken, an unfavorable alignment with respect to
the actual driving behavior can be expected from safety and economic points of
view, because of the high accident risk and the resulting high accident costs, see
Tables 1 and 2.

Figure 7: Relation design background for Germany [1]

The following considerations and recommendations for the application of


Relation Design Backgrounds according to fig. 7 are important:

(1) When selecting radii of curve sequences, the minimum radius of curve must
at least correspond to the selected design speed of the roadway.

(2) In hilly topography, strong curvilinear alignments, based on the developed


relation design backgrounds, may be very favorable and influence traffic
safety positively.
34 How to make two-lane rural roads safer

(3) By way of contrast, in flat or mountainous topography curvilinear


alignments often do not satisfy the demands for an environmentally friendly
and economic design. However, relation design, as understood here, means
more. Relation design should not only be directed towards achieving good
curvilinear alignments (the main idea so far), but should also include
selecting sound transitions between independent tangents and curves. For
example, for flat topography this could mean the achievement of well-
balanced transitions between independent tangents and curves, whether or
not interspersed with curvilinear alignment sections. For mountainous
topography limiting design parameters may take precedence over relation
design issues, to the extent where a curve design with hair-pin bends may
be required.

(4) Also in the case of redesigns or RRR-projects curvilinear alignments may


not lead to favorable solutions, as many case studies conducted by the
authors revealed.

(5) Safety research [1, 15, 16] has revealed that relation design in the fair range
for radii of curve of R 200 m has to be avoided and for 200 < R 350 m
should be used only in exceptional cases, because of the expected high
accident risk.

(6) For the case of a transition: independent (long) tangent clothoid


circular curve the good design range should always apply. Based upon the
calculated results for establishing the relation design backgrounds of the
cited countries, the radii of curves following independent tangents should at
least be

R 300 m (calculated) for Australia,


R 300 m (calculated) for Canada,
R 500 m (calculated) for Germany,
R 400 m (calculated) for Greece,
R 300 m (calculated) for Lebanon, and
R 300 m (calculated) for the USA.

Similar conclusions were established in several other studies for Germany


and the United States, which revealed that radii of curve of 350 to 500 m provide
a certain cross-point in safety on circular curves and in the corresponding
transition sections.
Therefore, when transition curves are incorporated in the horizontal
alignment design, it is recommended that independent tangents should be
followed by curves with radii of at least 400 meters. When transition curves are
not incorporated between curves and independent tangents, it is suggested, that
curves should have a radius of at least 800 meters.
Three quantitative safety criteria for highway geometric design 35

3 Classification of safety criterion III


The safety investigations in most countries mostly attempt to address questions
related to the improvement of geometric road design. However, questions related
to the improvement of skid-resistance or tangential and side friction are not as
often considered, even though many publications indicate that adequacy of
friction represents an important safety aspect. Studies show that the likelihood of
a highway curve becoming an accident black spot increases with decreasing
skid-resistance. Consequently, modern highway geometric design should clearly
emphasize the need for sufficient friction between tire and road surface,
especially on curved roadway sections. Thus, following the development in the
previous chapters of Safety Criteria I and II, an important question remaining is
the evaluation of the driving dynamic aspects, especially when driving through
curves.
For this reason, Safety Criterion III was introduced to achieve driving
dynamic consistency at curved sites. It compares side friction assumed (fRA) for
curve design with side friction demanded (fRD) for cars riding through the curve
at the 85th-percentile speed level. Based on skid-resistance databases from
Germany, Greece, and the USA as well as the data assumed for tangential
friction factors from the guidelines of five countries, the overall regression
equation (eqn. (12a) in Table 9) between tangential friction factor (fT) and design
speed (Vd) was developed, [1].
By new research with respect to stopping sight distances, conducted by
Harwood et al. [17], the above relationship for tangential friction, this time
concerning nine countries can be expressed by eqn. (12b). Since the differences
between equations (12a) and (12b) are only marginal, both equations can be used
by applying this book.
The formula for side friction assumed corresponds to eqn. (13), where n
expresses the permissible utilization ratio for side friction assumed in
comparison to tangential friction, and the factor 0.925 represents tire-specific
influences. As can be seen, different utilization ratios are suggested for new
designs, separated according to hilly/mountainous and flat topography, as well as
for existing (old) alignments in Table 9, based on in-depth safety and economic
considerations.
Whereas side friction assumed is related to the design speed (Vd), (see
eqns. (12) and (13)), the side friction demanded is related to the 85th-percentile
speed (eqn. (14)) with respect to the radius of curvature and its associated
superelevation rate.
The quantitative ranges for the differences between side friction assumed and
side friction demanded are listed in Table 10 [1, 7]. They are based on the
previously mentioned skid-resistance and friction databases, again arranged for
good, fair (tolerable) and poor design practices.
The value of 0.04 for poor design, being the difference between side friction
assumed and demanded, suggests for practical design tasks that at such a curved
site already 4 per cent of superelevation would be missing for a safe ride through
the curve at the 85th-percentile speed level.
36 How to make two-lane rural roads safer

In several countries it is believed, based on practical experience and


beginning research, that the assumptions so far for tangential and side friction
values are too conservative, and higher values could be justified. If these
considerations prove to be reliable, the authors propose to increase the utilization
ratio of side friction to n = 0.70 for existing (old) alignments. According to fig. 8,
with a utilization ratio of 70 per cent for side friction one can still expect that an
available tangential friction portion of 71 per cent remains. Note that these
statements about possible future developments, if available, should not be
applied for new designs. Since a high friction supply guarantees a low accident
risk and thus an important safety issue.

Table 9: Listing of formulas with respect to safety criterion III, [1].

fT = tangential friction factor in modern highway geometric design [-]

= 0.59 4.85 103 Vd + 1.51 105 Vd2 (12a)

fT = 0.58 4.92 103 Vd + 1.81 105 Vd2 (12b)

fRA = side friction assumed [-]

= n 0.925 fT (13)

n = utilization ratio of side friction [%/100]

= 0.40 for hilly/mountainous topography; new designs

= 0.45 for flat topography; new designs

= 0.60 for existing old alignments

fRD = side friction demanded [-]

2
V 85
= e (14)
127 R

R = radius in the observed circular curve [m]

e = superelevation rate [%/100]


Three quantitative safety criteria for highway geometric design 37

Table 10: Classification of safety criterion III

Classification
Design Symbol* Differences of side friction Design (CCRS) class
values [gon/km]
[-]
Good + fRA fRD +0.01 CCRSi 180
Fair o 0.04 fRA fRD < + 0.01 180 < CCRSi 360
Poor fRA fRD < 0.04 CCRSi > 360

* The Symbols in Tables 6, 7 and 10 are needed for the comparison in Chapter 3.

Figure 8: Relationship between utilized side friction factor and remaining


available tangential friction [1]

The design procedure according to Safety Criterion III is presented in fig. 9


[7].
With the design speed, the 85th-percentile speed as well as side friction
assumed (fRA) and demanded (fRD), known, the safety-evaluation process
according to the ranges for the three safety criteria can be conducted, as
summarized in Table 11.
38 How to make two-lane rural roads safer

CCRS Vd

CCRSi = 0 fT = 0.59 4.85 10-3 Vd + 1.51 10-5 Vd2


or
fT = 0.58 4.92 10-3 Vd + 1.81 10-5 Vd2
yes no

tangent
curved site

V85i e

2 fRA = n 0.925 fT
V85
f = i e
RD 127 R

n = 0.40 for hilly / mountainous


topography, new designs
n = 0.45 for flat topography, new
designs
fRA - fRD +0.01 n = 0.60 for existing (old)
alignments
yes no

fRA fRD 0.04

yes no

+ o
SC III non- good fair design poor design
relevant design SC SC III SC III
III

Figure 9: Design procedure according to Safety Criterion III.


Three quantitative safety criteria for highway geometric design 39

4 Safety criteria vs. alignment design


The flow chart for the component alignment is presented in fig. 10 revealing
the connections and interconnections with the newly developed Safety Criteria I
to III, and other important safety aspects.
As can be seen, four of the five alignment design levels are either
controlled by the three quantitative safety criteria (Table 11) or by additional
direct or indirect safety-related aspects.
These include:

The selection of an appropriate design speed for new and existing (old)
alignments.
The analysis of independent or non-independent tangents.
The establishment of an operating speed and relation design background for
the country under study.
The introduction of sound driving dynamic assumptions.

Thus, a quantitative safety evaluation process has been developed that regards
three safety criteria for examining alignment design as good, as fair (minor
safety deficiencies) and as poor (major safety deficiencies). The process is
completed by providing sound permissible tangential and side friction factors as
well as other important safety-related issues.

Table 11: Summary of the quantitative ranges for Safety Criteria I to III for
good, fair and poor design classes.

Design classes
Safety Good Fair Poor
Criterion Permissible Tolerated Non-permissible
differences differences differences
10 km/h <
I V85i Vd V85i Vd V85i Vd
10 km/h 20 km/h > 20 km/h

10 km/h <
II V85i V85i+1 V85i V85i+1 V85i V85i+1
10 km/h 20 km/h > 20 km/h

0.04
III fRA fRD +0.01 fRA fRD fRA fRD
< +0.01 < 0.04

The safety process influences quantitatively the design elements of the four
subcomponents:
40 How to make two-lane rural roads safer

- horizontal alignment,
- vertical alignment,
- cross section / alignment, in respect of superelevation rate and superelevation
runoff, and
- sight distance (fig. 10).

Only the subcomponent three-dimensional alignment has so far not been


incorporated into the overall safety-evaluation process. It represents still the
weakest link in highway geometric design.

5 Safety module
Finally, for a simplified general overview of the safety-evaluation procedure, for
example, for network investigations, the previously discussed three safety
criteria will be combined in the following in an overall safety module.
Based on the classification ranges of Table 11, Safety Criteria I to III indicate
that roadway sections could exhibit different design safety levels with respect to
the individual safety criteria. The reason for this is that each of the safety criteria
represents a separate safety aspect in highway geometric design. It may happen,
for example, that the transition section between an independent tangent and a
curve could correspond to poor design according to Safety Criterion II, whereas
Safety Criterion I with respect to the design speed or Safety Criterion III with
respect to side friction assumed, or both, provide acceptable values for the
observed curved site.
Therefore, for a fast, comprehensive overview of new or existing (old)
alignment or road networks, Safety Criteria I to III will be combined into a safety
module.
In this connection, Eberhard [10] developed a simple calculation scheme in
Table 12 for evaluating the safety module according to good, fair, and poor
design levels. Based on comparative analyses of the actual accident situation, it
can be proved in the Chapter 3, that the three individual safety criteria can be
equally weighted. However, as can be seen from Table 12, a specific weighting
factor is assigned to each design level. Good design is classified by the
weighting factor of +1.0, fair (tolerable) design is described by the factor 0.0
and for poor design the factor 1.0 becomes relevant. Summing up the
weighting factors for the individual safety criteria, as is shown, based on
different calculation examples in Table 12, the calculated average value
represents, in combination with the given limiting ranges, an evaluation scheme
for the safety module.
The results of the application of the safety module are illustrated in the Case
Studies in Chapter 6.
According to the calculation examples in Table 12 every design element of a
roadway section can now be classified by the overall safety module. For the
evaluation of a curved site all three safety criteria become relevant, whereas for
the classification of independent tangents (IT) the evaluation of Safety
Criterion III is trivial, since no centrifugal forces exist on tangents. As explained
Three quantitative safety criteria for highway geometric design 41

New design: New design

Component: Component:
Road network Cross section
- Road function - Standard cross section
- Category group - Traffic volume
- Road category - Traffic quality

New and
COMPONENT: ALIGNMENT
Existing
Alignments
Table 3 or
Selection of
eqns. (7)
Design Speed (Vd) and (8)

Analysis of DESIGN LEVELS Relation Design


Independent Background,
Tangents Horizontal Alignment
fig. 7
eqns. (9)-(11) Determination of
- Tangents (T),
- Circular Curves (R),
- Transition Curves (A). Driving Dynamic
Assumptions for
Side Friction
Operating Speed Factors,
Background (V85) Table 9
- Curvature
Vertical Alignment
Change Rate of
the Single Curve Determination of Controls:
- (Lane Width), - Grades (G), Safety Criteria I to
see fig. 3, - Crest Vertical Curves (RVC), III,
Table 4, - Sag Vertical Curves (RVS). Table 11
eqns. (5) and (6)

Cross Section / Alignment


Control:
Determination of Safety
- Superelevation Rate in Tangents Criterion III,
(emin), Table 10
Three- - Superelevation Rate in Curves (e),
Dimensional - Negative Superelevation Rate (e),
- Superelevation Runoff (s),
Alignment - Pavement Widening.
Determination of Driving
- Elements of Dynamic
Three - Sight Distance
Assumptions
Dimensional Determination of
for
Alignment, - Stopping Sight Distance (SSD), Tangential
- Design of the - Passing Sight Distance (PSD), Friction
Driving Space.
- Sight Distance Analysis.

Figure 10: Flow chart for alignment design vs. safety [1]
42 How to make two-lane rural roads safer

in Section 2.1 a non-independent tangent will not be regarded in the design


process and, thus, does not have any influence concerning Table 12.
If the safety module reveals good design, no adaptations or corrections of the
existing alignment are necessary.
If the safety module expresses fair design, the installation of sound traffic
control devices for good visual guidance, whether or not combined with
appropriate speed regulations, is normally sufficient to alleviate existing safety
deficiencies. Sometimes, a cost-effective safety improvement can be achieved by
reconstructing the pavement for increasing skid resistance and superelevation
rate. However, in the fair design range, higher accident rates and cost rates are to
be expected than in the good design range (see Tables 1 and 2), and [18].
If the safety module indicates poor design, upgrading of the existing
alignments or full-blown RRR-projects are normally recommended. According
to Table 12, this most critical case includes, with respect to the individual safety
criteria:

Poor design according to two criteria and fair design according to one, or
poor design according to all three criteria.
Three quantitative safety criteria for highway geometric design 43

Table 12: Evaluation scheme for the safety module, based on the three safety
criteria [10]

Design Level Symbol Weighting Factor


SC I, SC II, SC III

Good + + 1.0
Fair o 0.0
Poor 1.0

Limiting Ranges for the Average


Value of the Safety Module

x 0.5 good (+)


- 0.5 < x < 0.5 fair (o)
x 0.5 poor ()

Calculation Examples

1 2 3 4 5 6
Sum Safety Module
Design Average
SC I SC II SC III Class
Element Value

Curve + 1.0 1.0 1.0 1.0 0.33 0

Curve 0.0 1.0 1.0 2.0 0.67

Tangent 0.0 + 1.0 IT (n.r.)* + 1.0 + 0.50 +

Curve + 1.0 + 1.0 + 1.0 + 3.0 + 1.00 +

Legend:
* The expression describes an independent tangent for which Safety
Criterion III is not relevant, since no centrifugal force exists.
Therefore, Safety Criterion III will not be considered in the
evaluation scheme for independent tangents.
44 How to make two-lane rural roads safer
Chapter 3

Comparative analyses of the actual accident


situation with the results of the safety criteria

In numerous master theses and scientific publications [7, 10, 18 23], it was
proved that the safety concept of Lamm et al. [1] is appropriate for the safety
classification of roadway sections according to good, fair (tolerable), and poor
design practices and sensible results can be expected. The goal of the above
investigations was to show the level of agreement between the outcome of the
three safety criteria and the actual accident situation, expressed by the accident
rate (accident risk) and the accident cost rate (accident severity). While Schmidt
[19], Eberhard [10] and Zumkeller [20] already demonstrated a strong tendency
for a good agreement between safety criteria and accident rates, Schneider [21]
was the first to express a level of agreement in numbers.
In order to do so in the classification Table 6 for Safety Criterion (SC) I, in
Table 7 for SC II and in Table 10 for SC III the symbols + (good design), o
(fair design), and poor design) were already introduced. For comparative
reasons Schneider developed a similar system with respect to accident rates and
cost rates to differ between a low, medium and high endangerment. In this
connection,he defined a low accident rate, if no more than 1 accident occurred on
an element sequence (curved site or independent tangent), a medium accident
rate, if no more than 2 accidents occurred, and a high accident rate, if more than
2 accidents were present within an investigation time period of three years (for
the definition of the accident rate, see eqn. (2), Schneider considered in his
investigations the accident types Run-Off-The-Road and Deer accidents.
Both types are directly related to the operating speed, and thus best represent the
assumptions of the Safety Criteria.
With respect to the accident cost rate the process is far more complicated,
since here not only do the number of accidents count, but the accident costs
(either property damage or light, or severe, or fatal injury) also play an important
role (Chapter 1, Section 2.1). However, space does not permit examination of
Schneiders assumptions [21] in more detail, since the individual boundaries for
accident rates and cost rates change for every new accident and design database.
46 How to make two-lane rural roads safer

In order to regard Accident Rate (AR) and Accident Cost Rate (ACR),
equally, Schneider combined both rates, equally weighted, in a 3 by 3 matrix
(Table 13), which represents three endangerment levels (+, o, ), similar to the
safety criteria (see, Tables 6, 7 and 10).
Table 13: Combination of Accident Rates and Accident Cost Rates for
Different Endangerment Levels [21]
ACR

low medium high

+ o

low + + + o

AR medium o + o

high o

Thus, when a safety criterion reveals good design (symbol +) a level of full
agreement is reached, if the combination of AR and ACR in Table 13 also
shows the symbol +. However, if a safety criterion falls into the range of poor
design, a full agreement is reached if the combination of AR and ACR shows the
symbol .
Partial agreement arises when a safety criterion reveals +, but the
combination according to Table 13 results in an accident situation that would be
represented by the symbol o. A disagreement is defined, if the comparison
between individual Safety Criteria and Table 13 differ by two steps, i.e. from +
to or vice versa.
For the aggregation of agreement percentages, full agreement is regarded as
having a weight of 2, partial agreement a weight of 1 and disagreement a
weight of 0. The aggregated agreement percentages are calculated by dividing
the actual summation of weighted agreement levels by the maximum agreement
possibilities for the observed individual accident- and design database. The
above procedure has to be conducted for each safety criterion individually. Thus,
the following results are always valid for one safety criterion.
The reader who is interested in a more detailed discussion about the levels of
agreement between safety criteria and accident rates should consult the master
theses of Schneider [21] and Ruscher [7].

1 Data base: Schneider [21]


The investigations were based on two different relatively small databases of
different origin, which overall encompassed 50 roadway sections with a length
of 136 kilometers. For database I, which consisted of 20 roadway sections
Comparative analyses of the actual accident situation with the results of the Safety Criteria 47

representing the design- and accident data of the master theses of Schmidt [19],
Beck [9] and Zumkeller [20], a level of agreement of 72% could be reached for
Safety Criterion (SC) I, 69% for SC II and 66% for SC III in comparison with
the actual accident situation, expressed by the combination of accident rates and
cost rates according to Table 13.
The database II of Schneider, which encompassed 30 roadway sections,
provided by the Ministry of Environment and Transportation, State Baden-
Wuerttemberg, Germany, revealed for SC I 76%, for SC II 72%, and for
SC III 69% level of agreement.
Both databases combined resulted in levels of agreement of

- SC I: 74%
- SC II: 71%
- SC III: 68%.

These results can be considered significant, and the developed safety criteria
can, in the future, strongly support the work of traffic-safety officials, when
making decisions about good, fair or poor design practices.

2 Data Base: Ruscher [7]


Although the previously conducted master theses and the additional research
work of the Institute have shown a good agreement between design classes,
safety criteria and accident rates, it may be possible that these former
investigations could be statistically disputed, because of too small and perhaps
not consistent design and accident databases. Therefore, during the master thesis
of Ruscher [7] the developed Safety Criteria were re-examined once more in
relation to the actual accident situation, with this work being based on new,
independent and statistically sound databases. His investigations focused on
236 roadway sections, consisting of 2726 individual element sequences (curved
sites or independent tangents) with an overall length of 490 kilometers. For the
evaluation of the ISE-Safety Concept with respect to a broad database, two cases
with different accident types were conducted.
In the first case, 1000 accidents of the types Run-Off-The-Road and Deer
were included. In the second case, 1384 accidents of the types ROR, Head-
on/Rear end and Deer were incorporated.
With respect to the first case for Safety Criterion I a level of agreement of
81%, for SC II of 77% and for SC III of 72% in comparison with the actual
accident situation, (expressed by the combination of accident rates and cost rates
according to Table 13) was reached.
Despite the consideration of the additional accident type Head-on/Rear end
in the second case the results revealed insignificantly lower levels of agreement.
SC I showed an agreement level of 79%, SC II of 75% and SC III of 71%. Thus,
the relatively small databases of Schneider [14] and now supported by the broad
database of Ruscher [7] clearly indicate a statistically significant relationship
48 How to make two-lane rural roads safer

between the results of the three individual safety criteria and the actual accident
rates to identify good (low endangerment), fair (medium endangerment) and
poor (high endangerment) design practices. This is true for new designs,
redesigns, and RRR-practices, as well as for the examination of existing (old)
alignments.
Unfortunately, most highway agencies must rely on accident histories,
collected over several years, to identify roadway sections requiring safety
improvements. The great advantage of the ISE-Safety Concept is that the analyst
can, in the design stages, predict the endangerment (low, medium, high) for new
alignments. Additionally, the three safety criteria are also appropriate to render
judgements about the safety conditions of existing (old) roadway sections or
whole road-networks. In consequence, the highway and traffic-safety engineer
can quantitatively evaluate the expected accident situation. Deficiencies in new
designs can be corrected prior to construction and sound countermeasures can be
planned for highly endangered existing or old alignments.
As an additional effect the research results of Schneider [21] and Ruscher [7]
prove that the assumption in Chapter 2, Section 5 with respect to weighting all
three safety criteria equally for the assessment of the safety module, has to be
regarded as correct.
Chapter 4

Case studies

In the following chapter, four Case Studies will be discussed, one from South
Africa, one from Germany, one from Greece, and one from Italy, in order to
reveal how simple and sound the practical application can be conducted.

1 Example I

The horizontal alignment in Columns 1 to 5 of Table 14 shows an existing two-


lane rural State Route in South Africa [24]. Accident analysis indicates a high
accident frequency and severity at Curve 3. The main accident cause, recorded
by the police, was improper speed estimation in the curve itself and in the
transition sections. This suggests that the main accident type was ROR. The
longitudinal gradients are less than 2 per cent and the AADT values were
6800 vehicles/day in 1998. It was interesting to note that in the curved and
tangent sections preceding and following the critical Curve 3, no relevant
accidents were recorded (for example because of overtaking vehicles).

1.1 Results of the safety criteria

Before any remedial measures to increase traffic safety were to be conducted, the
safety evaluation processes of Criterion I (Table 6), Criterion II (Table 7) and
Criterion III (Table 10) were performed to analyze the endangerment of the
observed roadway section.
All necessary information about the design elements 1 to 5 are listed in
Table 14. The relevant descriptions include: radius of curve and parameters of
clothoids (Col. 2), length of the design element (Col. 3), Curvature Change Rate
of the Single Curve (Col. 4), and superelevation rate (Col. 5). To perform the
safety evaluation processes of Criterion I (Cols. 69), Criterion II (Cols. 1017),
and Criterion III (Cols. 1821), the following basic calculations were made.
They are discussed in detail under the following issues 1 to 9:
50 How to make two-lane rural roads safer

1. Curvature change rate of the single curve (Table 14, Col. 4). According to
Chapter 1, Section 1, the basic formulae for CCRS correspond to eqn. (1), and are
explained in fig. 1. It follows then for:

Element Sequence 1 (circular curve with following clothoid)

L L
cr1 + cl2 63, 700
R 2R
CCR S1 =
1 1
[gon/km] (1)
L

R1 = 250 m, Lcr1 = 78.4 m


L = Lcr1 + Lcl2 = 272 m
A2 = 220 m, Lcl2 = 193.6 m

Control:

2 2
A2 (220)
LCl2 = = = 193.6 m
R 250

78.4 193.6
250 + 2 250 63, 700
CCR S1 =
= 164 gon/km
272

Element 2 (tangent)

R= CCRS2 = 0 gon/km

Element 3 (circular curve, one radius)

L cr3
63, 700
R3 63, 700
CCR = =
S3 L R3

L = Lcr3; R3 = 50 m CCRS3 = 1274 gon/km

Element 4 (tangent)

R= CCRS4 = 0 gon/km
Table 14: Numerical data for the safety evaluation process (Example I)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Database Application of the three safety criteria
No. Parameter CCRSi e Safety Criterion I Safety Criterion II Safety Criterion III
A, R, T Li V85i Vd |V85i-Vd| SC I TLmin TLmax NIT Case V85r V85i |V85i-85i+1| SC II fRA fRD fRAfRD SC III
[] [m] [m] [gon/km] [%] [km/h] [km/h [km/h] [m] [m] [km/h [km/h] [km/h] [-] [-] [-]
] ]
1 R=250 272 164 8.0 94 85 9 + 94 0.16 0.2 0.04 o
ESC - n.r.
2 T,R = 65 0 2.5 - - - n.r. 301 NIT 1 - - n.r. n.r. n.r. n.r.
ESC 47 -
3 R=50 80 1274 8.0 47 85 38 47 0.16 0.27 0.11
ESC 49 -
4 T,R= 340 0 2.5 96 85 11 o 290 508 IT 3 96 96 n.r. n.r. n.r. n.r.
ESC 3 +
5 R=250 260 185 8.0 93 85 8 + 93 0.16 0.19 0.03 o

Legend:

IT = Independent Tangent good design [+]


NIT = Non-Independent Tangent fair design [o]
SC = Safety Criterion poor design []

Case studies 51
ESC = Element (Sequence)Change
n.r. = not relevant, because either SC III does not exist, or the tangent is too short (NIT)
52 How to make two-lane rural roads safer

Element Sequence 5 (compound circular curve, two radii)

Lcr4 Lcr5
+ 63, 700
R R
CCR S5 =
4 5
L
R4 = 250 m, Lcr4 = 100 m
L = Lcr4 + Lcr5 = 260 m
R5 = 450 m, Lcr5 = 160 m

100 160
+ 63, 700
CCR S5 =
250 450
= 185 gon/km
260

The calculated CCRSi-values can be found in Col. 4 of Table 14.

2. 85th-percentile speed (Table 14, Col.6). Since for South Africa an individual
operating speed background, as in fig. 3 or Table 4, is not known so far, eqn. (5)
for worldwide application is selected for determining the operating speed (V85).
Thus, the formula for V85 is given as follows:

-5 2
V85 = 105.31 + 2 10 CCR Si 0.071 CCR [km/h] (5)
i Si

Element Sequence 1
CCRS1 = 164 gon/km
V851 = 105.31 + 2 105 (164)2 0.071 (164) = 94 km/h.

Element 2
CCRS2 = 0 gon/km

The determination of operating speeds in tangents (V85T) will be discussed in


detail under Issue 4 in the following.

Element 3
CCRS3 = 1274 gon/km
V853 = 105.31 + 2 105 (1274)2 0.071 (1274) = 47 km/h.

Element 4
CCRS4 = 0 gon/km

The determination of operating speeds in tangents (V85T) will be discussed in


detail under Issue 4 in the following. Accordingly, it follows that
V854 = V85T = 96 km/h.
Case studies 53

Element Sequence 5
CCRS5 = 185 gon/km
V855 = 105.31 + 2 105 (185)2 0.071 (185) = 93 km/h.

The calculated V85i-values are listed in Table 14, Col. 6.

3. Assessment of an appropriate design speed (Table 14, Col. 7). The design
speed of the existing (old) alignment is unknown. Thus, in order to derive an
estimate of an appropriate design speed, the procedure described in Chapter 1,
Section 2.3.4, was followed:

i =n
CCR Si L i
CCR S = i =1
i =n
[gon/km]. (7)
Li
i =1
For the existing alignment, it follows according to Issues 1 and 2 or Table 14:
164 272 + 1274 80 + 185 260
CCR = = 318 gon/km
S 272 + 80 + 260
(without regarding tangent lengths).
From this CCRS it follows from eqn. (5), that the average 85th-percentile
speed is:

V85 = 105.31 + 2 105 (318)2 0.071 (318) = 85 km/h.

Based on this average 85th-percentile speed, a design speed of

Vd = 85 km/h

(see Table 14, Col. 7) was selected for the whole roadway section, and was
regarded as a reliable speed estimation for the existing alignment.

4. Evaluation of tangents (Table 14, Cols. 10-13). Elements 2 and 4 of the


existing alignment are tangents. The evaluation of tangents in the design process
is discussed in detail in Chapter 2, Section 2.1. Since Element 2 with a length of
65 meters is relatively short and Element 4 with 340 meters (Col. 3) is relatively
long, it was found that Element 2 corresponds to a non-independent tangent
(NIT), and Element 4 is an independent tangent (IT), as shown in the following
calculations.

Tangent Element 2
TL = 65 m

To decide whether or not the present tangent is independent or non-independent


the three cases discussed in Chapter 2, Section 2.1 have not be considered. The
54 How to make two-lane rural roads safer

calculation of TLmin and TLmax is necessary for the differentiation between


the cases (see fig. 5),

2 2
(V851 ) (V853 )
TL = (9a)
min 22.03

V851 corresponds to the operating speed in Element Sequence 1 and V853


corresponds to the operating speed in Element 3, see fig. 11 and Table 14.

2 2
(94) (47)
TL min = = 301 m.
22.03

Since
TL < TLmin
65 m < 301 m,

Case 1 is relevant (see Chapter 2, Section 2.1), and the tangent is non-
independent. With respect to Safety Criterion II, the sequence Curve 1 to
Curve 3 becomes relevant.

Tangent Element 4 (fig. 5)


TL = 340 m

2 2 2 2
(V85 ) (V85 ) (47) (93)
3 5
TL min = =
22.03 22.03
TLmin = 290 m.

2 2 2
2 (V85Tmax ) (V85 ) (V855 )
TL max = 3 (10a).
22.03

V85Tmax corresponds to eqn. (5) for CCRS = 0 gon/km

V85Tmax 105 km/h

2 2 2
2 105 47 93
TL max = 508 m .
22.03

Since
TLmin < TL < TLmax
Case studies 55

290 m < 340 m < 508 m,


Case 3 becomes relevant, and the present tangent is characterized as
independent.
For the calculation of the operating speed (V85T) eqn. (11a) for Case 3 in
Chapter 2, Section 2.1 is valid:

2
V85T = 11.016 (TL TL min ) + (V855 ) (11a)
V855 > V853
(always use the larger value).

2
V85 = 11.016 (340 290) + (93)
T

V85T = V854 = 96 km/h.

Finally, the operating speed V85T of Element 4 is listed in Col. 14 of


Table 14, as well as in Cols. 6 and 15.

5. Side friction assumed / demanded (Table 14, Cols. 18 and 19). The
equations for side friction assumed/demanded are shown in Chapter 2, Section 3.
According to Table 9, the formula for side friction assumed on existing
alignments is:
fRA = 0.06 0.925 fT, (13)
where
fT = 0.59 4.85 10-3 Vd + 1.51 10-5 Vd2 (12a)

It follows:
fRA = 0.6 0.925 (0.59 4.85 103 Vd + 1.51 105 Vd2)
fRA = 0.33 2.69 103 Vd + 0.84 105 Vd2
For the selected design speed of Vd = 85 km/h, side friction assumed is:
fRA = 0.33 2.69 103 (85) + 0.84 105 (85)2
fRA = 0.16.
The value is listed for the curved sites 1, 3 and 5 in Col. 18 in Table 14.
The actual side friction demanded is calculated from the following equation
in Table 9:
2
V85
f RD = e. (14)
127 R
Thus, the side friction factors demanded can be calculated for:

Element 1
2
94
f RD = 0.08 = 0.20
1 127 250
56 How to make two-lane rural roads safer

Element 2
In tangents fRD is zero, because no centrifugal force exists.

Element 3
2
47
f RD = 0.08 = 0.27
3 127 50

Element 4
Explanation, see Element 2.

Element 5
2
93
f RD = 0.08 = 0.19
5 127 250

The values for side friction demanded are listed in Col. 19 of Table 14.

6. Results of Safety Criterion I (Table 14, Cols. 8 and 9). The classification of
SC I is explained in Chapter 2, Section 1 and in Table 11. Accordingly the
following solutions can be expected for:

Element 1
|V851 Vd| = |94 85| = 9 km/h (good design)

Element 2
This element is characterized as NIT and is not relevant for the developed safety
concept.

Element 3
|V853 Vd| = |47 85| = 38 km/h (poor design)

Element 4
|V854 Vd| = |96 85| = 11 km/h (fair design)

Element 5
|V855 Vd| = |93 85| = 8 km/h (good design)

The results are also presented as a graphical layout in fig. 11.

7. Results of Safety Criterion II (Table 14, Cols. 16 and 17). The


classification of SC II is explained in Chapter 2, Section 2 and in Table 11.
Accordingly, the following solutions can be expected.

Transition between Elements 1 and 3 (Element 2 as NIT drops out):


Case studies 57

|V851 V853| = |94 47| = 47 km/h (poor design).

Transition between Elements 3 and 4:


|V853 V854| = |47 96| = 49 km/h (poor design).

Transition between Elements 4 and 5:


|V854 V855| = |96 93| = 3 km/h (good design).

The results are also presented as a graphical layout in fig. 11.

8. Results of Safety Criterion III (Table 14, Cols. 20-21). The classification of
SC III is explained in Chapter 2, Section 3 and in Table 11. Accordingly, the
following solutions can be expected for:

Element 1
f RA f RD = 0.16 0.20 = 0.04 (fair design)
1

Element 3
f RA f RD = 0.16 0.27 = 0.11 (poor design)
3

Element 5
f RA f RD = 0.16 0.19 = 0.03 (fair design)
5
The results are also presented as a graphical layout in fig.11.

9. Final evaluation of the investigated two-lane rural roadway section. From


the listings and/or calculations of the input data in Table 14, the safety-
evaluation processes for Criterion I, Criterion II, and Criterion III were carried
out in conjunction with the quantitative ranges of Table 11, in order to
distinguish good design from fair and poor design practices. For tangents, Safety
Criterion III is not relevant, since no centrifugal force exists.
An analysis of the critical curve (Element 3 in Table 14) indicates that the
difference between V85 and Vd exceeds 20 km/h and corresponds to poor design
according to the ranges of Safety Criterion I, presented in Table 11. The same is
true for Safety Criterion II with respect to the V85-speed differences between
Elements 1 and 3 as well as between Elements 4 and 3. These transitions thus
also fall into the ranges of poor design, presented in Table 11. A poor design
level is also revealed for Element 3 (Table 11) with respect to Safety
Criterion III, regarding the difference between side friction assumed fRA and side
friction demanded fRD.
The numerical data of Table 14 is difficult to describe and these listings as
valuable as they are may be too complex for easy comprehension of an exact
evaluation overview.
58 How to make two-lane rural roads safer

A graphical presentation of the numerical results in Table 14 was thus


developed and is presented in fig. 11. The different design levels, based on the
individual Safety Criteria I to III, can be recognized visually by using different
symbols. The symbols for Safety Criterion II are arranged vertically to the road
axis, whereas the symbols for Safety Criterion I are located on the left side and
those for Safety Criterion III are located on the right side, parallel to the axis.
From the graphic layout of fig. 11, it can immediately be recognized that the
critical curve (Element 3) corresponds to poor design practices in terms of all
Safety Criteria. This result supports the previous statements about the serious
accident situation at this curve site and the corresponding transitions.

Figure 11: Graphical Presentation of the Three Safety Criteria for Good, Fair or
Poor Design Practices (Example I)
Case studies 59

In addition, it can be seen that the two curved sites (Elements 1 and 5) can be
evaluated only as fair designs with respect to Safety Criterion III, despite the
existence of a superelevation rate of 8 per cent (Table 14). For these cases, the
speed behavior should be lowered through the application of speed limits and/or
appropriate traffic control devices (for example: chevrons [20]).
However, with respect to Curve 3 the expected severe accident situation
suggests that redesign may be called for.

1.2 Results of the safety module

The theoretical background of the safety module was discussed in Chapter 2,


Section 5. Furthermore, an evaluation scheme for practical applications was
developed in Table 12, based on the three safety criteria.
Whereas each criterion describes only one safety aspect (alignment, operating
speed or driving dynamics), the safety module connects them all and thus,
represents, to a certain extent, an overall averaged evaluation. For example, the
safety module could be of great advantage for road-network evaluations [25, 26].
For the presentation of the Safety Module the results of the individual Safety
Criteria in Table 14, characterized by the symbols +, 0, and , are listed in
Table 15 (Cols. 6 to 8). Note that the outcome for the Safety Module is always
related to the observed driving direction.
A graphical layout of the investigated road axis including the individual
elements is presented in fig. 12. The endpoints of the elements (respectively,
element sequences) are determined by the capital letters A to E to clarify the
situation.
The evaluation of the Safety Module for the driving direction A to E
according to fig. 12 is shown by the table in the upper part of Table 15, whereas
the table in the lower part represents the driving direction E to A.
Exemplarily, the calculation process for the Safety Module for the driving
direction A to E is conducted. The results can be found in Column 9 of Table 15
and with respect to the axis of fig. 12.

Section A to B consists of Elements 1 and 2 (fig. 12)


Table 15 reveals:

SC I = + Weighting Factor = +1.0 (Table 12)


SC II = Weighting Factor = 1.0 (Table 12)
SC III = 0 Weighting Factor = 0.0 (Table 12)
SUM 0.0 : 3 = 0.0

This result suggests according to Table 12 fair design, as presented in


Table 15 and fig. 12.

Section B to C consists of Element 3 (fig. 12)


Table 15 shows:
60 How to make two-lane rural roads safer

SC I = Weighting Factor = 1.0


SC II = Weighting Factor = 1.0
SC III = Weighting Factor = 1.0
SUM 3.0 : 3 = 1.0

The result implies according to Table 12 poor design, as presented in


Table 15 and fig. 11.

Section C to D consists of Element 4 (fig. 12)


Table 15 reveals:

SC I = 0 Weighting Factor = 0.0


SC II = + Weighting Factor = +1.0
SUM 1.0 : 2 = + 0.5

Safety Criterion III is not relevant.


According to Table 12 good design can be determined, as presented in
Table 15 and fig. 12.

Section D to E consists of Element 5 (fig. 12)


The calculation of the safety module is not possible, since the element that
follows point E, is unknown and therefore, Safety Criterion II cannot be
calculated.
Comparing the two parts of Table 15 and fig. 12, one can recognize clearly
that, obviously curve element 3 and the transitions are highly endangered for
both driving directions. This means the responsible agency has to conduct in this
area a safety project, probably a redesign or the establishment of stationary radar
devices.
According to the opinion of the authors the additional presentation of the
graphical layout for the opposite driving direction E to A would go too far and
will not increase the comprehension of the procedure.

2 Example II
The old existing horizontal alignment in Columns 1 to 5 in Table 16 shows a
two-lane rural State Route in southeast Germany. Accident blackspots exist
especially in curves 2 and 4 and the corresponding transitions. The accidents
reported by the police were mainly ROR accidents. The longitudinal grades are
less than 6 per cent; however, downgrades up to 5 per cent exist partially in the
direction of increasing stationing. The AADT values were 8940 vehicles/day in
1995. The example is related to Ref. [27].

2.1 Results of the Safety Criteria

To support the request for possibly redesign and reconstruction measures, the
safety evaluation processes of Criterion I (Table 6), Criterion II (Table 7) and
Criterion III (Table 10) were performed to analyze the critical issues with respect
to the roadway section, to be investigated.
Case studies 61

Table 15: Safety module for both driving directions (Example I)

DRIVING DIRECTION A to E
1 2 3 4 5 6 7 8 9 10
Safety
No. Station Parameter Safety Criteria Designations
module
according to
from to A, R, T Li SC I SC II SC III
fig. 12
[-] [km] [km] [m] [m]
R1=250,
1 0.378 0.650 272 + o
A2=220
ESC n.r.
2 0.650 0.715 T, R= 65 n.r. n.r. o A to B
ESC
3 0.715 0.795 R3=50 80
B to C
ESC
4 0.795 1.135 T, R= 340 o n.r.
+ C to D
ESC +
R4=250,
5 1.135 1.395 260 + o D to E
R5=450

DRIVING DIRECTION E to A
1 2 3 4 5 6 7 8 9 10
Safety
No. Station Parameter Safety Criteria Designations
dule
according to
from to A, R, T Li SC I SC II SC III
fig. 12
[-] [km] [km] [m] [m]
R1=250,
1 0.378 0.650 272 + o B to A
A2=220
ESC n.r.
2 0.650 0.715 T, R= 65 n.r. n.r.
C to B
ESC
3 0.715 0.795 R3=50 80
ESC
D to C
4 0.795 1.135 T, R= 340 o n.r.
ESC +
R4=250, + E to D
5 1.135 1.395 260 + o
R5=450
Legend:

SC Safety Criterion
ESC Element (Sequence) Change
+ good design
o fair design
poor design
n.r. not relevant
Driving Direction
No statement possible, since the preceding or succeeding element is
unknown
62 How to make two-lane rural roads safer

Figure 12: Graphical presentation of the safety module for Example I


(driving direction A to E).

All necessary data about the alignment are listed in Table 16. In order to
perform the safety evaluation process the following calculations have to be
conducted.
Case studies 63

1. Curvature change rate of the single curve (Table 16, Col. 4)

Element 1 (tangent)

R= CCRS1 = 0 gon/km

Element sequence 2 (compound circular curve with clothoid in front and


clothoid behind):

A2 = 100 m, Lcl2 = 77 m

R3 = 130 m, Lcr3 = 24 m
L = Lcl2 + Lcr3 + Lcr4 + Lcl5 = 176 m
R4 = 70 m, Lcr4 = 46 m

A5 = 45 m, Lcl5 = 29 m

Lcl2 Lcr3 Lcr4 Lcl5


+ + +
2 R 3 R 3 R 4 2 R 4 200 3
CCR S2 = 10 [gon/km]
L

77 24 46 29
+ + +
CCR S2 =
2 130 130 70 2 70 200
10
3
176

= 488 gon/km

Element 3 (tangent):

R= CCRS3 = 0 gon/km

Element 4 (circular curve, one radius, two clothoids):

A7 = 45 m, Lcl7 = 44 m

R8 = 46 m, Lcr8 = 32 m L = Lcl7 + Lcr8 + Lcl9 = 98 m

A9 = 32 m, Lcl9 = 22 m

L cl7 L cr8 L cl9


+ +
2R 8 R 8 2R 8 200 3
CCR S4 = 10 [gon/km]
L
64 How to make two-lane rural roads safer
Table 16: Numerical data for the safety evaluation process (Example II)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Database Application of the three safety criteria
No. Parameter CCRSi e Safety Criterion I Safety Criterion II Safety Criterion III
A, R, T Li V85i Vd |V85i-Vd| SC I TLmin TLmax NIT Case V85r V85i |V85i-85i+1| SC II fRA fRD fRAfRD SC III
[] [m] [m] [gon/km] [%] [km/h] [km/h] [km/h] [m] [m] [km/h] [km/h] [km/h] [-] [-] [-]
1 T,R = 386 0 2.5 112 80 32 149 548 IT 3 112 112 n.r. n.r. n.r. n.r.
ESC 30
A=100,
R=130,
2 176 488 7 82 80 2 + 82 0.17 0.69 0.52
R=70,
A=45
ESC 7 +
3 T,R = 234 0 2.5 89 80 9 + 119 816 IT 3 89 89 n.r. n.r. n.r. n.r.
ESC 25
A=45,
4 R=46, 98 917 7 64 80 16 o 64 0.17 0.63 0.46
A=32

Legend:

IT = Independent Tangent good design [+]


NIT = Non-Independent Tangent fair design [o]
SC = Safety Criterion poor design []
ESC= Element (Sequence) Change
n.r. = not relevant, because either SC III does not exist, or the tangent is too short (NIT)
Case studies 65

44 32 22
+ +
246 46 246 200 3
CCR S4 = 10
98
= 917 gon/km

The calculated CCRSi-values can be found in Col. 4 of Table 16.

2. 85th-percentile speed (Table 16, Col. 6) For Germany an individual


operating speed background was established by ISE according to fig. 3 or
Table 4. Thus, the formula for the German operating speed background reads:

V85 = 106 / (8270 + 8.01 CCRSi) [km/h] (4a)

Element 1
CCRS1 = 0 gon/km

Operating speeds in tangents (V85T) are subject of Issue 4.


Element Sequence 2
CCRS2 = 488 gon/km
V852 = 106 / (8270 + 8.01 488) = 82 km/h.

Element 3
CCRS3 = 0 gon/km

Operating speeds in tangents (V85T) are subject of Issue 4.

Element 4
CCRS4 = 917 gon/km
V854 = 106 / (8270 + 8.01 917) = 64 km/h.

The calculated V85i-values are listed in Table 16, Col. 6.

3. Assessment of an appropriate design speed (Table 16, Col. 7). The design
speed of the existing alignment is unknown. Thus, an appropriate design speed
has to be estimated according to Chapter 1, Section 2.3.4:

i= n
CCR Si Li
(7)
CCR S = i=1 [gon/km]
i=n
L
i=1 i
It follows:
488 176 + 917 98
CCR S = = 641 gon/km
176 + 98
66 How to make two-lane rural roads safer

and
V85 = 106 / (8270 + 8.01 641) = 75 km/h

It was decided to select a design speed of

Vd = 80 km/h

because of the partially long independent tangents, where higher operating


speeds can be expected. In cases of more curvilinear alignment designs, the
selection of the lower design speed level of Vd = 75 km/h would have been more
appropriate. The design speed of 80 km/h is listed in Col. 7 of Table 16.

4. Evaluation of tangents (Table 16, Cols. 10 13). The tangent Elements 1


and 3 are evaluated in the following. In this respect, the calculation of TLmin
and TLmax is necessary for differentiating between independent and non-
independent tangents (see Chapter 2, Section 2.1).

Tangent Element 1
TL = 386 m

The curve in front of tangent Element 1 is already examined and was evaluated
as fair design. Therefore traffic warning devices seem to be sufficient and no
reconstruction work is planned. The operating speed in the preceding curve was
determined to be about V850 = 100 km/h.

It follows that:
2 2
( V850 ) ( V852 ) 2
100 82
2
TL min = = (9a)
22.03 22.03

TL min =149 m.
2 2 2
TL max =
(
2 V85Tmax ) ( V850 ) ( V852 ) (10a)
22.03
V85Tmax corresponds to eqn. (4a) for CCRS = 0 gon/km

V85Tmax 120 km/h


2 2 2
2 120 100 82
TL max = 548 m
22.03
Since
TLmin < TL < TLmax
Case studies 67

149 m < 386 m < 548 m,


Case 3 becomes relevant, and the present tangent is characterized as
independent.

For the calculation of the operating speed (V85T) eqn. (11a) for Case 3 in
Chapter 2, Section. 2.1 is valid:
2
(
V85T = 11.016 TL TL min + V850 ) ( ) (11a)
V850 > V852
(always use the larger value).

V85T = 11.016 ( 386 149 ) + 100


2

V85T = V851 = 112 km/h.

Tangent Element 3

TL = 234 m
2 2
( V852 ) ( V854 ) 2
82 64
2
TL min = =
22.03 22.03

TLmin = 119 m
2 2 2
TL max =
(
2 V85Tmax ) ( V852 ) ( V854 )
22.03

2 2 2
2 120 82 64
= = 816 m
22.03

Since
TLmin < TL < TLmax
119 m < 234 m < 816 m,

Case 3 becomes again relevant, and the present tangent is characterized as


independent.
For the calculation of the operating speed (V85T) eqn. (11a) for Case 3 in
Chapter 2, Section 2.1 is valid:
2
(
V85T = 11.016 TL TL min + V852 ) ( ) (11a)
68 How to make two-lane rural roads safer

V852 > V854

V85T = 11.016 ( 234 119 ) + 82 ( 2)


V85T = V853 = 89 km/h.

Finally, the operating speeds V85T are listed in Col. 14 of Table 16, as well
as in Cols. 6 and 15.

5. Side friction assumed/demanded (Table 16, Cols. 18 and 19). The


equations for side friction assumed / demanded are shown in Chapter 2,
Section 3.

According to Table 9, the formula for side friction assumed on existing


alignments is

fRA = 0.06 0.925 fT (13)


where
fT = 0.59 4.85 103 Vd + 1.51 105 Vd2 (12)
It follows.
fRA = 0.33 2.69 103 Vd + 0.84 105 Vd2

For the selected design speed of Vd = 80 km/h, side friction assumed is


fRA = 0.33 2.69 103 (80) + 0.84 105 (80)2
fRA = 0.17.
The value is listed for the curved sites 2 and 4 in Col. 18 in Table 16.
The actual side friction demanded is calculated from the following equation in
Table 9:
2
V85
f RD = e (14)
127 R
Thus, the side friction factors demanded can be calculated. In tangents fRD is
zero, because no centrifugal force exists.

Element 2
2
82
f RD = 0.07 = 0.69
2 127 70
Element 4
2
64
f RD = 0.07 = 0.63
4 127 46

The values for side friction demanded are listed in Col. 19 of Table 16.
Case studies 69

6. Results of Safety Criterion I (Table 16, Cols. 8 and 10). The


classification of SC I is explained in Chapter 2, Section 1 and in Table 11.
Accordingly the following solutions can be expected for:

Element 1
|V851 - Vd| = |112 - 80| = 32 (poor design)

Element 2
|V852 - Vd| = |82 - 80| = 2 (good design)

Element 3
|V853 - Vd| = |89 - 80| = 9 (good design)

Element 4
|V854 - Vd| = |64 - 80| = 16 (fair design)

The results are also presented as a graphical layout in Figure 13.

7. Results of Safety Criterion II (Table 16, Cols. 18 - 19). The classification


of SC II is explained in Subchapter 4.2 and Table 11. Accordingly the following
solutions can be expected for:

Transition between Elements 1 and 2:


|V851 V852| = |112 82| = 30 km/h (poor design).

Transition between Elements 2 and 3:


|V852 V853| = |82 89| = 7 km/h (good design).

Transition between Elements 3 and 4:


|V853 V854| = |89 64| = 25 km/h (poor design).

The results are also presented as a graphical layout in fig.13.

8. Results of Safety Criterion III (Table 16, Cols. 20 and 21). The
classification of SC III is explained in Chapter 2, Section 3 and in Table 11.
Accordingly, the following solutions can be expected for:

Element 2

f RA f RD = 0.17 0.69 = 0.52 (poor design)


2
Element 7

f RA f RD = 0.17 0.63 = 0.46 (poor design)


7

The results are also presented as a graphical layout in fig. 13.


70 How to make two-lane rural roads safer

9. Final evaluation of the investigated two-lane rural roadway section. The


results of Table 16 and the graphical layout in fig. 13 clearly indicate two
dangerous transitions between independent tangent Element 1 and curve Element
Sequence 2 as well as between independent tangent Element 3 and curve
Element Sequence 4. In addition the enormous side friction deficit in Element
Sequences 2 and 4, caused by the extremely narrow radii of curve and the
relatively high German operating speed background lead, with a high
probability, to a critical endangerment of parts of the investigated roadway
section.
In conclusion, the poor design range of Safety Criterion I for Element 1
(independent tangent) already indicates that the design and operating speed
differences between tangents and curves cannot be adjusted to the extent
necessary for a sound alignment. Therefore, as a solution, only a redesign with
radii of curve of at least 350 m can be recommended or the installment of
stationary radar devices to break the high operating speeds [1].

2. 2 Results of the safety module

Because of the shortness of the investigated roadway section consisting of two


curves and two tangents which means only three element (sequence) changes
the establishment of the safety module does not make much sense. In addition,
the main reasons for the high endangerment of parts of the investigated roadway
section are already distinctly clarified by the three safety criteria according to
Section 2.1 of this Chapter.

3 Example III
On the island of Crete/Greece a road survey was conducted for a major rural road
network in order to re-establish the missing alignment data and to evaluate the
road characteristics. In the context of the road-evaluation process the three safety
criteria, as presented in this book, were implemented in order to describe the
quality of the alignment.
The following example refers to a two-lane rural roadway section within the
network between stations 7.519 km and 10.123 km in flat topography with
grades less than 3 per cent, the lane width corresponds to 3.50 m and the AADT
values were 8000 vehicles/day in 1998.
The evaluation of the roadway section according to the three safety criteria is
presented, as follows.

3.1 Results of the safety criteria

The safety-evaluation processes according to Criterion I (Table 6), Criterion II


(Table 7) and Criterion III (Table 10) were performed to analyze the safety or
unsafety of the observed roadway section.
All necessary information about the design Elements 1 to 5 are listed in
Table 17. To perform the safety-evaluation processes the following basic
calculations have to be conducted. They are discussed in detail under Issues 1 to
9:
Case studies 71

Layout-Scheme

SC II

: SC I SC III

Legend:
IT: Independent Tangent
NIT: Non-Independent Tangent
SC: Safety Criterion

good design (+)

fair design (0)

poor design (-)

n.r. = not relevant, because either SC III does not exist or the tangent is too short (NIT)

Figure 13: Graphical presentation of the three safety criteria for good, fair or
poor design practices (Example II).

1. Curvature change rate of the single curve (Table 17, Col. 4). According
to Chapter 1, Section 1, the formulae for CCRS correspond to eqn. (1), and are
explained in fig.1. It follows then for:
72 How to make two-lane rural roads safer

Element Sequence 1 (radius with clothoids)

L L L
cl1 + cr1 + cl2 63, 700
2R R 2R
CCR S1 =
1 1 1
[gon/km] (1)
L

R1 = 495 m, Lcr1 = 352 m

A1 = 269 m, Lcl1 = 146 m L = Lcl1 + Lcr1 + Lcl2 = 585 m

A2 = 220 m, Lcl2 = 87 m

Control:

2 2
A1 (269)
L Cl1 = = = 146.2 m
R 495

2 2
A2 (208)
L Cl2 = = = 87.4 m
R 495

146.2 352 87.4


2 495 + 495 + 2 495 63, 700
CCR S1 =
= 103 gon/km
585

Element 2 (tangent)

R= CCRS2 = 0 gon/km

Element 3 (radius with clothoids)

L L L
cl3 + cr3 + cl4 63, 700
2R R 2R
CCR S3 =
3 3 3
[gon/km] (1)
L

R3 = 605 m, Lcr3 = 173 m

A3 = 287 m, Lcl3 = 136 m L = Lcl3 + Lcr3 + Lcl4 = 383 m

A4 = 212 m, Lcl4 = 74 m
Table 17: Numerical data for the safety evaluation process (Example III).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Database Application of the three safety criteria
No. Parameter CCRSi e Safety Criterion I Safety Criterion II Safety Criterion III
A, R, T Li V85i Vd |V85i-Vd| SC I TLmin TLmax NIT Case V85r V85i |V85i85i+1| SC II fRA fRD fRAfRD SC III
[] [m] [m] [gon/km] [%] [km/h] [km/h] [km/h] [m] [m] [km/h] [km/h] [km/h] [-] [-] [-]
A=269,
1 R=495, 585 103 4.5 91 95 4 + 91 0.15 0.09 0.06 +
A=220
ESC 8 +
2 T,R = 718 0 99 95 4 + 17 121 IT 2 99 99 n.r. n.r. n.r. n.r.
ESC 6 +
A=287,
3 R=605, 383 76 1.1 93 95 2 + 93 0.15 0.10 0.05 +
A=212
ESC 6 +
4 T,R = 582 0 99 95 4 + 17 88 IT 2 99 99 n.r. n.r. n.r. n.r.
ESC 4 +
A=289,
5 R=928, 335 48 1.7 95 95 0 + 95 0.15 0.06 0.09 +
A=324

Legend:
IT = Independent Tangent good design [+]
NIT = Non-Independent Tangent fair design [o]

Case studies 73
SC = Safety Criterion poor design []
ESC= Element (Sequence) Change
n.r. = not relevant, because either SC III does not exist, or the tangent is too short (NIT)
74 How to make two-lane rural roads safer

Control:

2 2
A3 (287)
L Cl3 = = = 136.2 m
R 605

2 2
A4 (212)
L Cl4 = = = 74.3 m
R 605

136 173 74
2 605 + 605 + 2 605 63, 700
CCR =
= 76 gon/km
S3 383

Element 4 (tangent)

R= CCRS4 = 0 gon/km

Element Sequence 5 (radius with clothoids)

L L L
cl5 + cr5 + cl6 63, 700
2R R 2R
CCR S5 =
5 5 5
[gon/km] (1)
L

R5 = 928 m, Lcr5 = 132 m


A5 = 289 m, Lcl5 = 90 m L = Lcl5 + Lcr5 + Lcl6 = 335 m
A6 = 324 m, Lcl6 = 113 m

Control:
2 2
A5 (289)
L Cl5 = = = 90.0 m
R 928

2 2
A6 (324)
L Cl6 = = = 113.1 m
R 928

90 132 113
2 928 + 928 + 2 928 63, 700
CCR S5 =
= 48 gon/km
335
The calculated CCRSi-values can be found in Col. 4 of Table 17.
Case studies 75

2. 85th-percentile speed (Table 17, Col. 6). For Greece, the 85th-percentile
speed is determined according to fig. 3 or Table 4:

6
10
V85 = [km/h] (4b)
i 10150.1 + 8.529 CCR
Si

Element Sequence 1
CCRS1 = 103 gon/km
V851 = 106 / (10150.1 + 8.529 103) = 91 km/h.

Element 2
CCRS2 = 0 gon/km

The determination of operating speeds in tangents (V85T) will be discussed in


detail under Issue 4 in the following. Accordingly, it follows that
V852 = V85T = 99 km/h.

Element 3
CCRS3 = 76 gon/km
V853 = 106 / (10150.1 + 8.529 76) = 93 km/h.

Element 4
CCRS4 = 0 gon/km

The determination of operating speeds in tangents (V85T) will be discussed in


detail under Issue 4 in the following. Accordingly, it follows that
V854 = V85T = 99 km/h.

Element Sequence 5
CCRS5 = 48 gon/km
V855 = 106 / (10150.1 + 8.529 48) = 95 km/h.
The calculated V85i-values are listed in Table 17, Col. 6.

3. Assessment of an appropriate design speed (Table 17, Col. 7). The design
speed of the existing alignment is unknown. Thus, in order to derive an estimate
of an appropriate design speed, the procedure described in Chapter 1, Section
2.3.4, was followed:

i= n
CCR Si Li
CCR S = i=1 [gon/km] (7)
i=n
L
i =1 i

For the existing alignment, it follows according to Issues 1 and 2 or Table 18:
76 How to make two-lane rural roads safer

103 586 + 76 383 + 48 335


CCR = = 81 gon/km
S 586 + 383 + 335
(without regarding tangent lengths).

From this CCRS it follows from Eq. 4b, that the average 85th-percentile speed
is

V85 = 106 / (10150.1 + 8.529 81) = 92 km/h

Based on this average 85th-percentile speed, a design speed of

Vd = 95 km/h

(see Table 17, Col. 9) was selected for the whole roadway section, and was
regarded as a reliable speed estimation for the existing alignment.

4. Evaluation of tangents (Table 17, Cols. 12-15). Elements 2 and 4 of the


existing alignment are tangents. The evaluation of tangents in the design process
is discussed in detail in Sec. 4.2.1. Since Element 2 with a length of 718 meters
and Element 4 with 582 meters (Col. 5) are both relatively long, it was found that
both elements correspond to independent tangents (IT), as shown in the
following calculations.

Tangent Element 2
TL = 718 m
To decide whether or not the present tangent is independent or non-independent
the three cases discussed in Chapter 4.2.1 have to be considered. However,
because of the present long tangents of Elements 2 and 4 only the calculation of
TLmax is necessary to compare this value with the actual tangent length (TL):

2 (V85Tmax ) 2 (V85 ) 2 (V853 ) 2


TL max = 1 (10a)
22.03
V85Tmax corresponds to Eq. 4b for CCRS = 0 gon/km

V85Tmax 99 km/h

2 (99) 2 (91) 2 (93) 2


TL max = 121 m .
22.03
Since
TL TLmax
Case studies 77

718 m 121 m,
Case 2 becomes relevant, and the present tangent is characterized as
independent.

The 85th-percentile speed on this tangent will be:

V85T = V852 = 99 km/h.

Tangent Element 4

TL = 582 m

2 2 2
2 (V85Tmax ) (V85 ) (V855 )
TL max = 3 (10a)
22.03

V85Tmax corresponds to eqn. (4b) for CCRS = 0 gon/km

V85Tmax 99 km/h

2 2 2
2 (99) (93) (95)
TL max = 88 m .
22.03

Since
TL TLmax
582 m 88 m,
Case 2 becomes relevant, and the present tangent also is characterized as
independent.

The 85th-percentile speed on this tangent will be:

V85T = V855 = 99 km/h.

Finally, the operating speed V85T is listed in Col. 14 of Table 17, as well as V852
and V854 in Cols. 6 and 15.

After the design speed and the 85th-percentile speeds along the investigated
roadway section are known, the safety-evaluation process with respect to
Criteria I and II can begin.

5. Side friction assumed/demanded (Table 17, Cols. 18 and 19). The


equations for side friction assumed/demanded are shown in Chapter 2, Section 3.

According to Table 9, the formula for side friction assumed on existing


alignments is:
78 How to make two-lane rural roads safer

fRA = 0.06 0.925 fT (13)


where
fT = 0.59 4.85 103 Vd + 1.51 105 Vd2 (12a).

It follows that:
fRA = 0.33 2.69 103 Vd + 0.84 105 Vd2
For the selected design speed of Vd = 95 km/h, side friction assumed is

fRA = 0.33 2.69 103 (95) + 0.84 105 (95)2


fRA = 0.15.

The value is listed for the curved sites 1, 3 and 5 in Col. 18 in Table 17.

The actual side friction demanded is calculated from the following equation in
Table 9:
2
V85
f RD = e (14)
127 R
Thus, the side friction factors demanded can be calculated for:

Element
2
91
f RD = 0.045 = 0.09
1 127 495

Element 2
In tangents fRD is zero, because no centrifugal force exists.

Element 3
2
93
f RD = 0.011 = 0.10
3 127 605

Element 4
Explanation, see Element 2.

Element 5
2
95
f RD = 0.017 = 0.06
5 127 928
The values for side friction demanded are listed in Col. 19 of Table 17.

6. Results of Safety Criterion I (Table 17, Cols. 8 and 9). The classification of
SC I is explained in Chapter 2, Section 1 and in Table 11. Accordingly the
following solutions can be expected for:
Case studies 79

Element 1
|V851 Vd| = |91 95| = 4 km/h (good design)

Element 2
|V851 Vd| = |99 95| = 4 km/h (good design)

Element 3
|V853 Vd| = |93 95| = 2 km/h (good design)

Element 4
|V854 Vd| = |99 95| = 4 km/h (good design)

Element 5
|V855 Vd| = |95 95| = 0 km/h (good design)

The results are also presented as a graphical layout in fig.14.

7. Results of Safety Criterion II (Table 17, Cols. 16 and 17). The


classification of SC II is explained in Chapter 2, Section 2 and in Table 11.

Transition between Curve 1 and the Independent Tangent 2 or vice versa:


|V851 V852| = |91 99| = 8 km/h (good design).

Transition between the Independent Tangent 2 and Curve 3 or vice versa:


|V852 V853| = |99 93| = 6 km/h (good design).

Transition between Curve 3 and the Independent Tangent 4 or vice versa:


|V853 V854| = |93 99| = 6 km/h (good design).

Transition between the Independent Tangent 4 and Curve 5 or vice versa:


|V854 V855| = |99 95| = 4 km/h (good design).

The results are also presented as a graphical layout in fig. 14.

8. Results of Safety Criterion III (Table 17, Cols. 20 and 21). The
classification of SC III is explained in Chapter 2, Section 3 and in Table 11.
Accordingly, the following solutions can be expected for:

Element 1

f RA f RD = 0.16 0.09 = 0.07 (good design)


1
Element 3

f RA f RD = 0.16 0.10 = 0.06 (good design)


3
80 How to make two-lane rural roads safer

Element 5

f RA f RD = 0.16 0.06 = 0.10 (good design)


5
The results are also presented as a graphical layout in fig. 14.

9. Final evaluation of the investigated two-lane rural roadway section.


From the listings and/or calculations of the input data in Table 17 and the
graphical layout in fig. 14, the safety-evaluation processes according to Criteria I
to III reveal, in comparison with the quantitative ranges of Table 11,
continuously good design practices.
The investigated roadway section in Crete/Greece represents an existing, but
relatively new alignment. The results confirm good design practices with respect
to design, operating speed, and driving dynamic consistency. This means only
low endangerments, if at all, can be expected from an alignment point of view.
However, the cross-sectional elements have also to be selected, reasonably.
Unfortunately, this is often not the case in Greece, since the paved shoulders
between 2.0 and 2.5 meters are too wide and may often lead to a 3-lane or even
4-lane driving behavior without any protection by medians on an original two-
lane rural road.
A graphical presentation of the numerical results in Table 17 is shown in fig.
15. The different design levels, based on the individual Safety Criteria I to III,
can be recognized visually by using symbols. The symbols are arranged for
Safety Criterion II vertically to the road axis, whereas the symbols for Safety
Criterion I are located on the left side and those for Safety Criterion III are
located on the right side, parallel to the axis.

3.2 Results of the safety module

The theoretical background of the safety module was discussed in Chapter 2,


Section 5. Furthermore, an evaluation scheme for practical applications was
developed in Table 12, based on the three safety criteria.
As already explained in Section 1.2 of this Chapter, each criterion describes
one safety aspect, while the safety module represents, to a certain extent, an
overall averaged evaluation. While for the evaluation of individual design
elements and the corresponding transitions, the individual safety criteria are to be
preferred, the safety module could be of great advantage for road-network
evaluations [25, 26].
The evaluation of the Safety Module for the driving direction A to F
according to fig. 15 is shown by the table in the upper part of Table 18, whereas
the table in the lower part represents the driving direction F to A.
Exemplarily, the calculation process for the Safety Module for the driving
direction A to F is conducted in the upper part of Table 18. The results can be
found in Col. 9 of Table 18 and parallel to the axis of fig.15.

Section A to B consists of Element 1 (fig. 15),


Table 18 reveals:
Case studies 81

SC I = + Weighting Factor = +1.0 (Table 12)


SC II = + Weighting Factor = +1.0 (Table 12)
SC III = + Weighting Factor = +1.0 (Table 12)
SUM +3.0 : 3 = 1.0

This result suggests according to Table 12 good design, as presented in


Table 18 and fig. 15.

Section B to C consists of Element 2 (fig. 15),


Table 18 shows:

SC I = + Weighting Factor = +1.0


SC II = + Weighting Factor = +1.0
SUM +2.0 : 2 = 1.0
Safety Criterion III is not relevant.

The result implies, according to Table 12, good design, as presented in


Table 18 and fig.15.

Section C to D consists of Element 3 (fig. 15),


Table 18 reveals:

SC I = + Weighting Factor = +1.0


SC II = + Weighting Factor = +1.0
SC III = + Weighting Factor = +1.0
SUM +3.0 : 3 = 1.0

According to Table 12, good design can be determined, as presented in


Table 18 and fig.15.

Section D to E consists of Element 4 (fig. 15),


Table 18 reveals:

SC I = + Weighting Factor = +1.0


SC II = + Weighting Factor = +1.0
SUM +2.0 : 2 = 1.0
Safety Criterion III is not relevant.

According to Table 12, good design can be determined, as presented in


Table 18 and fig.15.
Section E to F consists of Element 5 (fig. 15).
The calculation of the safety module is not possible, since the element that
follows point F, is unknown and therefore, Safety Criterion II cannot be
determined.
82 How to make two-lane rural roads safer

Figure 14: Graphical presentation of the three Safety Criteria for good, fair or
poor design practices (Example III).
Case studies 83

Comparing the upper and the lower parts of Table 18 and fig. 15, it becomes
obvious that the safety module represents good design for all elements,
respectively, for all elements sequences of the investigated roadway section.
This means that no corrective measures or safety improvements have to be
considered.
According to the opinion of the authors the additional presentation of the
graphical layout for the opposite driving direction F to A would go too far and
will not increase the comprehension of the procedure.

4 Example IV
In the following, a short roadway section of a two-lane rural road in Sicily,
Italy,will be examined according to the three safety criteria because of a
relatively high accident situation at curved sites. The accident data reveal
8 accidents with injuries and/or fatalities between 1997 and 2001.
The section length is about 1 km, the lane width is 3.75 m plus paved shoulder,
1.50 m wide on both sides. The road is located in flat topography with
longitudinal grades less than 3 %. The AADT-values were 10.850 vehicles per
day in 2001. The horizontal alignment is composed of a sequence of circular
curves and tangents without clothoids as transition curves.
4.1 Results of the Safety Criteria

The safety-evaluation processes according to Criterion I (Table 6), Criterion II


(Table 7) and Criterion III (Table 10) were performed to analyze the safety or
unsafety of the observed roadway stretch. All necessary data are listed in
Table 19.

1. Curvature change rate of the single curve (Table 19, Col. 4)

Element 1 (tangent)

R= CCRS1 = 0 gon/km

Element 2 (circular curve)

R2 = 300 Lcr2 = 155 L = Lcr2

L cr2
63, 700
R2
CCR S2 = [gon/km] (1)
L

155
63, 700
CCR S2 = 300 = 212 gon/km
155
84 How to make two-lane rural roads safer

Element 3 (tangent)

R= CCRS3 = 0 gon/km

Element 4 (circular curve)

R4 = 180 Lcr4 = 151 L = Lcr4

L cr4
63, 700
R4
CCR S4 = [gon/km]
L

151
63, 700
CCR S2 = 180 = 354 gon/km
151

The calculated CCRSi-values can be found in Col. 4 of Table 19.

2. 85th-percentile speed (Table 19, Col. 6). For Italy an individual operating
speed was established according to fig. 3 and Table 4. Thus the formula of the
Italian operating speed background reads:
V85 = 118.9 0.062 CCRSi [km/h] (4h)
Element 1 (tangent)
CCRS1 = 0 gon/km
Operating speeds in tangents (V85T) are subject of Issue 4.
Accordingly, it follows, that V851 = V85Tmax 119 km/h.

Element 2 (circular curve)


CCRS2 = 212 gon/km
V852 = 118.9 0.062 212 = 106 [km/h]

Element 3 (tangent)
CCRS3 = 0 gon/km
Operating speeds in tangents (V85T) are subject of Issue 4.
Accordingly, it follows, that V853 = V85T 107 km/h.

Element 4 (circular curve)


CCRS4 = 354 gon/km
V854 = 118.9 0.062 354 = 97 [km/h]

The calculated V85i-values can be found in Col. 6 of Table 19.


Case studies 85

Table 18: Safety module for both driving directions (Example III).

DRIVING DIRECTION A to F
1 2 3 4 5 6 7 8 9 10
Safety
No. Station Parameter Safety Criteria Designations
module
according to
from to A, R, T Li SC I SC II SC III
fig. 12
[-] [km] [km] [m] [m]
A1=269
1 7.520 8.105 R1=495 585 + +
+ A to B
A2=220
ESC +
2 8.105 8.823 T, R= 718 + n.r.
+ B to C
ESC +
A3=287
3 8.823 9.206 R3=605 383 + +
A4=212 + C to D
ESC +
4 9.206 9.788 T, R= 582 + n.r.
+ D to E
ESC +
A5=289
5 9.788 10.123 R5=928 335 + + E to F
A6=324
DRIVING DIRECTION F to A
1 2 3 4 5 6 7 8 9 10
Safety
No. Station Parameter Safety Criteria Designations
module
according to
from to A, R, T Li SC I SC II SC III
fig. 12
[-] [km] [km] [m] [m]
A1=269
1 7.520 8.105 R1=495 585 + + B to A
A2=220
ESC +
+ C to B
2 8.105 8.823 T, R= 718 + n.r.
ESC +
A3=287 + D to C
3 8.823 9.206 R3=605 383 + +
A4=212
ESC +
+ E to D
4 9.206 9.788 T, R= 582 + n.r.
ESC +
A5=289 + F to E
5 9.788 10.123 R5=928 335 + +
A6=324
Legend:
SC Safety Criterion n.r. not relevant
ESC Element (Sequence) Change Driving Direction
+ good design No statement possible, since the
o fair design preceding or succeeding element is
poor design unknown
86 How to make two-lane rural roads safer

Figure 15: Graphical presentation of the safety module for Example III (driving
direction A to F).
Case studies 87

3. Assessment of an appropriate design speed (Table 19, Col. 7). The design
speed of the existing alignment is unknown. Thus, an appropriate design speed
has to be estimated according to Chapter 1, Section 2.3.4:

i= n
CCR Si Li
CCR S = =1
i (7)
i=n
L
i =1 i

212 155 + 354 151


CCR S = = 282 gon / km
155 + 151

V85 = 118.9 0.062 CCRS = 101 km/h and


V85 = 118.9 0.062 282 = 101 km/h
It was decided to select a design speed of
Vd = 100 km/h

4. Evaluation of tangents (Table 19, Col. 10-13). Elements 1 and 3 of


the existing alignment are tangents. The evaluation of tangents in the design
process is discussed in detail in Chapter 2, Section. 2.1.

Tangent Element 1
TL = 591 m
The operating speed of the preceding element of the tangent was determined to
be about V850 = 110 km/h.

It follows that:
2 2 2 2
(V850 ) (V852 ) 110 106
TL min = = (9a)
22.03 22.03
TLmin = 39 m
V85Tmax corresponds to eqn. (4h) for CCRS = 0 gon/km

V85Tmax = 119 km/h

2 2 2 2 2 2
2 (V 85T max ) (V 850 ) (V 852 ) 2 119 110 106
TLmax = = (10a)
22.03 22.03

TLmax = 226 m
88 How to make two-lane rural roads safer
Table 19: Numerical data for the safety evaluation process (Example IV).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Database Application of the three safety criteria
No. Parameter CCRSi e Safety Criterion I Safety Criterion II Safety Criterion III
A, R, T Li V85i Vd |V85i-Vd| SC I TLmin TLmax NIT Case V85r V85i |V85i85i+1| SC II fRA fRD fRAfRD SC III
[] [m] [m] [gon/km] [%] [km/h] [km/h] [km/h] [m] [m] [km/h] [km/h] [km/h] [-] [-] [-]
1 T,R = 591 0 2.5 119 100 19 o 39 226 IT 2 119 119 n.r. n.r. n.r. n.r.
ESC 13 o
2 R=300 155 212 6 106 100 6 + 106 0.15 0.23 0.08 -
ESC 1 +
3 T,R = 94 0 2.5 107 100 7 + 83 346 IT 3 107 107 n.r. n.r. n.r. n.r.
ESC 10 +
4 R=180 151 354 7 97 100 3 + 97 0.15 0.34 0.19 -

Legend:

IT = Independent Tangent good design [+]


NIT = Non-Independent Tangent fair design [o]
SC = Safety Criterion poor design []
ESC = Element (Sequence) Change
n.r. = not relevant, because either SC III does not exist, or the tangent is too short (NIT)
Case studies 89

Since
TL > TLmax
591 m > 226 m,
Case 2 becomes relevant and the present tangent is characterized as
independent.
The existing tangent element is long enough to allow an acceleration and
deceleration maneuver up to the maximum operating speed (V85Tmax) on tangent
element 1.

V85T = V85Tmax = 119 km/h.

Tangent Element 3
TL = 94 m

It follows that:
2 2 2 2
(V 852 ) (V 854 ) 106 97
TLmin = =
22.03 22.03
TLmin = 83 m

V85Tmax corresponds to eqn.(4h) for CCRS = 0 gon/km


V85Tmax = 119 km/h

V85Tmax = 119 km/h

2 (V85 Tmax ) 2 (V85 2 ) 2 (V85 4 ) 2 2 119 2 106 2 97 2 (10a)


TL max = =
22.03 22.03

TLmax = 346 m

Since
TLmin < TL < TLmax
83 m < 94 m < 346 m,
Case 3 becomes relevant and the present tangent is characterized as
independent.
For the calculation of the operating speed (V85T) eqn. (11a) for Case 3 in
Chapter 2, Section 2.1 is valid:

2
V 85T = 11.016 (TL TLmin ) + (V 852 ) (11a)
V852 > V854 (always use the larger value)

2
V 85T = 11.016 (94 83) + (106)

V85T = V853 = 107 km/h.


90 How to make two-lane rural roads safer

Finally, the operating speed V85T element 3 is listed in Col. 14 of Table 19, as
well as in Cols. 6 and 15.

5. Side friction assumed/demanded (Table 19, Cols. 18 and 19). The


equations for side friction assumed/demanded are shown in Chapter 2, Section 3.
According to Table 9, the formula for side friction assumed on existing
alignments is:

fRA = 0.06 0.925 fT, (13)


where
fT = 0.59 4.85 103 Vd + 1.51 105 Vd2 (12a)
It follows that:
fRA = 0.33 2.69 103 Vd + 0.84 105 Vd2
For the selected design speed of Vd = 100 km/h, side friction assumed is
fRA = 0.33 2.69 103 100 + 0.84 105 1002
fRA = 0.15
The value is listed for the curved sites 2 and 4 in Col. 18 in Table 19.

The actual side friction demanded is calculated from the following equation
in Table 9.
2
V 85
f RD = e (14)
127 R
Thus, the side friction factors demanded can be calculated. In tangents fRD
is zero, because no centrifugal force exists.

Element 2
2
106
f RD = 0.06 = 0.23
2 127 300

Element 4
2
97
f RD = 0.07 = 0.34
4 127 180
The values for side friction demanded are listed in Col. 19 of Table 19.

6. Results of Safety Criterion I (Table 19, Cols. 8 and 9). The classification of
SC I is explained in Chapter 2, Section 1 and in Table 11. Accordingly, the
following solutions can be expected for:

Element 1
V851 Vd = 119 100 = 19 km/h (fair design)
Element 2
V852 Vd = 106 100 = 6 km/h (good design)
Case studies 91

Element 3
V853 Vd = 107 100 = 7 km/h (good design)

Element 4
V854 Vd = 97 100 = 3 km/h (good design)

The results are also presented as a graphical layout in fig. 16.

7. Results of Safety Criterion II (Table 19, Cols. 16 and 17). The


classification of SC II is explained in Chapter 2, Section 2 and in Table 11.
Accordingly, the following solutions can be expected for:

Transition between Elements 1 and 2:


V851 V852 = 119 106 = 13 km/h (fair design)

Transition between Element 2 and 3:


V852 V853 = 106 107 = 1 km/h (good design)

Transition between Element 3 and 4:


V852 V854 = 107 97 = 10 km/h (good design)

The results are also presented as a graphical layout in fig. 16.

8. Results of Safety Criterion III (Table 19, Cols. 20 and 21). The
classification of SC III is explained in Chapter 2, Section 3 and in Table 11.
Accordingly, the following solutions can be expected for:

Element 2
f RA = f RD = 0.15 0.23 = 0.08 (poor design)
2
Element 4
f RA = f RD = 0.15 0.34 = 0.19 (poor design)
4
The results are also presented as a graphical layout in fig. 16.

9. Final evaluation of the investigated two-lane rural roadway section. From


the listings and/or calculations of the input data in Table 19, the safety-
evaluation processes for Criterion I, Criterion II and Criterion III were carried
out in conjunction with the quantitative ranges of Table 11, in order to
distinguish good design from fair and poor design practices. In tangents, Safety
Criterion III is not relevant, since no centrifugal force exists. The results of
Table 19 and the graphical layout in fig. 16 clearly indicate that on Element 1 the
difference between V85 and Vd is 19 km/h and corresponds to fair design
according to the ranges of Safety Criterion I, represented in Table 11. The same
is true for Safety Criterion II with respect to the V85-speed differences between
Elements 1 and 2.
92 How to make two-lane rural roads safer

In addition, with respect to Safety Criterion III the side friction deficit in
Elements 2 and 4 is caused by the relatively high Italian operating speed
background, which leads to a high side friction demand, often not covered,
especially under wet road surface conditions and/or inadequate tires. Thus, the
relatively high accident situation with injuries and fatalities at curved sites
becomes understandable. Therefore, the speed behavior should be lowered
through the application of speed limits, strongly controlled by the police, and
appropriate traffic-control devices (for example: chevrons). The best solution
would be warranted by stationary radar devices. Reconstruction or RRR
strategies do not seem to be necessary at the moment.

4.2 Results of the Safety Module

Since the establishment of the Safety Module was already explained on the basis
of two former examples, its development will not be discussed here again.
Case studies 93

Figure 16: Graphical presentation of the three Safety Criteria for good, fair or
poor design practices (Example IV).
94 How to make two-lane rural roads safer
Chapter 5

Influence of road equipment on traffic safety

On the basis of about 50 case studies in different countries and continents, it can
be demonstrated that the presented concept soundly represents the safety
classification of roadway sections according to good, fair (tolerable), and poor
design practices, and sensible results can be expected. Furthermore, it can be
demonstrated, based on large accident databases, that a strong tendency for a
good agreement exists between safety criteria and the actual accident situation.
However, based on new research work of Beck [9], it was expected that besides
the design parameters, the road equipment also has an influence on the accident
situation. Therefore, the basic relationships between highway geometric design,
accident situation and road equipment should additionally be clarified, and
through field investigations it was found that typical levels of road equipment
can be defined, as follows (fig. 17):

Level 1 Road Markings: edgeline marking, solid centerline, broken


centerline, etc.

Level 2 Traffic-Control Devices: curve warning sign, reverse turn warning


sign, hill warning sign, speed limit sign, chevron alignment sign with up to
3 arrows (individual or on one board), as well as combinations.

Level 3 Traffic-Control Devices: road equipment that exceeds level 2, for


example, multiple chevron alignment signs with more than 3 arrows (individual
or on one board), as well as combinations with level 2.

The following investigations include 79 sections of two-lane rural roads with


an overall length of 212 kilometers, which consist of 1466 individual elements
(curves or tangents). The overall number of recorded Run-Off-the-Road
accidents and Deer accidents was 723 within three years [18, 20].
The influence of the three levels of road equipment on the accident rate and
the accident cost rate was individually investigated for the design parameters:
pavement width, radius of curve, and curvature change rate of the single curve.
96 How to make two-lane rural roads safer

1 Pavement Width
Figure 18 reveals the relationships between pavement width, accident rate and
accident cost rate for the three levels of road equipment. As can be seen, a U-
shaped relationship can be expected, whereby level 3 represents the highest and
level 1 the lowest course. The adaptation of the regression curves for levels 1 to
3 in the range between 6.50 m and 7.50 m is remarkable. Considering the
intensifying of road-equipment measures at critical roadway sections the positive
effect especially of level 3 but also of level 2 becomes obvious.

Figure 17: Recorded road markings and traffic signs [18, 20]
Influence of road equipment on traffic safety 97

Based on the similar trends of accident rates and accident cost rates in fig. 18,
one can recognize that accident frequency and severity are closely connected, at
least for the pavement width a very interesting result. In conclusion, it can be
stated that for two-lane rural roads especially pavement widths between 6.50 m
and 7.50 m represent favorable results regarding the accident situation in general
[9] and with respect to the three levels of road equipment [20]. In the following,
only the results of the accident rate are shown in the graphs, since they reveal
comparable trends with respect to the accident cost rate [18].

16,00

14,00

Level 1; R = 0.913
12,00
Level 2; R = 0.954
veh.-km]

10,00 Level 3; R = 0.935


6

8,00
AR [ACC/10

6,00

4,00

2,00

0,00
400 450 500 550 600 650 700 750 800 850 900
Pavement Width [cm]

140,00

120,00
ACR [German Marks/100veh.-km]

Level 1; R = 0.956
100,00
Level 2; R = 0.939

80,00 Level 3; R = 0.945

60,00

40,00

20,00

0,00
400 450 500 550 600 650 700 750 800 850 900
Pavement Width [cm]

Figure 18: Relationships between accident rate, accident cost rate and pavement
width for the three levels of road equipment [18, 20].
98 How to make two-lane rural roads safer

2 Radius of Curve
Figure 19 shows the relationships between radius of curve and accident rate for
the three levels of road equipment.
It is interesting to note that for radii of curve R < 150 m, level 1 does not play
any role, which means only the road-equipment levels 2 and 3 are present in
endangered radii of curve-ranges. This indicates that sensible safety devices are
used in those endangered roadway sections by the responsible authorities.
According to fig. 19 levels 2 and 3 reveal strong decreasing trends, with
increasing radii of curve. In this connection, level 3 offers significantly more
safety in contrast to level 2 up to radii of curve of R 250 m. The same is true
for the accident severity, expressed by the accident cost rate [18].
According to fig. 19 and the above statements, the use of road equipment-
level 3 is urgently recommended, at least for radii of curve less than 250 m,
while for greater radii up to about 400 m levels 1 and 2 appear to be sufficient.

3 Curvature change rate of the single curve


Figure 20 shows the relationships between the accident rate and the curvature
change rate of the single curve for the three road equipment-levels. Up to about
300 gon/km (this corresponds, without considering transition curves, to roughly
the radii of curve of R > 220 m) the regression curves of levels 1 and 2 are nearly
identical, which means in more critical areas the signing according to level 2
lowers the accident rates down to a classification according to level 1.
Correspondingly, levels 2 and 3 reveal also a nearly identical course between
300 gon/km and 500 gon/km (R 200 m to R 130 m). Note, in this case the
road equipment-level 3 reduces the high safety deficiencies to those comparable
to level 2. Field investigations have shown that such a success could be reached
especially through the repetition of multiple chevron alignment signs with more
than three arrows (individual or more than one board, according to fig. 21).
Beginning with CCRS 450 gon/km (which corresponds roughly without
considering transition curves to R 150 m) level 3 reveals significant
improvements in contrast to level 2 signing. This leads to the request, that
inconsistencies in the alignment have to be either redesigned or reconstructed, at
least for CCRS-values greater than 450 gon/km. If that is not possible, they
should be secured by signing according to level 3, see fig. 21, as an example.
In this connection, the equipment with multiple chevron alignment signs and
guardrails throughout the curve can significantly improve the optical guidance,
especially at night and under wet surface conditions.

4 Road equipment and design (curvature change rate) classes


With respect to the new database [20], additionally including the three road
equipment levels, the mean accident rates, respectively mean accident cost rates
were again calculated for the good, fair and poor design classes in Table 20 in
accordance with Tables 1 and 2. As can be seen, Col. 2 shows the results for the
overall database, Col. 3 for level 1 road equipment, Col. 4 for level 2 road
equipment and Col. 5 for level 3 road equipment.
Influence of road equipment on traffic safety 99

10,00

8,00

Level 1; R = 0.618
veh.-km]

Level 2; R = 0.864
6,00

Level 3; R = 0.972
6
AR [Acc./10

4,00

2,00

0,00
0 50 100 150 200 250 300 350 400 450 500
Radius of Curve [m]

Figure 19: Relationships between accident rate and radius of curve for the three
levels of road equipment [18, 20],

10,00

Level 1; R = 0.772

8,00 Level 2; R = 0.973


veh.-km]

Level 3; R = 0.994

6,00
6
AR [Acc./10

4,00

2,00

0,00
0 50 100 150 200 250 300 350 400 450 500 550 600 650 700
Curvature Change Rate of the Single Curve [gon/km]

Figure 20: Relationships between accident rate and curvature change rate of the
single curve for the three levels of road equipment [18, 20].
100 How to make two-lane rural roads safer

Figure 21:Curve, guided by multiple chevron alignment signs and guardrails


[18].

It is to be noted that level 3 does not exist in the good design range. The same is
true for level 1, which does not exist in the poor design range. This is
understandable, since normally level 1 is not able to improve poor design
practices, and the application of level 3 in the good design range would make no
sense. For the investigated equipment levels (Table 20, Cols. 35), there always
exist increases in accident rates and accident cost rates between the good, fair,
and poor design classes when compared. This means that even the strongest road
equipment level, for example one with multiple chevron alignment signs,
guardrails, etc. (see, fig. 21), is not able to influence an originally poor design in
such a way that accident rates and accident cost rates could reach values
representing the good design levels, for example, of the overall database
(Table 20, Col. 2).
Further important results with respect to the superimposition of design
classes and road equipment levels according to the conducted research [20] and
Table 20 are:

1. Obviously, the individual road equipment measures are used by the


responsible authorities according to the actual local accident history, at least
in Germany.

2. There exists a strong superimposition between the design classes (good,


fair, poor) and the road equipment levels 1 to 3.
Influence of road equipment on traffic safety 101

Table 20: Mean accident rates and cost rates for different design (CCRS) classes
and road equipment levels [18, 20]

CCRS/design Mean AR* Mean AR Mean AR Mean AR


classes overall Level 1 Level 2 Level 3
[gon/km] database
(1) (2) (3) (4) (5)
> 35 0.48 0.46 0.47 -
180 good
> 180 0.77 0.80 0.84 0.56
360 fair
> 360 poor 1.69 - 1.92 1.84

CCRS/Design Mean ACR** Mean ACR Mean ACR Mean ACR


Classes overall Level 1 Level 2 Level 3
[gon/km] database
(1) (2) (3) (4) (5)
> 35 2.94 2.58 3.16 -
180 good
> 180 5.96 6.59 6.20 4.05
360 fair
> 360 poor 14.57 - 20.12 12.85
6
* AR = accident rate [acc./10 veh.-km]
** ACR = accident cost rate [German Marks/100 veh.-km]

3. The responsible authorities attempt to compensate increasing


endangerments within the individual design classes by a more intensive
road equipment level.

4. The design class poor is still decisively more dangerous than the design
class good (compare, Cols. 4 and 5 with Col. 2, for example), although
level 3 has obviously a better impact on the accident development, as
compared with levels 1 or 2.

5. On dangerous roadway sections level 3 leads to better results than level 2


with respect to accident risk and accident severity, as Cols. 4 and 5 of
Table 20 clearly reveal.

Thus, fundamental knowledge could be gained between design (CCRS)-


classes and sensible road equipment levels, in such a way that for endangered
and dangerous curved sites, level 2 and especially level 3 may lead to strong
reductions, respectively, adaptations with respect to accident risk and severity.
Therefore, an appropriate application of road equipment levels normally
influences traffic safety positively.
102 How to make two-lane rural roads safer

Note that level 2 or even level 3 signing has proved to improve traffic safety,
however not to a level which would correspond to good design practices
according to the discussed Safety Criteria I to III. Therefore, as an interim
solution signing can be recommended, but normally the reduced accident
situation remains nevertheless at a fair or even poor design level and only
redesigns combined or not with RRR-strategies promise, if at all, safe
solutions.

5 Road equipment and Safety Criteria


In the last part of this Chapter the evidence of the results between safety criteria
(Table 11) and actual accident situations were examined.
As the most important result, it can clearly be confirmed that the three
quantitative Safety Criteria are suitable for the classification of roadway sections
according to good, fair (tolerable) and poor design practices. In this connection,
Table 21 proves that relatively low accident rates and accident cost rates can be
expected for good designs, whereas relatively high accident rates and accident
cost rates normally represent poor design practices.

Table 21: Relationships between good/poor design practices and mean accident
rates and cost rates for 99 curved sections [18, 20].

Safety evaluation Mean AR Mean ACR Number of investigated


curves
Good design 0.23 1.56 69

Poor design 1.94 15.81 30

Note that curved roadway sections, which are classified by the safety criteria
as good design in comparison to those, classified as poor design represent
for poor design still about 10 times higher accident rates and cost rates than for
good design, despite the application of the most stringent traffic control devices
according to level 3. Besides, it was found that curved sections with low
endangerment potential, in general, are equipped according to level 1, whereas
curved sections with relatively high endangerment potential reveal for the most
part traffic-control devices according to level 3. Nevertheless, even level 3-road
equipment is often not able to sufficiently diminish the danger of accidents at
critical roadway sections. This means furthermore that in those cases, redesign,
reconstruction or RRR strategies are in the forefront for improving traffic safety
or the installment of stationary radar devices becomes necessary, for example, to
reduce excessive speeds.
Conclusion and outlook

The belief, generally, is that a road designed to standard is safe. However, when
a road goes into operation, the accident experience afterwards is the only
indicator of the safety performance of the road. Therefore, it is necessary to
create an explicit correlation between the standard of the road as designed and its
anticipated safety.
It has been found that 50 per cent or even more of the fatalities occur on two-
lane rural roads. At least half of these occur on curved roadway sections. To
address these fatalities, a practical procedure, which considers driving behavioral
and safety rules for the evaluation of new designs, redesigns, and RRR-projects,
was developed in this book. The methodology is based on the new design
parameter Curvature Change Rate of the Single Curve. This parameter was
tested against several databases of accident rates and accident cost rates and
found to be the major descriptor of the safety of the road. The same is true with
respect to operating speeds. With respect to the curvature change rate of the
single curve and on the basis of large operating speed and accident databases in
Europe and the USA, design classes were developed to classify, from a traffic
safety point of view, roadway sections as good, fair or poor designs. The design
classes are tuned with three safety criteria to develop an overall quantitative
safety-evaluation procedure for new designs, existing and old alignments of two-
lane rural roads. The safety criteria are introduced to analyze and evaluate by

Criterion I design consistency which corresponds to relating the design


speed with the actual driving behavior, which is expressed by
the 85th-percentile speed of passenger cars under free-flow
conditions;
Criterion II operating-speed consistency which seeks uniformity of 85th-
percentile speeds through successive elements of the road; and
Criterion III consistency in driving dynamics which relates side friction
assumed with respect to the design speed to that demanded at
the 85th-percentile speed.

Criterion I is a measure of the consistency of the alignment. Criterion II


reflects the harmony (or disharmony) between operating speeds on successive
design elements. Criterion III refers to the adequacy of the dynamic safety
provided. All three criteria are evaluated in terms of three ranges, described as
104 How to make two-lane rural roads safer

Good, Fair and Poor. Cut-off values between the three ranges are
developed.
The term operating speed is nowadays well defined and is used in
conjunction with the new design parameter Curvature Change Rate of the
Single Curve to describe the road characteristics in combination with operating
speed backgrounds for many countries. However, for the overwhelming majority
of existing roadways the design speed is not known or was roughly assessed in
the past. Therefore, a new procedure, which takes into account the overall
characteristics of the roadway, was developed in order to assign sound design
speeds for new designs, redesigns or RRR projects. Based on these calculated
design speeds, redesigns or RRR (Resurfacing, Restoration, and Rehabilitation)
projects can be undertaken by changing the alignment to the extent necessary to
remedy any detected individual or combined safety problem (like design,
operating speed or driving dynamic deficiencies), simultaneously regarding
important economic and environmental issues.
Furthermore, for the good, fair and poor design ranges of the developed
safety criteria relation design backgrounds could be established. By applying
relation design good curvilinear alignments can be achieved. However, in the
case of redesigns, curvilinear aspects do not often lead to optimum solutions. For
this reason, the relation design considerations are extended in this book in order
to achieve additionally a well-balanced design between independent (long)
tangents and curves.
Finally, recommendations are given for adequate radii of curve with or
without transition curves following tangents, and for tangents that should be
regarded as independent or non-independent elements in the design process.
Thus, the safety-evaluation process expresses the need for achieving design
and operating speed, and driving dynamic consistency. With respect to the
design component Alignment it could be shown that the three safety criteria
and the introduction of sound tangential and side friction factors have significant
safety impacts on four out of five alignment subcomponents. These are:

- Horizontal alignment
- Vertical alignment
- Cross section/alignment
- Sight Distance.

Only the subcomponent three-dimensional alignment could not be


incorporated directly into the overall safety-evaluation process. It also represents
the weakest link in highway geometric design.
In order to compare the actual accident situation with the results of the three
safety criteria numerous existing alignments were analyzed. The results
convincingly demonstrate that the proposed classification system for evaluating
good and for detecting fair and poor design practices is valid and that it can
strongly support the work of the highway engineer to reduce accident risk and
severity.
Conclusion and outlook 105

Finally, the results of the three safety criteria are compared with the actual
accident situation. The results confirmed in a convincing manner that a
statistically significant relationship exists between the outcome of the three
safety criteria and the actual accident rates. By using the good ranges for the
three safety criteria sound alignments in plan and profile can be achieved, which
are well tuned to the expected driving behavior of the motorists and may reduce
significantly accident risk and severity.
So far, for easing the danger of accident spots, accidents already had to have
occurred, in order to determine that the spot or the roadway section is dangerous,
for example, for future decision making of countermeasures. The great advantage
of the safety concept is that the analyst can, in the design stages, predict the
endangerment (low, medium, high) for new alignments. Additionally, the three
safety criteria are also appropriate to render judgments about the safety
conditions of existing (old) roadway sections or whole road networks. In
consequence, the highway and traffic safety engineer can quantitatively evaluate
the expected accident situation. Deficiencies in new designs can be corrected
prior to construction and sound countermeasures can be planned for highly
endangered existing or old alignments.
Many case studies where analyzed. The results clearly indicate a statistically
significant relationship between the results of the three individual safety criteria
and the actual accident rates to identify good (low endangerment), fair (medium
endangerment) and poor (high endangerment) design practices. This is true for
new designs, redesigns, and RRR-practices, as well as for the examination of
existing (old) alignments.
For a general evaluation process, the three safety criteria were combined into
an overall safety module. This module represents the current state of knowledge.
The discussed safety criteria constitute the core of the safety module proposed
for classifying road networks and roadway sections existing or planned as
good, fair or poor designs.
Finally, three road equipment levels with respect to individual design
parameters and relative accident numbers were investigated. It was found that the
application of signing and guardrails is obviously conducted by the responsible
authorities according to the level of endangerment of the respective roadway
section, at least in Germany. In particular, when regarding the new design
parameter Curvature Change Rate of the Single Curve with respect to the
accident rate in fig. 20, the sensible classification of road equipment according to
Levels 1 to 3 could be confirmed. Both the design (CCRS)-classes and the
derived safety-evaluation process are superimposed by the road equipment
levels. This often leads to a reduction, respectively, to an adaptation of accident
risk and accident severity, however, certainly not to a weakening or even to a
questioning of the developed safety conception.
106 How to make two-lane rural roads safer
References
[1] Lamm, R., Psarianos, B., Mailaender, T., Choueiri, E. M., Heger, R., &
Steyer, R. Highway Design and Traffic Safety Engineering Handbook,
McGraw-Hill, Professional Book Group, New York, N.Y., USA, 1999,
932 pages, ISBN 0-07-038295-6, Language Editors: J.C. Hayward, E.M.
Choueiri, & J.A. Quay.

[2] Feuchtinger, M.E., & Christoffers, C., Driving Dynamic Investigations


as Measure of Road-Traffic-Safety, Journal for Traffic Safety
(Zeitschrift fuer Verkehrssicherheit), vol. 1, Germany, 1953.

[3] Bitzl, F., The Safety Level of Roads, Research Road Construction and
Road Traffic Technique (Forschung Strassenbau und
Strassenverkehrstechnik), vol. 28, Minister of Transportation, Bonn,
Germany, 1964.

[4] Krebs, H.G., & Kloeckner, J. H. Investigations of the Effect of Highway


and Traffic Conditions Outside Built-Up Areas on Accident Rates,
Research Road Construction and Road Traffic Technique (Forschung
Strassenbau und Strassenverkehrstechnik), Minister of Transportation,
Bonn, Germany, 1977, vol. 223, pp. 163.

[5] Schlichter, H.G., Road Characteristics - An Analytical Evaluation Road


and Interstate (Strasse und Autobahn), vol. 2, Germany, 1976, pp. 55-58.

[6] Lamm, R., Driving Dynamics and Road Characteristics - A Contribution


for Highway Design with Special Consideration of Operating Speeds,
Thesis for Appointment as University Lecturer, Publication of the
Institute for Highway and Railroad Engineering, University of Karlsruhe
(TH), vol. 11, Germany, 1973.

[7] Ruscher, T., Evaluation of the Safety Concept of the Institute for
Highway and Railroad Engineering (ISE) with the Actual Accident
Situation for a Broad Data Basic, Master Thesis, Institute for Highway
and Railroad Engineering, University of Karlsruhe (TH), Germany, 2001.
108 How to make two-lane rural roads safer

[8] Cafiso, S., Experimental Survey of Safety Conditions on Road Stretches


with Alignment Inconsistencies, Proceedings of the 2nd International
Symposium on Highway Design, Transportation Research Board (TRB),
and Road and Transportation Research Association (FGSV), Mainz,
Germany, June 14 June 16, 2000, Conference Proceedings, pp. 377
387.

[9] Beck, A., Analysis and Evaluation of Relationships between Traffic


Safety and Highway Design on Two-Lane Rural Roads, Master Thesis,
Institute for Highway and Railroad Engineering, University of Karlsruhe
(TH), Germany, 1998.

[10] Eberhard, O., Development of an Operating Speed Background for


Roadway Sections with Grades 6 Percent, as well as Analysis and
Evaluation of Selected Road Sections, Based on Three Safety Criteria,
Master Thesis, Institute for Highway and Railroad Engineering,
University of Karlsruhe (TH), Germany, 1997.

[11] Lamm, R., Psarianos, B., Drymalitou, D., and Soilemezoglou, G.


Guidelines for the Design of Highway Facilities, vol. 3: Alignment,
Ministry for Environment, Regional Planning and Public Works, Athens,
Greece, 1995.

[12] Lamm, R., Psarianos, B., & Cafiso, S. Safety Evaluation Process for
Two-Lane Rural Roads A Ten Year Review, Transportation Research
Record, Journal of the Transportation Research Board, No. 1796,
Geometric Design and the Effects on Traffic Operations, Paper No. 02-
2178, USA, 2002, pp. 5159 and CD-ROM of the Proceedings.

[13] Lamm, R., Choueiri, E. M. & Hayward, J. C. Tangent as an Independent


Design Element, Transportation Research Record, vol. 1195, USA,
1988, pp. 123131.

[14] Lamm, R., & Smith, B. L. Curvilinear Alinement: An Important Issue


for a More Consistent and Safer Road Characteristic, Transportation
Research Record, vol. 1445, U.S.A., 1994, pp. 1221, (1993 Best of
Session Award for the session entitled Cross-Section and Alinement
Design Issues, Awarded by Committee A2A02, Geometric Design,
Transportation Research Board, National Academy of Science,
Washington, D.C., USA, 1994.

[15] Lippold, C., To the Relation Design of Two-Lane Rural Roads,


Commemorative Volume to the 60th Birthday of Univ.-Prof. Dr.-Ing.
Walter Durth, Technical University of Darmstadt, Department: Road
Design and Road Operation, Darmstadt, Germany, 1995, pp. 121132.
References 109

[16] Hanke, H., Effects of Alignment Design on Traffic Safety, VSVI


Journal, vol. 1, Wiesbaden, Germany, 1995.

[17] Harwood, D.W., Fambro, D. B., Fishburn, B., Joubert, H., Lamm, R. &
Psarianos, B. International Sight Distance Design Practices,
International Symposium on Highway Geometric Design Practices,
Transportation Research Board, Boston, Massachusetts, U.S.A.,
August 30 September 1, 1995; Conference Proceedings, 1998, pp. 321
to 3223.

[18] Lamm, R., Zumkeller, K., & Beck, A. Traffic Safety The Relative
Effectiveness of a Variety of Road Markings and Traffic Control
Devices, International Conference: Road Safety on Three Continents,
CSIR, TRB, VTI, BASt, Pretoria, South Africa, 20 22 September 2000,
Conference Proceedings, Session 3.

[19] Schmidt, G., Analyses and Evaluation of Roadway Sections with


Respect to Three Safety Criteria, Master Thesis, Institute for Highway
and Railroad Engineering, University of Karlsruhe (TH), Germany, 1995.

[20] Zumkeller, K., Analysis and Evaluation of Possible Relationships


between Traffic Safety, Road Design and Road Equipment with Road
Markings, Traffic Control Devices and Guardrails, Master Thesis,
Institute for Highway and Railroad Engineering, University of Karlsruhe
(TH), Germany, 1998.

[21] Schneider, B., Development of a Superior Safety Module for the


Evaluation of the Danger of Two-Lane Rural Roads in Tune with the
Actual Accident Situation, Master Thesis, Institute for Highway and
Railroad Engineering, University of Karlsruhe (TH), Germany, 1999.

[22] Lamm, R., Beck, A., & Zumkeller, K. Analysis of Relationships


Between Traffic Safety and Highway Design on Rural Roads, Road and
Construction (Strassen- und Tiefbau), vol. 12, Germany, 1999, pp. 612.

[23] Lamm, R., Beck, A. & Zumkeller, K. Analysis and Evaluation of


Interrelationships Between Traffic Safety and Highway Geometric Design
on Two-Lane Rural Roads, Proceedings of the 2nd International
Symposium on Highway Design, Transportation Research Board (TRB),
and Road and Transportation Research Association (FGSV), Mainz,
Germany, June 14 June 16, 2000, Conference Proceedings, pp. 557
570.

[24] Lamm, R., Wolhuter, K., Beck, A. and Ruscher, T. Introduction of a


New Approach to Geometric Design and Road Safety, The 20th Annual
South African Transport Conference, 16 20 July 2001, Pretoria, South
Africa, Abstracts No. 26:11, p. 33 and CD-ROM of the Proceedings.
110 How to make two-lane rural roads safer

[25] Lamm, R., Guenther, A. K., & Choueiri, E. M. Safety Module for
Highway Design, Transportation Research Record, vol. 1512, USA,
1995, pp. 715.

[26] Lamm, R., Mailaender, T., Steffen, H., & Choueiri, E. M. Safety
Evaluation Process for Modern Highway Geometric Design on Two-Lane
Rural Roads, Research Report V for CTI Engineering, Co., Ltd., Tokyo,
Japan; Karlsruhe, Germany, 1993.

[27] Richter, P., Weise, G., Heger, R. & Wagner, T., Driving Behavior and
Psychological Activation of Motorists as Evaluation Criteria for the
Design Quality of Road Traffic Facilities, Research Report of the
German Research Association (DFG), Technical University of Dresden,
Germany, 1997.
Index

A element 1, 2, 11, 14, 17, 23-25,


accident 30, 32, 39-40, 43, 49, 70, 80,
cost rate 6-8, 10-11, 45-46, 95- 90, 103
98, 100-103 speed 1, 10-13, 19-20, 23-25,
rate 2, 6-10, 14, 24, 32, 42, 45- 29, 33, 35, 37, 39-41, 53, 55,
48, 95-103, 105 65-66, 68, 75-78, 87, 90, 103-
risk 8, 10, 14, 25, 32-36, 45, 104
101, 104-105 driving
severity 6, 7, 32, 45, 98, 101, behavior 1, 6, 12, 23-24, 33, 80,
105 103, 105
situation 1, 6, 8, 24, 40, 45-48, direction 59-62, 80, 83, 85-86
58-59, 83, 92, 95, 97, 102, 104- dynamics 1, 19, 59, 103
driver 18
105
run-off-the road 9, 11, 45, 47,
E
95
element sequence 4, 10, 23, 27, 45,
47, 50-54, 59, 61, 63-65, 70, 72-75,
C 85, 88
curve
circular 1, 2, 12, 25, 30, 34, 36, H
41, 50, 52, 63, 83, 84 horizontal alignment 8, 10, 14, 24-
transition 1, 2, 8, 32, 34, 41, 83, 25, 34, 40-41, 49, 60, 83, 104
98, 104
degree of 8 R
Curvature Change Rate (CCR) 1, 2, radius of curve 18, 33, 49, 95, 98-
6, 8, 14, 18, 20, 24, 41, 49-50, 63, 99
71, 83, 95, 98, 99, 103-105 road
section 14, 17-18, 23
D characteristics 6, 26, 70, 104
design two-lane rural road 1, 14-15, 25,
consistency 10, 23, 103 57, 70, 80, 83, 91, 95, 97
112 How to make two-lane rural roads safer

S speed
safety behavior 10, 12, 59, 92
Safety Criterion I 1, 10, 12, 23, consistency 10, 14, 17, 24-25,
29, 40, 47, 51, 56, 57-58, 64, 30, 103
69-70, 73, 78, 80, 88, 90, 91 data 17-18
Safety Criterion II 1, 10, 23-25, operating 1, 2, 6, 10, 12, 14-20,
27, 29-31, 35, 40, 51, 54, 57-58, 23-30, 39, 41, 45, 52, 54-55, 59,
60, 64, 69, 73, 79-81, 88, 91 65-68, 70, 75, 77, 80, 84, 87,
Safety Criterion III 1, 10, 35-38, 89, 90, 92, 103-104
40-41, 43, 51, 57-60, 64, 69, 73, 85th percentile speed 1, 2, 12,14-
79, 80-81, 88, 91, 92 15, 17-18, 20, 23-25, 28, 30, 35,
module 40, 42-43, 48, 59, 60- 37, 52-53, 65, 75-77, 84, 103
62, 70, 80-81, 83, 85-86, 92, superelevation rate 1, 10, 12, 18,
105 35-36, 40-42, 49, 59
evaluation process 12, 17, 19,
25, 28, 37, 39-40, 49, 51, 60, T
62, 64, 70, 73, 77, 80, 83, 88, traffic safety 6, 11, 18, 33, 47-49,
104-105 95, 101-103, 105
side friction
assumed 1, 35-37, 40, 55, 57, V
68, 77-78, 90, 103 vehicle 1, 6-7, 13-14, 17-18, 49, 60,
demanded 1, 35-37, 56, 68, 78, 70, 83
90
Personal Information
Authors

Ruediger Lamm (1937 - 2005)


1963: Diploma in Civil Engineering from the University of Karlsruhe
(Dipl.-Ing.), Germany
1967: Ph.D. in Civil Engineering from the University of Karlsruhe
(Dr.-Ing.), Germany
1973: Habilitation in Road Engineering from the University of
Karlsruhe (Dr.-Ing. habil.), Germany
1963 - 1977: Scientific Assistant, Chief Engineer, Scientific Councilor, and
Professor at the Institute for Highway and Railroad Transporta-
tion and Engineering, University of Karlsruhe, Germany
1978 - 1979: Acting Dean (Sarparast), Engineering Faculty, University of
Gilan, Rasht, Iran
1980: Visiting Professor, The Ohio State University, Columbus,
Ohio, U.S.A.
1983 - 1987: Full Professor, Faculty of Civil and Environmental
Engineering, Clarkson University, Potsdam, NY, U.S.A.
1978 - 2004: University-Professor at the Institute for Highway and Railroad
Engineering, Department of Civil Engineering, Geo- and
Environmental Sciences, University of Karlsruhe, Germany,
officially retired since October 2002; however still active in
teaching and research at the Institute
1988 - 2004: Consultant, Mailaender Ingenieur Consult, Karlsruhe, Germany
2002 - 2004: Consultant for Highway Engineering and Traffic Safety, Buehl,
Germany

R. Lamm authored and co-authored about 180 papers and research reports in the
transportation field. Consulting and elaboration of highway geometric design
guidelines for different countries.

1993: Best of Session Award, Cross-Section and Alinement Design


Issues, sponsored by the Committee on Geometric Design of
114 How to make two-lane rural roads safer

the Transportation Research Board, for a Paper entitled


"Curvilinear Alinement: An Important Issue for a More
Consistent and Safer Road Characteristic", Awarded by
Committee A2A02, Geometric Design, Transportation
Research Board, National Academy of Science
1999: Publication of the Book: "Highway Design and Traffic Safety
Engineering Handbook", McGraw-Hill, Professional Book
Group, New York, N.Y., U.S.A., 1999, 932 pages, ISBN 0-07-
038295-6, Co-Authors: B. Psarianos, T. Mailaender; Associate
Authors: E.M. Choueiri, R. Heger, R. Steyer; Language
Editors: J.C. Hayward, E.M. Choueiri, and J.A. Quay.
2000: Societ Italiana Infrastrutture Viarie, Be it known that
Ruediger Lamm has been duly elected a Honorary Member of
this Association
2001: Honorable Mention, Subject Category: Engineering
Handbooks, presented to McGraw-Hill, for: "Highway Design
and Traffic Safety Engineering Handbook", 1999, Professional
Scholarly Publishing Division, Association of American
Publishers, U.S.A.

Anke Bettina Beck (born in 1974)


2001: Diploma in Civil Engineering from the University of Karlsruhe
(Dipl.-Ing.), Germany
2001 - Present: Research Assistant at the Institute for Highway and Railroad
Engineering, Department of Civil Engineering, Geo- and
Environmental Sciences, University of Karlsruhe, Germany

Thomas Ruscher (born in 1974)


2001: Diploma in Civil Engineering from the University of Karlsruhe
(Dipl.-Ing.), Germany
2001 - 2003: Project Engineer at Weissenrieder Consult, Offenburg,
Germany
2003 2004: Projekt Engineer at Wald & Corbe Consult, Huegels-
heim/Baden-Baden, Germany

Theodor Mailaender (born in 1949)


1976: Diploma in Civil Engineering from the Technical College of
Darmstadt (Dipl.-Ing.), Germany
1976 - 1978: Project Engineer at KWU-Siemens, Erlangen, Germany
1978 - 1986: Railroad Engineer, Karlsruhe, Stuttgart, Germany
1986 - Present: President of Mailaender Ingenieur Consult, Karlsruhe,
Germany
Personal information 115

Co-Authors

Salvatore Cafiso (born in 1962)


1987 Degree in Civil Engineering from Catania University, Italy
1987 - 1991 Research Assistant at the Institute of Road, Railway and
Airport, University of Catania, Italy
1991 - 2000 Assistant Professor in Road, Railway and Airport Engineering,
University of Catania, Italy
1991 - Present Member of Italian Society of Transportation Infrastructures
(SIIV)
2000 - Present Coordinator of the Ph.D. Course in Transportation
Engineering, Catania University, Italy
2000 - Present Delegate of the Transport Infrastructures Section of the Depart-
ment of Civil and Environmental Engineering, Catania
University, Italy
2001 - Present Associated Professor at the Department of Civil and
Environmental Engineering, Catania University, Italy

Grazia La Cava (born in 1970)


1998 Degree in Civil Engineering from Catania University, Italy
2002 Ph.D. in Transportation Engineering from Catania University,
Italy
2002 Member of Italian Society of Transportation Infrastructures
(SIIV)
2002 - Present Research Assistant at the Department of Civil and
Environmental Engineering, University of Catania, Italy

Coworkers

Torsten Beck (born in 1967)


1994: Diploma in Civil Engineering from the University of Karlsruhe
(Dipl.-Ing.), Germany
1994 - 1996: Project Engineer at Ingenieurbro fr Verkehrstechnik IVT
GmbH, Karlsruhe
1996 - 2002: Scientific Assistant at the Institute of Urban and Regional
Planning, University of Karlsruhe, Germany
2002 - Present: Owner of engineering consultant beck-consult.de,
Berghausen, Germany

Ralf Heger (born in 1966)


1992 - 1999: Cooperation with Consulting Engineers Lindorf-Reichelt-
Streiber GbR, Dresden, Germany
116 How to make two-lane rural roads safer

1993: Diploma in Civil Engineering from the Dresden University of


Technology (Dipl.-Ing.), Germany
1993: Research Scholarship, Texas A&M University, College
Station, Texas, U.S.A.
1994 - 1999: Scientific Assistant at the Institute of Traffic Route
Construction, Dresden University of Technology, Germany
1999 - 2002: Executive Manager of the Consulting Engineers Lindorf-
Streiber-Heger GbR, Dresden, Germany
2001 - Present: Scientific Consultant for the Institute of Traffic Psychology,
Institute for Traffic Planning and Road Traffic, Dresden
University of Technology, Germany

Basil Psarianos (born in 1954)


1976: Diploma in Rural and Surveying Engineering from the
National Technical University of Athens (Dipl.-Ing.), Greece
1981: Ph.D. in Road Design from the University of Hannover
(Dr.-Ing.), Germany
1984 - 1987: Freelance Engineer, Greece
1988 - 1992: Lecturer for Transportation Engineering at the National
Technical University of Athens, Department of Rural and
Surveying Engineering, Greece
1993 - Present: Assistant, Associate and Full Professor for Transportation
Engineering at the National Technical University of Athens,
Department of Rural and Surveying Engineering, Greece

Supported by
John C. Hayward (born in 1947)
1969: Bachelor of Science in Civil Engineering from Ohio
University, Athens, Ohio, U.S.A.
1969 - 1974: Research Assistant, Pennsylvania Transportation Institute, The
Pennsylvania State University, University Park, Pennsylvania,
U.S.A.
1971: Master of Science in Civil Engineering from the Pennsylvania
State University, University Park, Pennsylvania, U.S.A.
1974: Ph.D. in Civil Engineering from the Pennsylvania State
University, University Park, Pennsylvania, U.S.A.
1974 - 1982: Manager of Transportation Planning, Michael Baker
Corporation, Beaver, Pennsylvania, U.S.A.
1982 - 1986: Assistant Vice President - GIS and CADD Services, Michael
Baker Corporation, Beaver, Pennsylvania, U.S.A.
1986 - 1995: Senior Vice President Transportation Engineering, Michael
Baker Corporation, Pittsburgh, Pennsylvania, U.S.A.
Personal information 117

1995: Advanced Management Program, Harvard Business School,


Harvard University, Cambridge, Massachusetts, U.S.A.
1995 - 2000: Executive Vice President Transportation Business Unit,
Michael Baker Corporation, Pittsburgh, Pennsylvania, U.S.A.
2000 - Present: Consultant, Principal-Engineering and Management Consulting
Practice, Beaver Falls, Pennsylvania, U.S.A.
2001 - Present: Head, Department of Engineering, PLS Professor of Logistics
Engineering, Robert Morris University, Moon Township,
Pennsylvania, U.S.A.

Max Eugen Rapp (born in 1954)


1982 Diploma in Civil Engineering from the University of
Karlsruhe, Germany
1985 Master of Business Administration from the University of
Lausanne (HEC), Switzerland
1982 - 1984 Ed. Zblin AG, Stuttgart, Germany, Technical employee
ECBM Manantali, Mali, Site Manager
1986 - 1989 Greschbach Industrie GmbH & Co., Herbolzheim, Germany,
Product Manager and Project Supervisor
1989 - 1997 Group VIVENDI / SGE / G+H Montage Group, Ludwigshafen,
Germany, G+H Montage GmbH (Subsidiary of VIVENDI /
SGE), Sales Manager, Department for Structural Engineering,
Ulm (07/89 - 01/91), General Manager, Department for
Structural Engineering, Berlin (02/92 - 06/96), Dwuzet
Fassaden GmbH, Berlin (Subsidiary of VIVENDI / SGE /
G+H): General Manager (02/91 - 03/92). Managing Director
(04/92 - 09/97)
1997 - Present Bureau of Engineering Max Eugen Rapp & Partner, Berlin /
Schramberg, Consultant and Free-Lancing Engineer (PE) for
Architectural and Civil Engineering Services
2002 - Present Instructor for Project Management and Quantity Surverying,
Faculty of Architecture and Civil Engineering, University for
Applied Studies, Karlsruhe, Germany
Member Chambre of Engineering, Baden-Wrttemberg, Germany, listed as
Expert for
- planning tasks
- safety-, and health protection coordination
- evaluation of real estate properties
- structural energy planning and design

Keith Wolhuter (born in 1935)


1959: Graduated B.Sc B.Eng. from University of Stellenbosch, South
Africa
118 How to make two-lane rural roads safer

1960 - 1968: Geometric Design Engineer, Cape Provincial Roads


Department, South Africa
1969 - 1981: Associate and then senior partner, Kantey and Templer
Consulting Engineers, South Africa
1981 - 1988: Research Engineer, Division of Roads and Transport
Technology, CSIR, South Africa
1988 - 1991: Secondment to National Department of Transport as Manager
of National Transportation Research Programs, South Africa
1991: M.Eng (Transportation) from University of Pretoria, South
Africa
1991 - Present: Technology Manager, Division of Roads and Transport
Technology, CSIR, South Africa
He has authored over thirty papers and reports in the
transportation field including various Geometric Design
Manuals for use in South Africa and the countries of the SADC
region.

All Authors, Associate Authors, and Editors are involved, as researchers and
practitioners, in the transportation fields: Highway Geometric Design, Traffic
Safety, Driving Behavior, and Driving Dynamics.
SI* (Modern metric)
conversion factors

Conversion from SI Units Conversion to SI Units


Lengths Lengths
1 cm = 0.3937 in 1 in = 2.54 cm
1m = 3.2808 ft 1 ft = 0.3048 m
1 km = 0.6214 mi 1 mi = 1.6093 km
Areas Areas
2 2
1 cm = 0.1550 in 1 in2 = 6.4516 cm2
1 m2 = 10.7639 ft2 1 ft2 = 0.0929 m2
1 km2 = 0.3861 mi2 1 mi2 = 2.590 km2
Volumes Volumes
1l = 0.2642 gal 1 gal = 3.7854 l
1l = 0.035315 ft3 1 ft3 = 28.3169 l
3 3 3
1m = 35.3133 ft 1 ft = 0.02832 m3
Velocities Velocities
1 km/h = 0.6214 mi/h 1 mi/h = 1.6093 km/h
Mass Mass
1 kg = 2.2046 lb 1 lb = 0.4536 kg
Force Force
1N = 0.2248 lb 1 lb = 4.4482 N
Pressure or Stress Pressure or Stress
1 N/m2 = 0.02088 lb/ft2
1 lb/ft2 = 47.880 N/m2
* SI is the Symbol for the International System of Measurement
Intelligent Road Design Urban Transport XII
M. K. JHA, Morgan State University, USA,
P. M. SCHONFELD, University of Urban Transport and the Environ-
Maryland, USA, J-C. JONG, Sinotech ment in the 21st Century
Engineering Consultants, Inc., Taiwan,
Edited by: C. A. BREBBIA, Wessex
R.O.C., E. KIM, The Korea Transport
Institute of Technology, UK,
Institute, South Korea
V. DOLEZEL, Tu Pardubice, Czech
Traditional textbooks on roadway designs Republic
focus on fundamental alignment elements
Transportation in cities, with its related
and design criteria. When applying these in
environmental and social concerns continues
practice, engineers must try various design
to be a topic of the utmost priority for urban
combinations, check whether the resulting
authorities and central governments around
alignments satisfy requirements, and
the world. This is reflected in the
evaluate their relative effectiveness.
proceedings of the Twelfth International
Introducing a systematic and efficient
Conference on Urban Transport and the
approach to optimize alignments, this
Environment in the 21st Century, stressing
essentially practical text emphasizes the use
the continuous steady growth and research
of artificial intelligence (AI) and Geographic
into the urban transport systems control
Information Systems (GIS) in extensively
aspects, information and simulation systems.
automated highway design. Based on a
Papers cover topics such as: Transport
series of research projects, it provides a
Logistics and Operations Research;
thorough introduction to the mathematical
Transport Modelling and Simulation;
models and solution algorithms for
Intelligent Transport Systems; Urban
optimizing highway alignments, including
Yransport Planning and Management; Road
horizontal, vertical, and three-dimensional
and Parking Pricing; Public Transport
alignments.
Systems; Environmental and Ecological
The text is ideally suited to senior
Considerations; Transport Sustainability;
undergraduate or graduate students
Infrastructure and Maintenance;
majoring in civil engineering or
Information Systems and GPS Applications;
transportation management. Practicing
Transport Security and Safety; Transport
highway design and transportation
Technology; Energy and Transport Fuels;
engineers will also find it of interest.
Land Use and Transport Integration.
Series: Advances in Transport Vol 19
WIT Transactions on The Built
ISBN: 1-84564-003-9 2005 448pp
Environment, Volume 89
147.00/US$264.00/220.50
ISBN: 1-84564-179-5 2006 960pp
290.00/US$530.00/435.50

We are now able to supply you with details of new


WIT Press titles via
E-Mail. To subscribe to this free service, or for
information on any of our titles, please contact
the Marketing Department, WIT Press, Ashurst Find us at
Lodge, Ashurst, Southampton, SO40 7AA, UK http://www.witpress.com
Tel: +44 (0) 238 029 3223
Fax: +44 (0) 238 029 2853 Save 10% when you order from our encrypted
E-mail: marketing@witpress.com ordering service on the web using your credit card.
Fuel Efficient Car Engineering; Electromagnetic Compatibility
and Lightning; Reliability, Availability,
Technology Maintainability and Safety (RAMS); Freight;
Advanced Train Control; Train Location;
M.L. POULTON, Automotive Consultant
CCTV/Communications; Operations Quality;
Presents measures designed to reduce fuel Timetables; Traffic Control; Global
consumption in passenger cars. Navigation using Satellite Systems; Online
Scheduling and Dispatching; Dynamics and
ISBN: 1-85312-447-8 1997 160pp
Wheel/Rail Interface; Power Supply; Traction
59.00/US$91.00/88.50
and Maglev; Obstacle Detection and
Collision Analysis; Railway Security.
WIT Transactions on The Built
Environment, Volume 88
Computers in Railways X ISBN: 1-84564-177-9 2006 1008pp
330.00/US$550.00/495.00
Computer System Design and
Operation in the Railway and
Other Transit Systems X
Edited by: J. ALLAN, Rail Safety and
Standards Board, UK, C. A. BREBBIA,
Wessex Institute of Technology, UK, Advances in City
A. F. RUMSEY, Parsons Transportation
Group, USA, G. SCIUTTO, Universita Transport
degli Studi di Genova, Italy, S. SONE,
University of Kogakuin, Japan,
Case Studies
C.J. GOODMAN, The University of Edited by: S. BASBAS, Aristotle University
Birmingham, UK of Thessaloniki, Greece

This book updates the use of computer-based Highlighting the highly topical subject of
techniques, promoting their general transport and the environment and the
awareness throughout the business closely related field of town planning, this
management, design, manufacture and book contains chapters concerning
operation of railways and other advanced developments in the transportation systems
passenger, freight and transit systems. of various cities all over the world. These
Including papers from the Tenth include Singapore, Sao Paulo, Santiago,
International Conference on Computer Bilbao, Eindhoven, Adelaide, Bangalore and
System Design and Operation in the Railway Thessaloniki.
and Other Transit Systems, the book will be The studies featured will be of interest to
of interest to railway management, postgraduate researchers in transport and the
consultants, railway engineers (including environment, engineers and planners
signal and control engineers), designers of working within transport and environment
advanced train control systems and computer ministries and local authorities, and
specialists. Themes of interest include: consultants.
Planning; Human Factors; Computer Series: Advances in Transport Vol 17
Techniques, Management and languages; ISBN: 1-85312-799-X 2005 208pp
Decision Support Systems; Systems 66.00/US$120.00/99.00
Urban Transport XI Innovations in Freight
Urban Transport and the Environ- Transport
Editors: E. TANIGUCHI, Kyoto University,
ment in the 21st Century Japan and R.G. THOMPSON, University
Edited by: C. A. BREBBIA, Wessex of Melbourne, Australia
Institute of Technology, UK,
L. C. WADHWA, James Cook University, Highlighting new ideas and best practice,
Australia this book examines innovations in modern
freight transport systems.
The continuing need for better urban Partial Contents: Intelligent Transport
transport systems and a healthier Systems; Vehicle Routing and Scheduling;
environment has led to an increased level of Logistics Terminals; Intermodal Freight
research around the world. This is reflected Transport; Underground Freight Transport
in Urban Transport XI, which features the Systems; E-Commerce and the
proceedings of the latest conference in this Consequences for Freight Transport; Future
well-established series. The subjects covered Perspectives.
are of primary importance for analysing the Series: Advances in Transport, Vol 11
complex interaction of the urban transport ISBN: 1-85312-894-5 2002 216pp
environment and for establishing action 76.00/US$118.00/114.00
strategies for transport and traffic problems.
Over 85 papers are included and these
highlight topics within the following areas:
Urban Transport Systems; Public Transport
Systems; Infrastructure and Maintenance;
Hybrid Vehicle Propulsion
Safety and Security; Transport C.M. JEFFERSON, University of the West
Sustainability; Accessibility and Mobility; of England, UK and
Environmental Impacts; Air and Noise R.H. BARNARD, University of Hertford-
Pollution; Energy and Fuel; Integrated Land shire, UK
Use and Transport; Travel Demand In this book, the authors review recent
Management; Traffic Control and progress in the development of a range of
Integration; Advanced Transport Systems; hybrid vehicles and describe the results of
Simulation; Economic and Social Impacts; field trials and operational experience.
and Cost and Investment Analysis. Numerous tables, graphs and photographs
WIT Transactions on The Built are included together with clear references.
Environment, Volume 77 The volume will be of great interest to
ISBN: 1-84564-008-X 2005 928pp engineering and technical staff working in
297.00/US$475.00/445.50 the road and rail vehicle industries, and final
year undergraduates and postgraduates
studying mechanical and automotive
engineering.
All prices correct at time of going to press but
subject to change.
Series: Advances in Transport, Vol 10
WIT Press books are available through your ISBN: 1-85312-887-2 2002 176pp
bookseller or direct from the publisher. 69.00/US$107.00/103.50