WCTR-Jul - 10 Framework For Evaluating The Economic Impact of Traffic Signals

Framework for evaluating economic impact of traffic signals
Chandra, Ashish; Firth, Keith; and Carrignon, David
FRAMEWORK FOR EVALUATING ECONOMIC THE IMPACT OF TRAFFIC SIGNALS

Ashish Chandra, Keith Firth, and David Carrignon, Colin Buchanan and Partners, London
ABSTRACT
In recent years there has been a sustained debate on the role of traffic signals in major cities. There is growing concern that there are too many traffic signals in London and that they cause unnecessary delays, particularly outside peak hours. Typically, aside from road safety, the form of junction control, and thus the use of traffic signals, is based on an evaluation of worst case, peak traffic flow conditions alone, without consideration to the impact over the complete day, week or year. This approach, although comprehensive in minimising queues and delays (and thus road user costs) during peak conditions, fails to differentiate the operational requirements and benefits during other times of the day. The present paper attempts to inform the debate on the use of traffic signal control and how alternative methods of control can be evaluated and where and when they might be implemented. It explores the options of introducing part-time control in the form of flashing amber signals, used widely in other European countries, as well as complete removal of traffic control regulations akin to shared space.
INTRODUCTION
There are many studies looking at the positive impact of ITS and traffic signals on traffic congestion. Despite the evidence of benefits shown in most studies, road users are frequently being complaining, and in recent years there has been a sustained debate within local authorities on the role of traffic signals. In London, the number of traffic signal installed on street has increased by 1,000 since the year 2000, with the total now standing at over 5,000 sites (including pedestrian crossings). These are equally distributed between Inner and Outer-London, with two thirds on non-strategic roads. Stand-alone signal controlled pedestrian crossings make up half of the total sites. There is a growing concern that there are too many traffic signals in London and that they cause unnecessary delays, particularly outside peak hours. Typically, aside from road safety, the type of junction control and thus the use of traffic signals is based on an engineering evaluation of the worst case traffic scenario, peak traffic flow conditions, with little consideration to the impact over the complete day, week or year. This approach, although comprehensive in minimising queues and delays (and thus road user costs) during peak conditions, fails to differentiate the operational requirements and benefits during other times of the day.
12 WCTR, July 11-15, 2010 Lisbon, Portugal
th

This paper is based on a detailed economic impact study conducted on behalf of Greater London Authority to understand the economic benefits of operating traffic signals throughout the day and night and to identify what-if scenarios, that seek to maximise operational benefits. The study included an evaluation of alternative strategies for removing signal control entirely or switching off traffic signals during parts of the day, exploring the feasibility and quantifying the cost and benefits of these strategies. This was achieved by developing micro-simulation models that replicated, as far as possible, conditions at a representative sample of junctions across London under these scenarios, and which represented various forms of geometric arrangement and complexity. Because traffic signals could be justified for several reasons, an assessment framework was developed. The proposed assessment framework is based on key selection criteria, as set out in, inter alia, Transport in The Urban Environment (Institution of Highways and Transport 1997), which states that traffic signals are used to control traffic movement through improved road safety, reduce congestion (when compared to alternative conventional solutions) and/or provide network management strategies to regulate the use of the road network for particular users (ie., bus priority).This framework enables the identification of the purpose(s) for each traffic signal control at each junction. It considers alternative control strategies like full time signal control, part-time signal control, or the complete signal switch-off, enabling to identify the optimum operating conditions for particular times of the day. This is then assessed against other forms of junction control that best replicate likely or specified road user behaviour when traffic signals are disabled for certain periods of the day. For instance, traffic might conform to regulations more commonly occuring at mini-roundabouts in the UK. The study commissioned by GLA and this paper attempt to inform the debate on the use of traffic signal control and how alternative methods of control can be evaluated and where and when they might be implemented. It explores the options of introducing part-time control in the form of flashing amber signals, used widely in other European countries, as well as complete removal of traffic control regulations akin to shared space. This study builds upon an initial exploratory study, using actual traffic flows at signalised junctions in London during different time periods. Junctions were evaluated using an assessment framework to assess the requirement for traffic signals and to define the considerations required to determine suitable alternative methods of control in place of existing traffic signals. Further analysis was undertaken using actual traffic flows at signalised junctions in London during different times of the day. Junctions were evaluated using an assessment framework to assess the requirement for traffic signals and to define the considerations required to determine suitable alternative methods of control in place of existing traffic signals. In addition, to evaluate the economic impact of traffic signals at these junctions, traffic models of these junctions were then tested for two scenarios, that is, with and without traffic signals and the results were compared to determine the net economic costs and benefits of traffic signals by location and time of day using TfL approved traffic simulation models.
th

The following sections describe the background of this study, the present the assessment framework used for evaluating junctions, the key assessment criteria, alternative methods of control, some example of junction assessment, the economic impact assessment and the conclusion. Although the main focus of the study was to assess economic benefits and disbenefits for road users under various control regimes, the economic results will have to be presented as part of the a separate paper. As such, this paper focuses on understanding the influencing factors that need to be considered will evaluating the suitability of a control regime at a traffic junction.
BACKGROUND ISSUE
Transport networks are not static; they rise and fall as time goes by and technology develops. The network evolution has been detailed in research by the Grbler, 1990 showing that various transportation networks follow each other like canals, rail, the surfaced road network, aviation and the high speed trains. Each network follows an S shape curve leading to the life cycle metaphor encompassing three main phases: birth, growth, and maturity (D. Levinson, Handbook of Transport strategy, Policy and Institutions). If this evolution cycle is true for the network as a whole, it is also likely to be true for the technology on the network itself. In recent years there has been a sustained debate on the role of traffic signals in major cities. In London, the number of traffic signal installations has increased by 1,000 since the year 2000, with the total is now standing at over 5,000 sites (including pedestrian crossings). This rapid growth of the traffic signal infrastructure corresponds to the maturity stage of development. The Urban Traffic Control (UTC) or Area Traffic Control (ATC) as it was called in the 1970s gave very good results to improve traffic conditions in congested urban networks. For this reason, ITS was identified from the very early days as a key technology to tackle traffic congestion, especially as the option of building new roads became increasingly difficult. Significant investments were done to develop ITS technology, leading in the UK to dynamic traffic control systems. The first Split, Cycle & Offset Optimisation Technique (SCOOT) was trialled in 1979 in the UK (The Story of How Traffic Management Grew up, TEC, Jan 2009). This use of ITS to improve traffic conditions however does not change the initial analysis made by Rees Jeffreys in 1949 for whom traffic lights for road safety purpose (p155) is: a restrictive measure, hampering traffic flow and causing a wastage of time and money. Traffic lights are improving traffic conditions in congested urban environment, but as a safety equipment, it delays roads users. This apparent contradiction leads to confusion as to the purpose of some equipment as the infrastructure moves on. The maturity stage also very often corresponds to a devolution process, the transfer of powers away from the Central Government to more local bodies of Government. In the UK, this took place with the 2004 Traffic Management Act. Part 2 of the Act is the Network Management Duty and states: "It is the duty of a local traffic authority to manage their road network with a view to achieving, so far as may be reasonably practicable having regard to their other obligations, polices and objectives, the following objectives: (a) securing the expeditious movement of traffic on the authority's road network; and
th

(b) facilitating the expeditious movement of traffic on road networks for which another authority is the traffic authority. The State therefore transfers its infrastructure and its duty to local authorities. The emphasis clearly to reduce the economic impact of congestion, but there is no mechanism or framework to assess and the infrastructure handed over. The network could have moved on and traffic flow redirected elsewhere in the case of a bypass, leaving the junction setup for addressing a congestion issue serving a safety issue only. Other method of control or design could therefore be envisaged.
ASSESSMENT FRAMEWORK
Overview
The present study follows on from our initial exploratory study which used a theoretical VISSIM model to test whether there is a theoretical threshold in terms of the level of traffic at which it is beneficial in economic terms to switch off the traffic signals. This initial study concluded that the economic benefits and disbenefits of traffic signals are heavily dependent on factors including traffic composition, vehicle occupancy, pedestrian volumes and time of day. The study also highlighted that any assessment of traffic signals should take into account a wider spectrum of influencing factors including safety and network management issues. It was proposed that before calculating the economic benefits or disbenefits of signalised traffic junctions, it is important to establish the utility of traffic signals and suitability of using an alternate method of control. An assessment framework was developed to assess the requirement for traffic lights and defines the considerations required to determine suitable alternative methods of control in place of existing traffic signals.
th

Figure 1: Methodology for assessment of traffic signals
th

KEY ASSESSMENT CRITERIA

It is noted that most of the junction appraisals which lead to the installation of traffic signals evaluate peak traffic flow conditions and generalise the use of signals over the complete day and year. This approach, although comprehensive in evaluating impact of traffic signals on safety and traffic flow in general, fails to differentiate the operational requirements and benefits during other times of the day. The time of the day parameter has been taken into account in the following analysis. According to Transport in The Urban Environment (IHT 1997), the traffic signals are used to control traffic movement through: 1. improved road safety; 2. major reductions in congestion; 3. specific strategies which regulate the use of the road network. The following sections set out guidelines based on the above considerations under key heads and also proposes alternative methods of control in the case where traffic signals are switched off for complete or partial time periods in the day. A brief outline of the assessment methodology is shown above in Figure 1
Individual site road safety assessment

It is proposed that any road safety analysis should be site and time of day specific. Consideration of the safety issues which could arise and which should be considered prior to any formal decision-making, are set out in the following sections of this report.
The Ratio Of Pedestrian And Cyclists To Vehicular Traffic

The proportion of vehicular to pedestrian / cyclist numbers may influence the decisions regarding safety when considering switching off traffic signals. The cost benefits of switching off the traffic signals, in terms of delay improvements, may be greater than the loss of the cost benefit of accident savings where the proportion of pedestrians or cyclists is very low relative to the proportion of vehicular traffic. Where the proportion of pedestrians or cyclists is relatively high (such as near schools or other obvious pedestrian desire lines), there are likely to be many crossing movements and therefore a high level of vehicle / pedestrian conflicts. The cost benefit of providing signal control in terms of a possible reduction in the number of potential vehicle / pedestrian accidents arising from this conflict may be greater than any delay savings resulting from switching off the signals. This would not be the case in situations where the overall traffic volume is very high. The number of potential gaps which could be used by pedestrians to cross safely are likely to be
th

far fewer in high traffic volume situations. Clearly, conditions would vary with time of day, and so an optimum traffic / pedestrian / cyclist volume and ratio would need to be identified. The time at which it may be most appropriate to switch off traffic signals would have to be decided on a site by site basis.
Carriageway Widths
The overall width of the carriageway, and number of traffic lanes, should be a consideration in deciding whether traffic signals may be switched off, as the greater the distance that a pedestrian has to travel in order to cross the road, the greater the potential exposure to risk.
Junction Layout And Geometry

The layout of the junction (excluding roundabout junctions) in terms of inter-visibility, the number of arms and permitted movements, as well as whether or not there is any obvious priority, may all be factors in influencing safety at a junction during periods of traffic signal switch-off. This can be outlined as follows: Inter-visibility; Where unobstructed junction intervisibility (and visibility splays which comply with those required for the type of junction which will be created if traffic signals are switched off) can be achieved, such junctions should score well, as any reduction in safety is less likely to be serious than at junctions where there are obstructions to visibility. Although TD 50/04 recommends unobstructed junction intervisibility, Relaxations or Departures from Standard sometimes result in the provision of signal controlled junctions which have poor junction inter-visibility. Visibility splays at priority junctions; Visibility splays at priority junctions are measured with two values known as distance x and distance y. The x distance is measured along the centreline of the major carriageway and should be 9 m (minimum of 2.4 m in difficult circumstances when traffic flows are low). From this point an approaching driver should be able to see clearly points to the left and right on the major road at a distance of 70m (distance y) for the design speed of major road of 50 kph, measured from its intersection with the centreline. The measurement of the x and y distances is shown in Figure Error! No text of specified style in document.2. DMSSD; Desirable Minimum Stopping Sight Distance (DMSSD) refers to drivers approaching the junction on the major and minor road and is chosen according to the design speed of this road. The DMSSD for the design speed of 50kph is 70m with one step below Desirable Minimum being 50m. Forward visibility at entry to roundabout; Drivers of vehicles approaching the give way line must be able to see objects of height between 0.26m and 2m on the full width of the circulatory carriageway for the Visibility Distance as specified in TD 16/ 93 Geometric design of Roundabouts.
th

Figure Error! No text of specified style in document.2: Visibility standards with a curved major road (TD 42/95 Geometric Design of Major / Minor Priority Junctions)
Visibility to the right at roundabouts; Drivers of all vehicles approaching the give way line must be able to see the full width of the circulatory carriageway to their right for the Visibility Distance as specified in TD16/93 Geometric design of Roundabouts. It also needs to be checked from the centre of the offside lane at a distance of 15m back from the give way line. Number of arms; The more arms a junction has, and the higher the number of permitted turning movements, the higher the number of potential conflict points, and the higher the number of directions a crossing pedestrian may need to check before making the decision as to whether or not it is safe to cross. Permitted movements; Where some movements may be banned, the number of potential conflict points and directions of approaching traffic will be reduced, which could be safer than situations where all traffic movements are permitted. However, the signing and markings at such junctions will need to be carefully reviewed, as drivers may (incorrectly) assume that movements which are not permitted during the hours of signal operation are permitted when the signals are switched off or operate differently than during peak hours. This could result in a serious reduction in safety if unexpected manoeuvres are undertaken at certain time of day. Obvious priority; At junctions where there is a clear priority (such as roundabouts or three-armed junctions at which a minor road joins a major road at right angles in a clear T form), the ease with which drivers adapt to the new situation is likely to be greater, and so may have less of a safety implication. However, where priority is not clear (such as Y junctions, or crossroads junctions where the flow on each arm is relatively equal), the potential for conflict and driver / pedestrian indecision will be increased.
th

Pedestrian and cyclist provision

The type and number of crossing facilities provided at a signal controlled junction, as well as the proximity of alternative crossing facilities on site, might influence safety, and should be considered. This can be outlined as follows: Controlled crossing facilities; When controlled crossing facilities are provided as part of the existing signal arrangement, and in the absence of other facilities, the greatest potential reduction in safety for pedestrians could be expected, in the event that signals are switched off or operate differently than in the peak hour. However, it would be expected that the reduction in safety would not be so great in locations where central refuges provide opportunity for pedestrians to cross in parts.. Signalised junctions where pedestrian crossing points are uncontrolled; Removing signal control at such junctions may have a negative impact on safety as the number of gaps created for pedestrians to cross would be reduced. Pedestrian crossing behaviour may, however, be different to that at controlled crossing points as users may be more used to utilising gaps in traffic rather than obeying signal control. Stand alone facilities nearby; The proximity of alternative crossing facilities (such as Toucan or Pelican crossings) would be expected to mitigate the safety impact of a switch-off. ASLs; At junctions where Advanced Stop Lines for cyclists are currently provided, the switching off of traffic signals is likely to result in traffic positioning itself closer to the junction (i.e. within the ASL) in order to maximise visibility. As a result, safety for cyclists at such junctions might be expected to decline during times when signals are switched off.
Characteristics of Traffic
Factors such as driver approach and through speed and the proportion of heavy goods vehicles travelling through the junction are also likely to have an impact on safety during periods when the signals are switched off. This can be outlined as follows: Approach speeds; Whilst it is possible that traffic signal switch-off could result in lower traffic speeds, as drivers may be unsure of priority (which could have a positive effect on safety), it is also possible that drivers who consider themselves on the priority arm or route will increase speed on approach. This will increase the risk to pedestrians trying to judge safe crossing gaps and drivers may be less willing to give way to traffic on other arms. This will, in turn, increase the potential for conflict with traffic on other arms. Through speeds; Traffic signal switch-off may enable left-turning vehicles to approach and turn left at the junction more efficiently. Depending on opposing traffic movements, there could be increased opportunities for left-turning traffic to undertake this movement without stopping. This will give cyclists less warning of a large vehicle
th

planning to turn left and potentially increase this already common collision type at junctions. Goods vehicles; The higher the proportion of such vehicles, the higher the potential for masking of traffic in adjacent lanes, and for an increase in the severity of any eventual collisions. Although possibly a minor factor, it will nevertheless have a negative impact on safety in the event of signal switch-off.
Collision History
Analysis of the collision history at a junction under consideration will be of great benefit in assessing the likely safety implications of switching off traffic signals.
Road network management

The use of traffic signals could, in parts of the road network, be dictated by traffic management imperatives over local congestion or road safety considerations. This study will need to take this specific use into account. The following key assessments would be required: Is the traffic signal part of any strategic network, e.g. TLRN; Is the signal used for enforcing speed or flow metering; or Is the signal part of group of signals, e.g. SCOOT.
Congestion and capacity assessment

The positive or negative impact of traffic signals on congestion will be assessed through: the degree of saturation; the traffic throughput and reserve capacity; vehicular delay; delay to passengers and other street users; scope for further signal timing optimisation; and requirements during different times of the day.
Figure 3 shows the conventional approach to choosing junction type based on the simple relationship between traffic flows on major / minor roads, using average daily traffic demand. This diagram does not compare the type of junction to time of the day, but it gives a good indication of possible alternatives. This part of the Framework will assess whether or not the level of traffic at a junction necessitates traffic signal control or not. As with the road safety section, this assessment will be site and time of day specific.
th
10

Figure 3: Type of Junction appropriate for different traffic flows on major / minor roads
Source: Transport in the Urban Environment
ALTERNATIVE METHODS OF CONTROL

The Framework defines alternate strategies in the event that signal performance is found to be sub-optimal or when the disadvantages due to the traffic signals outweigh the advantages during part of or a complete period. A detailed review of existing measures and provisions within the legislation to provide for alternatives to current junction design has been conducted. These strategies may include: Introduce parttime signal control; Introduce flashing amber signal to traffic during off-peak scenario; Optimising the signal settings for all periods of day, if found to be inappropriate.
Part-time signal control is currently in use in the UK at roundabouts, although the number of sites has been diminishing in recent years due to safety and design concerns. The traffic signals are switched off for most of the day. If the entry arms suffer long queues, the traffic signals are automatically switched-on to regulate conflicts. This method is also considered at standard crossroads and T-junctions, but flashing amber arrangements might be preferred.
th
11

Presenting a flashing amber signal to traffic does not currently constitute a possible alternative within the UK context for the moment, but is supported by the Vienna convention and this method of control is now common place in Europe and across the World. It is very difficult to predict what road-user response and behaviour would be during traffic signal switch-off. The closest condition that we have in the UK so far, save for a very few sites that have no technical evaluation of behaviour, is that which occurs during a signal failure. Attempts to standardise modelling/ forecasting techniques of this condition have, to date, not been particularly successful. This study considers a range of responses and possible methods of simplifying the assessment of these responses. Table 1 below presents the closest approximation to alternative conventional methods of control, or road-user responses, envisaged in the absence of formal traffic signal arrangement with the traditional green, red and amber signal. The following sections describe the plausible regulatory framework for junction control should it be required to discontinue traffic signals completely or disable signal control during different times of the day, week or year. As such, these are based on principles of how traffic is likely to behave, or can be encouraged to behave, at junctions when signals are disabled. These assumptions regarding behaviour are based on information available, engineering judgement and expected limitations in absence of clear evidence. At the time this report was being finalised, a live trial was underway at the Cabstand junction in Portishead, North Somerset to test the impact of disabling traffic signal control at all times. The trial was proving extremely successful in junction capacity terms, and provided an opportunity to witness first-hand the type of behaviour that occurs at junctions when traffic signal control is disabled completely, without any form of vehicular or pedestrian priority markings in place.
Amber Flashing Light at intersections

For this study, three alternative method of control were tested: Switching off traffic signals and not providing an alternative signing system; Introducing flashing amber with priority to the right; Introducing flashing amber with a minor/major road priority.
In accordance with the Convention, single amber flashing light or two amber lights flashing alternately shall mean that drivers may proceed but shall do so with particular care. Flashing amber light may be installed alone or may also be used in place of a three- colour system at times when traffic is light.
th
12

Table 1:
Alternative traffic control regimes
Control Type
Traffic Management Regime Degree of control Vehicle to vehicle control regulations None
Anarchic
Filter-in-turn (FiT)
Off-Side Priority Rule (OSPR)
Major/minor Priority Control
Regulated
Partial Control
None
Regulated
Partial Control
None
Regulated
Partial Control
None
Regulated
Partial Control
None
Filter-in-turn according to order of arrival
Give way to vehicles approaching from the right (as if mini-roundabout)
Minor road traffic gives way to major road traffic (as if conventional major/minor priority junction)
Method of Control
Traffic signal regime Vehicle to pedestrian control regulations
Switched off or Switched off or flashing amber flashing amber
None
Flashing amber with all-red Switched off or Switched off or pedestrian flashing amber flashing amber crossing stage Pedestrians Vehicles must Pedestrians have priority stop at red Filter-in-turn have priority (presumed that signal to allow according to (presumed that crossings revert pedestrians to order of arrival crossings revert to zebra) cross to zebra) No premeditated order, completely random, until pedestrian forces controlled crossing stage
Flashing amber with all-red Switched off or Switched off or pedestrian flashing amber flashing amber crossing stage Vehicles must Pedestrians stop at red have priority signal to allow None (presumed that pedestrians to crossings revert cross to zebra)
Modelled User Behaviour
Driver behaviour
No premeditated order, completely random
Drivers judge priority according to order of arrival
Drivers judge Drivers judge Drivers judge priority whether gap is priority Drivers judge according to sufficient to according to whether gap is order of arrival, accord with offorder of arrival, sufficient to until pedestrian side priority but must give accord with offforces rule, but must priority to side priority rule controlled give priority to pedestrians crossing stage pedestrians Pedestrians assume priority through presence on crossing
Pedestrian behaviour
Pedestrians Pedestrians Pedestrians Pedestrians Pedestrians assume priority assume priority cross in gaps or judge priority cross in gaps or through through force controlled according to force controlled presence on presence on crossing stage order of arrival crossing stage crossing crossing
Pedestrians cross in gaps
Flashing amber with all-red pedestrian crossing stage Vehicles must stop at red signal to allow pedestrians to cross Drivers judge Drivers judge Drivers judge whether gap is whether gap is whether gap is Drivers judge sufficient to sufficient to sufficient to whether gap is accord with accord with offaccord with sufficient to conventional side priority conventional accord with major/minor rule, until major/minor conventional priority rule, pedestrian priority rule, but major/minor until pedestrian forces must give priority rule forces controlled priority to controlled crossing stage pedestrians crossing stage Pedestrians Pedestrians Pedestrians assume priority cross in gaps or Pedestrians cross in gaps or through force controlled cross in gaps force controlled presence on crossing stage crossing stage crossing
Flashing amber with all-red Switched off or Switched off or pedestrian flashing amber flashing amber crossing stage Vehicles must Pedestrians stop at red have priority signal to allow None (presumed that pedestrians to crossings revert cross to zebra)
Note* - the traffic signals switched off will normally require hooding or advance signal failure warning sign (could be VMS). In addition, although priority to pedestrians under flashing amber may already covered by local regulations (e.g. RTRA and Highway Code in UK), traffic signals on flashing amber may require alteration to existing legislation (e.g. Road Traffic Regulation Act in UK).
th
13

Table 2:
Traffic signal colours as per the Vienna Convention on Road Signs and Signals
Colour Green Position At intersection At intersection, level crossing, swing bridge, airport, fire station or ferry terminal At intersection At intersection At intersection At intersection At intersection Above lane Above lane Above lane At level crossing, swing bridge, airport, fire station or ferry terminal Anywhere except intersection At intersection At crossing Meaning Proceed
Type
Shape
Amber Plain Red Red and amber Nonflashing Arrow pointing left Arrow pointing right Arrow pointing upwards Arrow pointing downwards Cross Arrow pointing diagonally downwards Green Green Green Green Red Amber or white Double Red
Stop if possible
Stop Signal is about to change Only traffic turning left may proceed Only traffic turning right may proceed Only traffic travelling straight ahead may proceed Traffic may continue in lane Traffic may not enter lane Lane closes shortly ahead, change lane Stop
Flashing
Plain
Amber Amber Lunar white
Proceed with caution The priority is determined by Main Road or Yield signs Proceed
The application of these methods to junctions in the UK, along with the behaviour that might be expected or which has been assumed, for simplicity, to occur under flashing amber is described below. This describes how road user behaviour has then been assessed within the micro-simulation environment.
th
14

Switch off, no traffic signal Description

In this scenario traffic signals would be temporarily switched off during specific time periods. At present, it is assumed that this could generally only be carried out at locations where parttime signal control might be expected and priority is immediately obvious, namely at roundabouts. Drivers are assumed to give priority to traffic approaching from the right. Without any evidence of whether or not part-time operation might be a solution for other forms of junction, it is not considered appropriate to apply this method of control to junction arrangements other than roundabouts for the purposes of this study. It is, however, a condition that is relatively commonplace when traffic signal faults occur.
Advantages
Disabling traffic signals at a roundabout would be expected to lead to drivers reverting to the form of behaviour normally expected at non-signalised roundabouts, namely giving priority to traffic approaching from the right. In this case there is unlikely to be any confusion regarding vehicular priority.
Disadvantages
If signals were switched off only during certain times of the day, this could lead to uncertainty between both drivers and pedestrians as to whether this was intentional or due to a mechanical fault. This problem would be resolved with the introduction of appropriate signage informing all road users of the temporary switch off and use of part-time signals, as is presently adopted in the UK. If pedestrians were provided with controlled pedestrian facilities at a signal controlled roundabout, the switch-off will impact on their amenity and might affect their safety, by forcing them to look for gaps in oncoming traffic rather than cross the carriageway during a pedestrian stage. This assumes that pedestrians are expected to defer to vehicles and that vehicle drivers do not alter their behaviour to provide priority, which would be possible with further advance signing instructing drivers to yield to pedestrians, or indeed re-education. The disabling of traffic signals removes the ability to manage fluctuations in demand, and capacity problems on one or more of the approaches might be encountered when there is a predominant flow of traffic in one direction.
th
15

Flashing amber, priority to the right Description

In this scenario, normal operation of traffic signals would be replaced at certain times of the day with the introduction of flashing amber to all approaches. Drivers are assumed to give priority to vehicles approaching the junction from the right. This method would be feasible at all forms of junction, but is assumed to apply only to crossroad junctions for the purposes of this study. If traffic demand is broadly balanced across all arms, then no single approach or movement is likely to force priority for itself and so behaviour would be similar to that found at all-way stop junctions in the US. If, however, there is a predominant movement on a main road through a junction, behaviour is likely to revert to a conventional major/minor priority arrangement. If there is a predominant right-turn movement (for example), behaviour is likely to revert to a conventional offside priority arrangement. This is likely, however, to be affected by the geometry of the junction and the number of approach lanes, and a junction that is clearly designed to have a major/minor road arrangement is then likely to influence traffic behaviour towards major/minor road priority behaviour.
Advantages
Introduction of flashing amber together with appropriate signage, akin to that found on the approach to mini-roundabouts in the UK, would give both drivers and pedestrians a good indication of the altered method of control at the junction, encouraging them to take appropriate care on the approach.
Disadvantages
Since priority is not immediately clear at a crossroad junction, an introduction of flashing amber could lead to some confusion and possible conflict between different streams of traffic. As well as signage, it would be beneficial to re-educate road users on appropriate behaviour at such locations, but this would be costly. The lack of specific signal control removes the ability to manage fluctuations in demand, and capacity problems on one or more of the approaches might be encountered when there is a predominant flow of traffic in one direction. This would be solved by introducing detection equipment that would switch the signals back on, yet this could not be used on a minute by minute basis. Pedestrian amenity will be affected and their safety could be affected if they were no longer provided with controlled facilities and there were insufficient gaps in traffic to cross the carriageway, assuming again that pedestrians would be expected to defer to vehicles. To avoid this, a pedestrian actuated crossing stage would be incorporated into the plan, yet it
th
16

would mean that these facilities would also be required under normal traffic signal operating hours. Thus, formal crossing facilities would need to be introduced to all junctions where flashing amber was to be adopted.
Flashing amber, major/minor priority Description

In this scenario, normal operation of traffic signals would be replaced at certain times of the day with the introduction of flashing amber. Drivers are assumed to apply priority rules of a major/minor priority junction. This method of control would be appropriate for T-junctions.
Advantages
An introduction of flashing amber in conjunction with appropriate signage , in the form of a standard UK advance warning sign for T junction and priority route, would make drivers aware of the altered method of control at the junction, encouraging them to take appropriate care on the approach.
Disadvantages
In a similar manner to crossroads, as well as signage it would be beneficial to re-educate road users on appropriate behaviour at such locations, but this would be costly. The lack of specific signal control removes the ability to manage fluctuations in demand, and capacity problems on one or more of the approaches might be encountered when there is a predominant flow of traffic in one direction. This would be solved by introducing detection equipment that would switch the signals back on, yet this could not be used on a minute by minute basis. Pedestrian amenity will be affected and their safety could be affected if they were no longer provided with controlled facilities and there were insufficient gaps in traffic to cross the carriageway, assuming again that pedestrians would be expected to defer to vehicles. To avoid this, a pedestrian actuated crossing stage would be incorporated into the plan, yet it would mean that these facilities would also be required under normal traffic signal operating hours. Thus, formal crossing facilities would need to be introduced to all junctions where flashing amber was to be adopted.
Modelling assumptions
Some of these alternative methods of control are not generally in use in the UK. In order to model the behaviour of vehicular and pedestrian traffic at a junction where signal control is to be disabled, it is necessary to make assumptions that cannot be verified with robust evidence. However, in absence of any before-and-after data from appropriate sites in the UK, modelling assumptions have been made for the present study based on generalised traffic behaviour.
th
17

In addition, it was assumed that if the traffic signals were to be switched off (or flashing amber was introduced) for all or part of the day, all road users would be informed of the alternative control regime before it was introduced. As such, the study assumes drivers to be informed about the new regulations and expected behaviour through regulating traffic signs, public information campaigns, re-training and through gradual learning and word of mouth. Further assumptions related to road safety and their plausible impacts have been considered seperately below.
JUNCTION ASSESSMENT
For the purpose of present study, to provide potential economic benefit/disbenefit results for Greater London, key junctions were selected which are representative of various characteristic mix of junctions present in London. These selected junctions were further analysed and assessed with the help of microsimulation models to produce the generalised economic analysis. Table 3 shows the classification and distribution of signalised junctions in Greater London by: number of arms (3/4 or more); location (inner/outer London); network (TLRN/non-TLRN). Table 3: Greater London junctions per number of arms, location & Network Number of junctions 538 387 638 220 215 176 286 72 2,532 % of the total number of junctions 21.2% 15.3% 25.2% 8.7% 8.5% 7.0% 11.3% 2.8% 100.0%
Junction type 3 arms Inner London non-TLRN 3 arms Inner London TLRN 3 arms Outer London non-TLRN 3 arms Outer London TLRN 4 and more arms Inner London non-TLRN 4 and more arms Inner London TLRN 4 and more arms Outer London non-TLRN 4 and more arms Outer London TLRN Total number of traffic junctions
Table 4 below shows the list of junctions selected for further evaluations. Due to study limitations, only five sets of junctions were considered and together they represent about 67.5% signalised junctions in London (excluding pedestrian crossings) based on the above criteria only.
th
18

Table 4:
List of selected junctions Peak major arm traffic flow (2-way morning peak) 615 910 884 941 907 952 987 1475 884 640 888 Outer London Outer London Outer London Inner London Inner London
Junction location
Key arms and intersections
Type of junction (no. of arms) 4-arm 4-arm 4-arm 4-arm 4-arm 4-arm 4-arm Roundabout 3-arm (T-junction) 4-arm 3-arm (T-junction)
Location
A5 Edgware Road (interconnected traffic signals)
Edgware Road/ Harrow Road/ Marylebone Road Edgware Road/ Praed Street/ Chapel Street Edgware Road/ Sussex Garden/ Old Marylebone Road Edgware Road/ Burwood Place/ Harrowby Street Edgware Road/ George Street/ Kendal Street Edgware Road/ Connaught Street/ Upper Berkeley Street Edgware Road/ Seymour Street
Church Road A13/ River Road East Barnet West Norwood
(A312/B455)-Target roundabout River Road - Bastable Avenue East Barnet Road/Margaret Road A215 Norwood Road - Palace Road
Example Junction Assessment

Edgware Road - Junction 9 (Praed Street/ Chapel Street) The Edgware Road/Praed Street/ Chapel Street junction is located in the London Borough of Westminster, close to Paddington and Edgware Road stations. The location of the Edgware Road/ Praed Street/ Chapel Street junction is shown in Figure 4. Edgware Road and its surrounding area offers many shops, cafes and restaurants and constitutes an important trip generator which is likely to attract high number of pedestrians and cyclists, especially during the hours of the day.
Traffic mix and characteristics

A classified traffic count survey was carried out at the junction on Wednesday, 4th July 2007 between 07:00- 10:00, 12:00- 14:00 and 16:00- 19:00. The morning and evening results of the traffic count survey are summarised in Table 5 and Table 6.
th
19

Figure 4: Edgware Road/ Praed Street/ Chapel Street junction
This image is the copyright of Google Earth, reproduced under licence to Colin Buchanan and Partners. Further reproduction of this image will constitute a breach of Google Earths copyright protected rights and may result in prosecution
From the classified turning counts presented above it can be seen that cyclists constitute 6.2% of vehicular traffic during the morning, and 4.9% during the evening period. Pedestrian count surveys were not available but the volume of pedestrians during the day and evening is known to be high. The junction is located within a 30mph speed limit zone. No data is currently available on the approach speeds. HGVs are estimated to constitute 6.6% of vehicular traffic during the morning survey period and 2.4% during the evening survey period (see Table 5 and Table 6). This is a relatively low proportion of general traffic and should not cause safety concerns in the event of a switch off.
th
20

Table 5:
Edgware Road/ Praed Street/ Chapel Street morning turning count survey (07:00- 10:00)
Approach Edgware Road(S) Edgware Road(N) Praed Street Chapel Street (one way out) Total
Table 6:
Total No Of Vehicles 3151 3685 1423 0 8259
Pedal Cycles 63 300 147 0 510
HGVs 216 209 117 0 542
% Pedal % Cycles HGVs 2 8.1 10.3 0 6.2 6.9 5.7 8.2 0 6.6
Edgware Road/ Praed Street/ Chapel Street evening running count survey (16:00- 19:00)
Approach Edgware Road(S) Edgware Road(N) Praed Street Chapel Street (one way out) Total Junction layout and geometry
Total No Of Vehicles 3266 3245 1983 0 8494
Pedal Cycles 185 97 130 0 412
HGVs 91 70 46 0 207
% Pedal Cycles 5.7 3 6.6 0 4.9
% HGVs 2.8 2.2 2.3 0 2.4
Compliance to visibility splays and Desired Minimum Stopping Sight Distances (DMSSD) with standards set out in TD 42/95 Geometric Design of Major/ Minor Junctions and TD 9/93 Highway Link Design is presented in Table 7. Table 7: Edgware Road/ Praed Street/ Chapel Street junction- DMSSD and visibility splays
Approach/ Visibility Edgware Road(N) Edgware Road(S) Praed St Chapel St
DMSSD=70m Yes Yes Yes Yes
Y= 70m (X=9m) No No No No
Y=70m (X=4.5m) Yes No Yes Yes
Y=70m (X=2.4m) Yes No Yes Yes
th
21

Table 7 above shows that it would not be possible to achieve visibility splays of 70m with the x distance of 9m on any of the approaches to the junction. The Edgware Road(S) approach could not achieve the visibility splay of 70m even with the x value decreased to 2.4m. On other approaches, the x value would need to be relaxed to 4.5m. The DMSSD of 70m is achievable on all approaches to the junction. It has to be noted that these results are indicative only and should be confirmed with site measurements. The number of lanes and carriageway widths at the Edgware Road/ Praed St/ Chapel Street junction are shown in Table 8. Table 8: Edgware Road/ Praed Street/ Chapel Street junction lanes and carriageway widths
Approach Edgware Road(S)
Lane
Width (m) 3 3 6 3 3.5 3.5 3.5 3.5 10 7
Nearside entry lane (ahead) Offside entry lane (ahead) 2 exit lanes Nearside entry lane (left) Praed St Middle entry lane (ahead) Offside entry lane (right) Nearside entry lane (ahead and left) Edgware Road(N) Offside entry lane (ahead) 3 exit lanes Chapel St 2 exit lanes
Central Refuge No Yes Yes No No Yes No
From the table above it can be seen that the number of lanes does not exceed three on any of the approaches/ exits of the junction. The presence of central refuges facilitates pedestrian movements across the junction. This is a four arm junction with two of the arms (Praed Street and Chapel Street) being one way only. This significantly reduces the number of possible conflicts and would facilitate a carriageway crossing for pedestrians in the event of a switch off. The right-turn from Edgware Road into Chapel Street is banned which further reduces the number of potential conflicts at the junction. The layout of the junction does not indicate clearly which movement would have priority if traffic signals were to be switched off. Judging from the volume of traffic (see Table 5 and Table 6). Edgware Road carries more traffic and should be given priority over Praed Street.
Pedestrian and cyclist provision

There are staggered signal controlled pedestrian crossings provided on the Praed Street and Edgware Road (N) arms of the junction. Switching off signals could decrease pedestrian
th
22

safety at these locations. However, the reduction in safety is not expected to be so great in locations where central refuges are provided. Uncontrolled straight across pedestrian crossings are located on the Chapel Street and Edgware Road (S) arms of the junction. In this case pedestrians are already used to exploiting gaps in traffic rather than obeying signal control. An ASL is provided on the Praed Street approach to the junction. In this case switching off traffic signals could have a negative impact on the safety of cyclists. Drivers would be expected to position themselves closer to the junction (within the ASL) to maximise visibility.
Collision history
A collision history was obtained from TfL for the Edgware Road/ Praed Street/ Chapel Street junction for the period of 36 months to November 2008. A summary of the collision history is shown in Table 9. Table 9: Edgware Road/ Praed Street/ Chapel Street collision history by year and severity
Severity/ Months To Fatal Serious Slight Total
12/11/2006 0 0 5 5
12/11/2007 12/11/2008 Total 0 0 0 0 1 1 7 3 15 7 4 16
The collision summary presented above shows that there were 16 accidents in total during the 36 month study period. Fifteen accidents resulted in slight injury and one in serious injury. No fatalities were recorded at the junction during this time period. The collision rate per year is 5.3 which is above the borough average of 2.32 at signalised junctions. Six of the accidents (37.5%) involved pedestrians which is above the borough average of 33.1% at signalised junctions and 30.1% at signal controlled junctions. None of the accidents involved cyclists. Five accidents (31.3%) occurred during the hours of darkness which is below the borough average of 36% at signalised junctions and slightly above 29.3% at priority junctions. None of the accidents involved right turning vehicles. The spread of the accidents over the three year time period shows an increase during 2007 from five to seven accidents, followed by a drop in 2008 to three accidents.
Road network management

The Edgware Road/ Praed Street/ Chapel Street junction forms part of the TLRN.
th
23

Traffic analysis
The analysis of road safety, network management and traffic at the Edgware Road/ Praed St/ Chapel Street junction is summarised in Figure 5. From the traffic analysis indicator table it can be seen that the overall tendency is against the switch off of traffic signals at this location. The main factors are high number of pedestrians/ cyclists, higher than average number of accidents with high percentage of pedestrian collisions and the fact that this junction forms part of the TLRN and its signal control is linked to other junctions in the corridor. Figure 5: Edgware Road/ Praed St/ Chapel Street junction- road safety, network management and traffic analysis indicator table
POTENTIAL RISK INDICATORS against switch off .for TRAFFIC MIX AND CHARACTERISTICS Volume of traffic Percentage of goods traffic Volume of cyclist movements Pedestrian activity JUNCTION LAYOUT AND GEOMETRY Visibility requirements of TD 42/95 NOT achieved Total number of traffic lanes/ overall carriageway width No. of arms Total number of permitted movements Unclear priority PEDESTRIAN AND CYCLIST PROVISION Central refuges/ islands not provided Controlled crossing facilities provided as part of signal control Absence of stand-alone crossings close to the junction Advanced Stop Lines provided for cyclists? COLLISION HISTORY (latest 36 months) High risk site? Increasing trend in collisions? %age of pedestrian Collisions > average %age of cyclist Collisions > average %age of dark Collisions > average % Right turning accidents > average NETWORK MANAGEMENT TLRN junction Part of corridor management OVERALL TENDENCY HIGH MEDIUM LOW HIGH/ YES PART LOW/NO
Edgware Road/ Praed Street/ Chapel Street
YES SOME/ PART PART NO
YES NO
th
24

ECONOMIC IMPACT ASSESSMENT

The above methodology can be used for economic assessment of traffic lights. For the purpose of present study, Microsimulation models of two comparative scenarios were developed, with and without signals based on the Alternate Methods of Control described above. An economic model was developed, using the outputs from the VISSIM modelling and the economic parameters set out in the DfTs WebTAG guidance for transport appraisals. For each of the five junctions, model outputs were produced for the following time periods: Morning peak (8.00 9.00) Inter-peak (12.00 13.00) Evening peak (17.00 18.00) Off-peak (22.00 01.00)
The following vehicle types are covered: Car, Light goods vehicle (LGV), High goods vehicle (HGV), Bus, Taxi, Motorbike, Bicycle. Each of the modelled junctions was used to analyse two sets of results: the existing or base (Do Minimum) scenario and a Do Something scenario where the traffic signals were replaced with an alternative measure of control as discussed earlier in the report. The following categories of benefit have been quantified and valued: Time savings; Vehicle operating costs (fuel); Vehicle operating costs (non-fuel); and Carbon emissions.
Some of the assumptions that have been used in the economic analysis are described in the next section, and a full assumptions register is provided in the Appendix to this report.
Methodology Time savings

The VISSIM model outputs show the average delay time and the number of vehicles for each vehicle type / time of day / junction, thus enabling total delay time to be calculated. The difference between total delay time in the Do Minimum (existing case with optimised traffic signals) and Do Something (without traffic signal control) shows whether there is a benefit or disbenefit as a result of removing traffic signals in each case.
th
25

The change in delay time can be valued by applying a value of time. Standard values of time per person from the DfTs WebTAG guidance have been used. These are then converted into values of time per vehicle by applying journey purpose splits and average vehicle occupancy rates from WebTAG and the London Area Transport Survey (LATS). The results for each individual time period are then scaled-up so that they represent a total for the whole year. It has been assumed that benefits can be scaled-up in proportion to the observed hourly flows at each junction. This means that if, for instance, the observed flow at a junction between 7.00 and 8.00 is 20% lower than the observed flow between 8.00 and 9.00, then the benefit / disbenefit for 7.00-8.00 is assumed to be 20% lower than the benefit calculated for 8.00-9.00 from the model results. It is not necessarily the case that there is a linear relationship between flow and benefit, although to prove otherwise would require an enormous amount of modelling to be undertaken.
Vehicle operating costs (fuel)

WebTAG guidance provides a formula that can be used to estimate the rate of fuel consumption by vehicles travelling at different speeds. This has been applied to the average speeds in the Do Minimum and Do Something scenarios in order to estimate differences in fuel consumption rates between the two scenarios. This in turn is then applied to the average distance travelled, that is, the distance covered by the area modelled, to calculate changes to total fuel consumption, and WebTAG values for the cost of fuel are applied to estimate the total change to fuel vehicle operating costs. The model results are factored up in the same way as the time savings to obtain an annual total.
Vehicle operating costs (non-fuel)

A very similar approach is used for the non-fuel operating costs again, a WebTAG formula is used to estimate the change to non-fuel vehicle operating costs as a result of different speeds between the two scenarios and the results are scaled-up and annualised accordingly. Emissions Emissions benefits are related to fuel consumption, which is estimated as part of the vehicle operating costs. WebTAG values for carbon emissions per litre of fuel consumed are applied to calculate total emissions, and then monetised also using WebTAG values.
Results
Individual junctions The charts in this section are based on the results for the individual junctions, ie they do not represent a total benefit / disbenefit for all London. Figures 5.1 to 5.5 below show the results for each of the five junctions that were modelled, split by benefit type and time period.1
1
It should be noted that the scale of the y-axis is different for each chart, as the benefits / disbenefits for some junctions are much larger than others
th
26

Figure 6: Impact of removing traffic signals on Edgware Road

200,000
100,000
0 Time saving Fuel VOC Non-fuel VOC Emissions Total
per year
-100,000
-200,000
-300,000
-400,000
-500,000 AM peak Off peak Inter peak Total PM peak
The results for Edgware Road show a disbenefit from removing traffic signals in the morning and evening peak periods. In the case of the evening peak the disbenefits are substantial over 400k a year - reflecting a high traffic flow. There is a benefit from removing traffic signals during the inter-peak of over 100k a year, and a slight benefit in the off-peak. Figure 7: Impact of removing traffic signals at Target Roundabout
100,000 0 Time saving -100,000 -200,000 -300,000 -400,000 -500,000 -600,000 -700,000 -800,000 -900,000 AM peak Off peak Inter peak Total PM peak Fuel VOC Non-fuel VOC Emissions Total
per year
th
27

Figure 5.2 indicates that there would be a significant disbenefit (approximately 660k a year) from removing traffic signals during the inter-peak at the Target Roundabout, and to a lesser extent during the evening peak (disbenefit of approximately 190k a year). There would be a slight benefit from removing traffic signals in the off-peak, with a neutral impact in the morning peak since the level of gridlock is such that the signalling system does not influence delay time. Overall the total for the whole day shows a large disbenefit. It should be noted that roundabouts are peculiar examples of traffic junctions where the delay for individual arms and that for overall traffic is highly dependent on the balance of flows and available gaps for major traffic movements. Unbalanced flows, for instance in the inter-peak scenario, can result in higher delay for all traffic but can be easily minimised by introducing demand operated traffic signals. Figure 8: Impact of removing traffic signals, A13 River Road
4,000
2,000
0 Time saving -2,000 Fuel VOC Non-fuel VOC Emissions Total
per year
-4,000
-6,000
-8,000
-10,000
-12,000 AM peak Off peak Inter peak Total PM peak
As shown in Figure 8, the A13 River Road has similar results to Edgware Road, albeit on a smaller scale, as there would be disbenefits from removing traffic signals during the morning and evening peak and a benefit during the inter-peak and off-peak. These are in line with the results that would be expected, with traffic signals required at busier times of day to regulate flows but less necessary at times when flows are lower. As shown in Figure 9, the East Barnet junction benefits from the removal of traffic signals at all times of day, with the largest benefit occurring during the morning peak. Again, the size of flow is an important factor in determining whether there is a benefit the flows at this junction are relatively low throughout the day therefore traffic signals are less necessary to regulate flows.
th
28

Figure 9: Impact of removing traffic signals at East Barnet junction

14,000
12,000
10,000
8,000
per year
6,000
4,000
2,000
0 Time saving -2,000 AM peak Off peak Inter peak Total PM peak Fuel VOC Non-fuel VOC Emissions Total
Figure 10:
12,000
Impact of removing traffic signals at Norwood Road junction
10,000
8,000
per year
6,000
4,000
2,000
0 Time saving -2,000 AM peak Off peak Inter peak Total PM peak Fuel VOC Non-fuel VOC Emissions Total
West Norwood has an overall benefit from removing traffic signals of approximately 11k a year, as shown in Figure 10. This is on a smaller scale than the benefits at some of the other junctions. There would be a slight disbenefit from removing traffic signals during the morning peak. Overall the results show that there are differences between individual junctions. For instance, the junctions at Edgware Road and A13 River Road indicate that there would be a disbenefit from removing traffic signals in the morning and evening peak, but a benefit from doing so
th
29

during the inter-peak. The East Barnet junction appears to benefit from removal of traffic signals at all times of day. One consistency is that all junctions benefit from the removal of traffic signals during the off-peak, when flows are typically lower. As tends to be the case in transport economic appraisals, the results are driven by the time savings, while the weighting given to the vehicle operating costs and emissions is comparatively smaller. The junctions with the largest scale of benefit / disbenefit are Edgware Road and Target Roundabout (Church Road); consequently the results for these junctions have the largest influence when a weighted average2 is produced, as shown in Figure 11. Figure 11: Impact of removing traffic signals, weighted average of five modelled junctions
10,000 0 Time saving -10,000 -20,000 Fuel VOC Non-fuel VOC Emissions Total
per year
-30,000 -40,000 -50,000 -60,000 -70,000 -80,000 AM peak Off peak Inter peak Total PM peak
The results in Figure 11 indicate that, on average, it would not be beneficial to remove traffic signals from junctions with the exception of the off-peak period. Other splits of results are also possible. Figure 12 to Figure 16 show results split by mode.
The five junction types modelled cover approximately 67% of junctions in London, but an adjustment has been made to effectively assume that they cover 100%. In other words, the results in Figure 6 act as a proxy for the average benefit / disbenefit per London junction.
th
30

Figure 12: Impact of removing traffic signals, car

200,000 100,000 0 -100,000 -200,000
per year
-300,000 -400,000 -500,000 -600,000 -700,000 -800,000
On the whole the disbenefits to cars of removing traffic signals outweigh the benefits. The disbenefits to cars are particularly large for Target Roundabout (Church Road); a disbenefit of approximately 690k a year which is equivalent to over 80% of the total Target Roundabout disbenefit. Figure 13: Impact of removing traffic signals, LGV
1,000
per year
The impact on LGV is relatively small. The biggest impact is at the Edgware Road junction, with disbenefits from removing traffic signals in the morning and evening peak and a benefit in the inter-peak and off-peak. LGVs are not present in the Target Roundabout model.
th
A4 1 A4 AM p 1 In eak t A4 er p e 1 PM ak A4 pe 1 Of ak Ch fp ur e Ch ch R A41 ak ur o ch ad Tot a A Ch Roa M l pe d ur ch In ak Ch Ro ter p a ur ch d P eak Ro M Ch ad pea k O Ea urch ff st Ro pea k a Ea Bar st n e d T o Ba t, A ta Ea rnet M p l st , e B Int ak Ea arne er p st e t Ba , PM ak rn pe e W Ea t, O ak es st t N Ba ff p e W es orw rnet ak tN o ,T o o W orw d A tal M es o t N od pe W or Int ak es wo er tN o p or d P eak M w W o es od pea tN O k or ff p w oo eak A1 d T ot 3 a A1 AM l pe 3 In ak t A1 er p e 3 PM ak A1 pe 3 Of ak fp A1 eak 3 To ta l
-500 -1,000 -1,500 -2,000 -2,500 -3,000 -3,500
A4 1 A4 AM p 1 In eak t A4 er p e 1 PM ak A4 pe 1 Of ak Ch fp ur e Ch ch R A4 ak 1 ur oa To ch d ta Ch Roa AM l pe d ur ch In a Ch Ro ter k ur ad p e ch P ak Ro M Ch ad pea k O Ea urch ff p st Ba Roa eak Ea st rne d T o Ba t, A ta Ea rne M p l st t, I nt eak B Ea arne er p st e t Ba , PM ak rn et pe W Ea a , es st Off k B t W No arn pea es r k t N wo et, T o W orw d A otal M es o t N od p W or Int eak es wo er tN o p or d P eak W woo M p es e t N d O ak or ff p w oo eak A1 d T ot 3 a A1 AM l pe 3 In ak t A1 er p e 3 PM ak A1 pe 3 Of ak fp A1 eak 3 To ta l
500 0
31
per year
-8,000 -6,000 -4,000 -2,000 -14,000 -12,000 -10,000
per year
A4 1 AM
0 50,000
2,000
-200,000 100,000 -50,000
-150,000
-100,000
A4
th
Figure 15: Impact of removing traffic signals, bus
Figure 14: Impact of removing traffic signals, heavy goods vehicle
Heavy goods vehicles have a disbenefit from removing traffic signals during the inter-peak and evening peak at the Target Roundabout. The scale of benefits / disbenefits is small at the other junctions.
32
p 1 In eak t A4 er p e 1 PM ak A4 pe 1 Of ak Ch fp ur ch A4 eak Ch ur Roa 1 T ch ot d a Ch Roa AM l pe d ur ch Int ak Ch Ro er p ur a ch d P eak Ro M ad pe Ch O ak Ea urch ff st Ro pea k a Ea Bar st ne d T ot Ba t, a A Ea rne M p l st t , I Ba nt eak Ea rn er p st et, e Ba PM ak rn et pe W Ea a , es st Off k B t W No arn pea es r k t N wo et, T o W orw d A otal M es o t N od p W or Int eak es wo er tN o p or d P eak W woo M p es e t N d O ak or ff p w ea oo k A1 d T ot 3 al A1 AM p 3 In eak t A1 er p e 3 PM ak A1 pe 3 Of ak fp A1 eak 3 To ta l
A4 1 A4 AM p 1 In eak t A4 er p e 1 PM ak A4 pe 1 Of ak Ch fp ur ch A4 eak Ch Ro 1 ur ch ad Tot a A Ch Roa M l p d ur ch Int eak Ch Ro er p a ur ch d P eak Ro M Ch ad pea k O Ea urch ff p st Ba Roa eak Ea st rne d To Ba t, A ta Ea rnet M p l st , e B Int ak Ea arne er p st e t, Ba PM ak rn et pe a W Ea , es st Off k B t W No arn pea es k r t N wo et, T o o W orw d A tal M es o t N od p W or Int eak w es oo er p tN or d P eak W woo M p es e t N d O ak or ff p w ea oo k A1 d T o 3 AM tal A1 p 3 In eak t A1 er p e 3 PM ak A1 pe 3 Of ak fp A1 eak 3 To ta l

Overall buses experience a significant disbenefit from removing traffic signals at the Edgware Road (A41) and Target Roundabout junctions, largely due to disbenefits in the evening peak. Figure 16: Impact of removing traffic signals, taxi
20,000
ea k
To ta l
ea k
pe a
pe a
fp ea
pe
pe
ea k
te rp
In te rp
AM
PM
AM
PM
Of fp
In
ne t,
A4
Ba r
Ba r
Ba r
Ba r
st
st
st
Ea
per year
-40,000
-60,000
-80,000
-100,000
-120,000
Taxis are only present in the Edgware Road (A41) and East Barnet models. There is a benefit to taxis from removing traffic signals at East Barnet, but this is negligible and does not show up on the chart. There is a relatively large disbenefit to taxis from removing traffic signals at the Edgware Road junction (just over 100k a year). Impacts on motorbikes and bicycles are negligible and are not shown here. The main observation from the charts for individual modes is that the largest benefits / disbenefits apply to car and bus. This is unsurprising as cars form the majority of vehicle flows and bus has the highest value of time per vehicle due to its level of passenger occupancy. Another split that can be shown is the benefits by junction type. In this case we have split by 4-arm junction, 3-arm junction and roundabout (a 4-arm junction but treated separately here). The 4-arm junction results are a weighted average of Edgware Road (A41) and East Barnet results; the 3-arm junction results are a weighted average of West Norwood and A13 River Road results. Only one roundabout (Target Roundabout, Church Road) has been modelled.
th
33
Ea
Ea
Ea
st
Ea s
tB ar
A4
A4
-20,000
ne t,
ne t,
ne t,
A4
ne t,
Of
A4
To
ta l
ak
ak

Figure 17: Impact of removing traffic signals, by junction type

100,000 0
pe ak
pe ak
pe ak
pe ak
ea k
ea k
al
ea k
ea k
al
pe ak
pe ak
ea k
To t
To t
ea
Of fp
In te rp
In te rp
Of fp
AM
AM
PM
PM
4ar m
3ar m
tP M
nt e
tI
un da bo u
bo u
bo u
4ar m
3ar m
4ar
3ar
bo u
un da
un da
Ro
-300,000
per year
-400,000 -500,000 -600,000 -700,000 -800,000 -900,000
The 4-arm and 3-arm junctions both show a disbenefit of removing traffic signals in the peak periods, with a benefit during the inter-peak and off-peak. The scale of impacts is larger for 4arm junctions, although this may be due to other individual junction characteristics rather than a reflection of a consistent difference between 4-arm and 3-arm junctions. Target Roundabout on the other hand shows a large disbenefit in the inter-peak, as well as a disbenefit in the evening peak.
Pedestrians
The analysis undertaken does not take into account the benefits and disbenefits to pedestrians. This, as discussed previously, is largely due to a lack of data, but also from the lack of validated methods of forecasting and modelling pedestrian behaviour when traffic signals are not in use. It is apparent from the analysis that there are disbenefits from removing traffic signals during the morning and evening peaks, and this generally coincides with periods when pedestrian numbers are also high. The inter-peak period is more complex, in parts of London both pedestrian and vehicle numbers are high during this time, but in other locations numbers are much reduced. Where it has been shown that there are benefits from switching off (introducing flashing amber) traffic signals during certain periods, it is possible that these benefits would significantly reduce if pedestrian actuation of an all-red pedestrian crossing stage was introduced, resulting in additional delay to vehicles. This is more likely to be an issue at inner-London sites and could therefore negate any benefits.
th
34
Ro
Ro
Ro
-200,000
un da
Ro
un da
4ar
4ar
3ar
3ar
bo u
tA
tO
-100,000
ff p
tT ot
rp
al

Road safety
The economic analysis of switching off traffic signals does not account for the possibility of benefits or disbenefits arising from the impact of the proposals on road safety, more specifically personal injury accidents (PIAs). There are very limited studies of the impact of removing formal control at junctions on road safety, and what data there is seems to provide mixed messages. A report written for TfL in 2006 (TRL PPR292 A Review of Simplified Streetscape Schemes) concluded that the collision data (from a number of European schemes) did not provide a safety case for simplified streetscapes one way or the other. This did not specifically deal with part-time switching off of traffic signal control, but provides a reference to sites with before and after data for the presence of traffic signals. As part of the study, accident data across London for periods when traffic signals were not in use due to a fault was examined. LRSU data showed that in three years there were around 180 PIAs (60 PIAs per annum) at sites where signal faults had occurred. For the year up to February 2009, there were over 2,700 faults. The length of time that signals were not in use varied considerably, with a modal average of 2 hours, yet a mean average of 21 hours. This gives 0.0010 PIAs per hour, or 9 PIAs per site per year if signals are always out, compared to an average of 2.4 accidents at signalised junctions in inner-London. It is difficult, however, to draw clear conclusions from the data as when signals are not working no alternative guidance in terms of priority is provided to drivers. As shown below in Table 10, the average value of a personal injury accident on urban roads is around 91k. The potential impact on the economic benefits of considering the effect of the change in PIAs as a result of switch-off could be very significant. With junction benefits at sites during certain periods of the day valued at around 100k or less, it would only take one additional injury accident occurring at the junction per year to negate all benefits. On the other hand, any single accident saving could easily double the benefits of switch-off.
th
35

Table 10: Collision costs by severity and type of road ( per accident, June 2007 prices)3
Type of collision
Fatal Serious Slight All injury collisions Damage-only collisions Average collision cost per injury collision (including an allowance for damage-only collisions)4
Urban roads 1,769,900 207,120 21,000 59,240 1,840 91,810
Rural roads 1,930,740 231,110 24,750 121,420 2,720 142,640
Motorways 2,145,280 235,690 29,490 91,930 2,620 111,810
All roads 1,876,830 215,170 22,230 75,610 1,970 104,900
Source: Issue 12 of Levels of Collision Risk in Greater London (Feb 09)
CONCLUSIONS
The above framework has been developed to help policy makers review existing signalised traffic junctions in new light. As such, the review should be carried out without prejuidice to improve operational performance of the junctions. New radical ideas such as designing shared space and more complicated control mechanisms can be easily tested using the above approach.
It was assumed that if the traffic signals were to be switched-off (or flashing amber was introduced) for all or part of the day, drivers will be informed about the alternative control regime before being introduced. As such, the study assumes drivers to be informed about the new regulations and expected behaviour through regulating traffic signs, public information campaigns, and training and through gradual learning and word of mouth. The set of junctions selected for this study represents a wide range of junctions from both outer and inner-London. They also represent a good range of traffic and pedestrian demand throughout the day. It should also be noted that different junctions can display very different characteristics. It is fair to say that there is no such thing as an average junction, so the results should be treated with caution. In particular, some junctions including roundabouts are peculiar examples of signalised junctions where the delay for individual arms and that for overall traffic is highly dependent on the balance of flows and available gaps for major traffic movements. Unbalanced flows, eg in the inter-peak scenarios, can result in higher delay for all traffic but can be easily minimised by introducing demand operated traffic signals.
3 4
Values taken from Road casualties Great Britain 2007 Department for Transport September 2008 Department for Transport figures from the in draft Accidents Sub-objective Unit on the Transport Analysis Guidance web site (www.dft.gov.uk/webtag)
th
36

Each of the modelled junctions was used to analyse two sets of results: the existing or Base (Do Minimum) scenario and a Do Something scenario where the traffic signals were replaced with an alternate measure of control. The main conclusions are as follows: Traffic signals generally provide significant benefits to road users; It would, however, be beneficial to switch-off traffic signals at some junctions at particular times of day; In particular, there would be a benefit at the junctions studied from switching off during the off-peak, after a full safety assessment.
The results do not include the net economic cost or benefits to pedestrians who are assumed to cross in gaps in the traffic or at stand alone pedestrian crossings. If delays to traffic are imposed by pedestrians calling up a pedestrian crossing stage during the period of flashing amber, this could have a significant impact on any benefits. The results do not include the net economic cost or benefit due to changes in accident numbers. The studies that were carried out to attempt to value this impact were inconclusive, yet with the average cost of a personal injury accident at over 90k, an increase or decrease in accident levels could have a significant impact on the results. This study has identified substantial economic benefits to road users from having traffic signals and this benefit needs to be more widely promulgated. The study has also demonstrated that for certain junctions at certain periods of the day there would, based on the assumptions made regarding traffic and pedestrian behaviour, be some benefit to switching off traffic signals (or introducing flashing amber). It is evident, however, that this is site-sensitive and can only be used as a broad guide to the type of sites that might deliver such a benefit. The study assumes that when traffic signal control is disabled, traffic behaviour would revert to some form of conventional priority control, which might even be stipulated through analysis of traffic demand and turning patterns and the use of advance signing. It is possible, however, that junctions could operate without any imposition of regulated traffic controls, with the expectation that road users would behave appropriately. This form of behaviour cannot, at present, be modelled yet it is recommended that scope for this form of uncontrolled arrangement is also investigated. This can only be achieved through live trials at a variety of sites. The results would have the potential of determining precisely how traffic would behave at shared space type environments and could provide unparalleled knowledge in this field. Such work would also need to monitor the behaviour of pedestrians.
th
37

WCTR-Jul - 10 Framework For Evaluating The Economic Impact of Traffic Signals

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WCTR-Jul - 10 Framework For Evaluating The Economic Impact of Traffic Signals

Uploaded by

Copyright:

Available Formats

Framework for evaluating economic impact of traffic signals

Chandra, Ashish; Firth, Keith; and Carrignon, David

FRAMEWORK FOR EVALUATING ECONOMIC THE IMPACT OF TRAFFIC SIGNALS

Framework for evaluating economic impact of traffic signals

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Framework for evaluating economic impact of traffic signals

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Figure 1: Methodology for assessment of traffic signals

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

KEY ASSESSMENT CRITERIA

Individual site road safety assessment

The Ratio Of Pedestrian And Cyclists To Vehicular Traffic

Framework for evaluating economic impact of traffic signals

Junction Layout And Geometry

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Pedestrian and cyclist provision

Framework for evaluating economic impact of traffic signals

Road network management

Congestion and capacity assessment

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Source: Transport in the Urban Environment

ALTERNATIVE METHODS OF CONTROL

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Amber Flashing Light at intersections

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Alternative traffic control regimes

Off-Side Priority Rule (OSPR)

Major/minor Priority Control

Filter-in-turn according to order of arrival

Give way to vehicles approaching from the right (as if mini-roundabout)

Traffic signal regime Vehicle to pedestrian control regulations

Switched off or Switched off or flashing amber flashing amber

Modelled User Behaviour

No premeditated order, completely random

No premeditated order, completely random

Drivers judge priority according to order of arrival

No premeditated order, completely random

Pedestrians cross in gaps

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Amber Amber Lunar white

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Switch off, no traffic signal Description

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Flashing amber, priority to the right Description

Framework for evaluating economic impact of traffic signals

Flashing amber, major/minor priority Description

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

12 WCTR, July 11-15, 2010 Lisbon, Portugal

Framework for evaluating economic impact of traffic signals

Key arms and intersections