You are on page 1of 8

International Journal of Traffic and Transportation Engineering 2015, 4(5): 123-130

DOI: 10.5923/j.ijtte.20150405.01

The Impact of Speed Radars on Drivers’ Behavior:


A Case Study in Dubai
Akmal Abdelfatah*, Mahmoud Taha, Ahmed Ramadan, Mohammed Masri,
Rami Kakish, George Kazma

Department of Civil Engineering, American University of Sharjah, Sharjah, United Arab Emirates

Abstract This paper investigates the impact of speed radars on driving behaviours. A spot speed study has been
performed to determine the level of driver compliance to the posted speed limits at locations between two radars and locations
without radars. Handheld pocket radars have been utilized to record the vehicle speeds on three different road segments. The
first segment is at the exact location of a speed radar, while the second segment is at a location between two speed radars, and
the third is at a road segment with no speed radars. After data collection, statistical analyses were carried out to determine the
descriptive statistics of the speed. In addition to that, the significance of the difference between speeds was examined through
tests of hypotheses. In conclusion, radars are effective in keeping the drivers within the enforcement speed limit, but far less
effective in keeping drivers within the posted speed limit. The Dubai Police Department is recommended to reduce the
difference between the posted speed limit and the enforcement speed.
Keywords Speed enforcement, Spot speed study, Traffic safety

causing the actual speeds on the major freeways (with


1. Introduction posted speeds of 120 km/h) to be around 140 km/h. This
paper presents the results of a spot speed study on three
Most of the major highways in the world and in Dubai segments on Dubai freeways.
specifically, are controlled by speed enforcement tools
(speed radars), which serve the purpose of controlling the
drivers’ speeds on a particular highway and issuing fines for 2. Literature Review
drivers who violate/exceed the speed limits. However, Spot speed studies should be conducted at times when the
radars are scattered over the length of an entire highway, road segment experiences “free-flow” driving conditions [1].
sometimes with enough distance between consecutive Furthermore, local law enforcement and other officials,
radars for drivers to exceed the speed limit and still not get such as Dubai Police Department, need to be contacted and
fined as these drivers would be out of radar’s range. This approvals should be permitted before conducting a spot
study aims at examining the driver compliance to posted speed study.
and enforcement speed limits at road segments between two According to a study conducted in 2002, there are around
radars and without radars in Dubai - United Arab Emirates 1 million annual fatalities worldwide due to traffic accidents,
(UAE), and provide statistical data showing compliance or and about 10 million permanently disabled every year. By
noncompliance. the year 2020 estimations show that traffic accidents will
This study investigates the effect of speed radars on road become the third leading epidemic in fatalities around the
drivers’ behavior and whether they actually help in reducing world, preceded by clinical depression and coronary heart
the drivers’ maximum speed on the road, which has a major disease. In the UAE, traffic accident fatalities significantly
impact on the number of accidents. The goals of this study exceed those of other industrialized countries of Europe and
are to determine existing roadway speeds and evaluate the the United States. Moreover, the study concluded that
effectiveness of speed enforcement programs (speed traffic accidents are the second most leading cause of
radars). fatalities in the UAE. Among top ranked potential causes of
The enforcement speeds on Dubai highways are typically traffic accidents is exceeding the speed limits and driving
higher than the posted speed limit by 20 km/h., which is carelessly [2].
Based on the traffic accidents that occurred between 1995
* Corresponding author:
akmal@aus.edu (Akmal Abdelfatah)
and 1999, Dubai has the highest fatality rates per 100,000
Published online at http://journal.sapub.org/ijtte population and registered vehicles compared to the other
Copyright © 2015 Scientific & Academic Publishing. All Rights Reserved countries in the group, as illustrated in Figure 1 [2].
124 Akmal Abdelfatah et al.: The Impact of Speed Radars on Drivers’ Behavior: A Case Study in Dubai

Table 1 illustrates the traffic accident data in Dubai for violations, speeding violations captured by radars were the
years 2003 to 2012 based on the data provided by the highest accumulating a total of about 150 thousand
General Department of Traffic Police (Dubai Police), and violations (about 67% of the total violations) [6]. The
Road and Transportation Authority (RTA) [3]. percentage of speeding violations is a good indication that
As illustrated, the total number of accidents has an speeding is a major cause of traffic accidents and fatalities.
increasing trend from 2003 till 2008. However, the trend The previous discussion shows the seriousness of the
from 2009 till 2012 is decreasing due to the implementation speeding problem in Dubai, as one of the major cities in
of the new traffic law amendments in 2008 in Dubai [3]. UAE. Therefore, the local authorities in Dubai are using
Some studies showed that traffic accidents to be among speed radars as a tool to reduce the speeds on Dubai’s road
the highest ten causes of death worldwide (9th cause of network. Accordingly, there is a need to understand the
death), with a total number of deaths equal to 1.3 million impact of speed enforcement on drivers’ behavior. This will
[4]. More recently, the worldwide fatalities because of help the local authorities in taking decisions regarding the
traffic accidents were estimated to be about 1.24 million applications to be used for speed enforcement.
[5]. The following paragraphs summarize some of the efforts
According to statistics issued by Dubai Statistics Center carried out, at different parts of the world, to study the
in 2012, traffic violations in a month have reached more effect of speed enforcement on drivers’ behavior.
than 225 thousand traffic violations. Among these traffic
Table 1. Traffic accident data of Dubai (2003-2012) [3]

Total number Accidents per 100,000 Number of Accidents per 100,000


Year Population
of accidents population registered vehicles registered vehicles

2003 2,287 1,141,487 200.4 435,722 52.5

2004 2,413 1,228,178 196.5 500,272 48.2

2005 2,794 1,321,453 211.4 630,000 44.3

2006 3,224 1,421,812 226.8 740,187 43.6

2007 3,335 1,529,792 218 868,592 38.4

2008 4,011 1,645,973 243.7 996,997 40.2

2009 3,576 1,770,978 201.9 1,021,880 35

2010 2,686 3,300,000 81.4 1,031,961 26

2011 2,636 3,361,183 78.4 1,055,412 25

2012 2,932 4,089,107 71.7 1,137,750 25.8

Figure 1. International traffic fatality rates, 1995-99 [2]


International Journal of Traffic and Transportation Engineering 2015, 4(5): 123-130 125

A study conducted in Barcelona evaluated the level of street showed only a 1.6% decrease with the enforcement.
reduction in the number of road collisions, number of These results confirm that adding speed radar enforcement
injuries, and number of cars affected by an accident due to affects the average speeds of vehicles traveling on the
the presence of speed cameras on Barcelona's beltway. The streets on which the enforcement is implemented [12].
study compared accidents on two road segments, where one According to a study conducted in British Columbia (BC),
road had no speed cameras installed and one with speed Canada, and using time series analysis, it has been found
cameras installed. The analyses of the study were based on that photo radars reduced traffic collisions by 14% and
collisions and accidents’ data. The results of the study collision fatalities by 26%. Non radar controlled and radar
concluded that speed cameras are effective in reducing the controlled sites were compared to assess traffic and
number of accidents and, therefore, the number of injured collision migration. It has been observed that the mean
people [7]. speed at non-photo radar monitoring sites decreased by 3%.
Oliver and Gernot [8] discussed the implementation of a The study concluded that the automated speed enforcement
smart camera traffic enforcement system at a railway program reduced collisions considerably at both radar
crossing as a tool to improve the traffic safety. The system controlled (-14% ± 11%) and non-radar controlled locations
has been designed to capture traffic violations. (-19% ± 10%). In conclusion, “the BC Photo Radar
In another study in London, UK, Walter et al. [9] Program was proved to be effective in reducing traffic
presented the impact of increasing the police enforcement speeds at enforcement and non-enforcement sites. Moreover,
within an urban area. The researchers reported the effect of crashes, injuries, and deaths were reduced as well [13].
setting speed radars and more static and mobile police Transport Quebec issued its evaluation report on the first
enforcement that are easily recognizable by the drivers. The year of its 18-month photo radar pilot project. The report
authors concluded that there was a systematic speed concluded that photo radars are effective in reducing
reduction along the considered corridor due to the vehicle speeds and road accidents. The pilot project
application of the enforcement measures. However, there comprised of 15 locations: 6 fixed photo radar sites, 3
was no effect of these enforcement measures on the use of mobile photo radar police vehicles and 6 traffic signal fixed
seatbelts and mobile phones. installations. In fixed radar locations, drivers reduced their
Dowling and Holloman [10] conducted a study to speeds by an average of 12 km/h. A 9km/h reduction in
determine the most effective way to position or display average speed was recorded in mobile radar locations.
enforcement speed controls, such as radars. They indicated Overall, there was a 99% reduction in excessive speeding, a
that the purpose of traffic enforcement is not only to issue 63% reduction in speeding and a 20-35% reduction in
tickets for drivers violating the posted speed, but also to accidents across the 15 locations. As for the technological
alter the drivers about the perceived risk associated with aspect of the radars, the report mentioned that the stationary
speeding behavior. radars were able to monitor 92% of all passing vehicles and
Another study was conducted on the rural roads in the mobile radars were able to monitor over 80% of passing
Dutch province Friesland. Enforcement was mainly vehicles [14].
implemented by means of movable radars. The evaluation In the US, Haibabaie et al. [15] evaluated the effect of
covered a period of 5 years of enforcement, where the speed using radars for speed enforcement in work zones. The
data showed a great decrease in mean speed and the study concluded that the speed radar did not have an impact
percentage of speed limit violators over time. The largest of on the headway at the radar location and it’s downstream.
these decreases was witnessed in the first year of the Liu et al. [16] analyzed the effect of controlling the speed
enforcement implementation, and also in the fourth year, on the road network using radars and cameras. The results
when enforcement was strengthened. Decreases in mean showed that thee radars have a limited effect on the speed.
speed and number of violators were also noticed on a On the other hand, the cameras showed much better effect
nearby roadway that was not subjected to enforcement on limiting the speeds. The effectiveness of speed
radars due to the spillover effect. The best approximation enforcement on freeways has been evaluated by Xu et al.
for the safety effect of the enforcement project is a [17]. They concluded that the effectiveness of automated
reduction of 21% in both the number of injury accidents and speed enforcement depends on the location of the
the number of serious casualties [11]. enforcement device. The effect on speeds may extend to
A study that examined the effect of speed radars on about 2 km, when the speed enforcement device is fixed on
traffic speeds was conducted in Vancouver, Canada. The cantilevers. However, road side speed enforcement showed
researchers conducted this study on two arterial streets, one an effect of about 1 km only.
with radar enforcement in place and the other with no speed As it is noticed from the literature review, there is a good
radar enforcement. The summary data for the study showed evidence that the speed radars have an effect on drivers’
the traffic volumes, average daily speeds, and the behavior. Also, the speed radars not only result in a
percentage of vehicles exceeding the speed limit. The study reduction in the average speed at their locations, but also
found out that average speeds on the radar enforced street they can result in a reduction in the average speed on roads
are reduced by 2.2%, on the other hand, the unenforced in the vicinity of the controlled segments.
126 Akmal Abdelfatah et al.: The Impact of Speed Radars on Drivers’ Behavior: A Case Study in Dubai

This paper is aiming at quantifying the level of speed obtain the equivalent leftmost lane speeds [18].
reduction that is achieved on Dubai’s road network due to This study revolves around a null hypothesis (Ho) that
the extensive use of speed radars on the majority of the states that “drivers do not actually comply with the
major roads in Dubai, especially on freeways. maximum radar speed limit on a road for the entirety of the
road, but rather they only slow down at the locations of
fixed speed radars”. The only way to accept or reject this
3. Methodology hypothesis is by testing its validity through “Test of
A handheld “Pocket Radar Model PR1000” is used Hypothesis” procedure [19].
throughout the study to measure the speeds of vehicles. The The null and alternative hypotheses can be written as
speeds are measured from the back of the vehicles in order follows:
to reduce the chances of drivers spotting and reacting to the Ho: µ < Radar speed limit (enforcement speed)
pocket radar. H1: µ ≥ Radar speed limit (enforcement speed)
A pilot study was conducted to determine the sample size
Where µ: Mean speed on a particular road.
required to obtain a representative sample of the entire
population during the actual study. In essence, the standard
deviation of the speed readings of the pilot study is needed 4. Results
to determine the sample size. The tolerance is chosen to
satisfy the accuracy requirements while at the same time This section presents the pilot study results and the spot
keeping the required sample size within a manageable and speed study results.
practical level.
4.1. Pilot Study Results
Readings of vehicle speeds travelling on 3 different
highway sections have been obtained (i.e. without a radar In the pilot study, 50 speed readings were collected on
control, at the exact location of fixed radar, and between Friday, March 1, 2013, from 11-11:50 A.M. It should be
two fixed radars). The sequence of vehicles’ speeds to be noted that Friday is an official weekend day in Dubai. The
measured is at random in order to eliminate any bias. selection of the morning time on a weekday is to guarantee
The speeds of vehicles travelling on left lanes were free-flow conditions on the freeway. Accompanied by
measured at the location between radars and at the location Dubai Police; the readings were recorded at a location
without a radar control only, since bridges overpassing the between two radars on Emirates Road (one of the major
highway are available at these locations. The speeds of the freeways in Dubai), above a bridge overpassing the
vehicles travelling on the highway underneath the bridge highway. The posted speed limit on this freeway is 120
can be measured safely. Due to the unavailability of an km/h and the enforcement speed (radar speed) is 140 km/h.
overpassing bridge at the location that is at a radar control, The speeds of vehicles on the leftmost, middle, and
the speeds of vehicles travelling on the right lane were rightmost lanes were measured in order to examine the
measured. The readings obtained from the right lane on the difference in speeds of vehicles on different lanes.
location that is at a radar control are multiplied by an As observed, the speeds’ average on the leftmost lane
adjustment factor in order to estimate the speeds of vehicles was the highest, and on the rightmost lane was the lowest,
travelling on the left lane in that location. The factor is as expected. The standard deviation of the speed
calculated from the ratio between the average speeds of measurements from the leftmost lane was found to be 9.905
vehicles travelling on the left lane and the vehicles km/h. Using a tolerance of 2 km/h, the required sample size
travelling on the right lane at the location between two was calculated to be approximately 100 vehicles.
radars. The test of hypothesis proved that speeds of vehicles
The collected data is summarized in the form of travelling on each lane group are different from each other
frequency and cumulative frequency curves to examine the (i.e. μ1 ≠ μ2, μ1 ≠ μ3, and μ2 ≠ μ3).
data distribution and the total percentage of drivers driving
below a certain speed [18]. 4.2. Spot Speed Study Results
As mentioned earlier, data taken from the right lane on After collecting all the readings for the three cases
the location that is at a radar control is multiplied by a considered (at radar, between two radars, and no radar
factor to obtain the equivalent left-lane speeds. Speeds of locations), the frequency and cumulative frequency
vehicles travelling at the left and right lanes in the location distribution curves were plotted. The curves help in the
between radars were collected at the same time, and the initial determination of the normalcy of data points for each
ratio between the average speeds of these two lanes is used of the three road cases.
as a factor to convert the right lane readings of the location From Figure 2, the cumulative frequency distribution
at a radar control to left lane readings. However, before curve of the location between radars, it can be deduced
proceeding with this step, a Chi-Square (χ2) test between without further analysis that the majority of drivers drive
the two sets of data is required to test the similarity of the within the enforcement speed. However, the same cannot be
pattern of the two samples to decide if it is possible to said about the posted speed, as the majority of the drivers
multiply the speeds from the right lanes by a factor to drive above the posted speed limit.
International Journal of Traffic and Transportation Engineering 2015, 4(5): 123-130 127

(χ2) test resulted in a calculated (χ2) value higher than the


100% decision value, using the same degrees of freedom and the
Cumulative % Frequency

80% same α. Therefore, the right lanes of the two different cases
do not exhibit the same distribution, making the initial
60% assumption invalid.
40% This third (χ2) test was performed in order to prove the
assumption that if the left lanes of two different cases
20% exhibit similar distributions of speeds, then it can be
0% assumed that the left lane speed distribution of the third
case, location at the radar, would exhibit a similar
118 128 138 148 158
distribution as well. Using the same decision value, α, and
Speeh (km/h) number of degrees of freedom, the calculated (χ2) value of
101.03 is greater than the decision value of 12.59, thus
Figure 2. Cumulative Frequency Distribution Curve – Location between indicating that the distributions of these two cases are not
Radars
similar and deeming the initial assumption invalid.
From Figure 3, the cumulative frequency distribution The results of the three (χ2) tests came to the result that
curve of a location with no radars, a different pattern is the speed distributions of all three cases are not similar.
observed. Although the observation that most of the drivers There is data for the left lanes in two of the cases, between
exceed the posted speed limit still holds, another radars and without radars, and the data of the right lane in
observation that a significant percentage of the drivers also the case at the location of radar. So it’s not possible to
do not drive below the enforcement speed limit. compare the former two cases with the latter case through a
test of hypothesis.
100% Having completed the (χ2) tests, three cases of tests of
Cumulative % Frequency

hypothesis were performed. The first case involved


80% comparing each of the speed averages of the three cases
60% studied by the group with their respective posted speed limit
and enforcement speed limit. To start the test, a pair of null
40% and alternative hypotheses for each case was defined, one
20% involving the posted speed limit, and the other involving the
enforcement speed limit.
0% It should be noted that the posted speed limit for the
100 120 140 160 cases between radars and at radar is 120 km/h on the
Emirates Road, while it is 100 km/h on Al-Khail road for
Speeh (km/h)
the case without radars. The enforcement speed limit,
Figure 3. Cumulative Frequency Distribution Curve - Location without however, is 140 km/h and 120 km/h, respectively. The null
Radar (Ho) and alternative (H1) hypotheses are defined in Table 2.
Since it was not possible to acquire left-lane data for the Table 2. The Null (Ho) and Alternative (H1) Hypotheses for the Three
Cases
case at the radar, the speeds on the most right lane were
measured. Accordingly, it was required to decide on a Left Lane, Left Lane,
Right Lane,
certain factor to convert the data into equivalent left-lane Between Without
At Radar
Radars Radars
data. The factor cannot be obtained and used unless the
distributions for the left lane and right lane are similar. The With Respect to
H0: µ = 120 H0: µ = 100 H0: µ = 120
Posted Speed
(χ2) test was performed in order to check the similarity H1: µ > 120 H1: µ > 100 H1: µ > 120
Limit
between the two distributions.
With Respect to
The Level of Significance, α, assumed to be 0.05, along Enforcement
H0: µ = 140 H0: µ = 120 H0: µ = 140
with the number of degrees of freedom 6 resulted in a Speed Limit
H1: µ > 140 H1: µ > 120 H1: µ > 140
decision (χ2) value equal to 12.59. The sum of Chi-Squares
Where µ: The speed average recorded on that road.
calculated is equal to 158.9, which is significantly greater
than the decision value. The test indicates that the two After defining the null and alternative hypotheses for
distributions are not similar, and thus a factor cannot be each of the cases, the test is performed using the z-value of
obtained to be then multiplied by the right lane at radar data the sample and the z-value at 95% confidence level.
in order to get the equivalent left lane data. Equation 1 was used to compute the z-value of the sample.
The (χ2) test was performed in an attempt to prove that
the distributions on the right lanes of two different cases to zsample = (µ-µ0 )/(s/√𝑛𝑛) (1)
be similar, then it can be assumed that the distributions on The value obtained from the equation would then be
the left lanes of the same two cases would be similar. The compared against the z-value at 95% level of confidence,
128 Akmal Abdelfatah et al.: The Impact of Speed Radars on Drivers’ Behavior: A Case Study in Dubai

which is equal to 1.645. If the zsample value is greater than After using equation (2), Zsample was equal to -9.4, which
the 95% z-value, Ho is rejected while H1 is accepted, and means that Ho is rejected and H1 is accepted.
vice versa. The results are shown in Table 3. Specific patterns were observed in the data points
collected during the spot speed study. These Patterns
Table 3. Results of the First Test of Hypothesis
enabled a better understanding of the drivers’ behavior in
Left Lane,
Left-Lane, Right Lane,
relation to the availability of a radar control on the roadway.
Between The patterns are shown in Figures 4 to 6 and discussed in
Without Radar At radar
radars
the following section.
µ=132.07 µ=129 µ=114.06
Parameters
s=5.762 s=17.215 s=12.05
With 5. Discussion of the Results
Respect to
ZSample=20.95 ZSample=16.85 ZSample= -4.93
Posted In the roadway section with no radar controls, it is
Speed Limit observed that the majority of drivers (68%) drive at or
ZSample > Z95% ZSample > Z95% higher than the enforcement speed of the radar as shown in
Result of ZSample < Z95%
Test
Reject Ho, Reject Ho, Figure 4. Also, the percentage of drivers speeding driving at
Accept Ho
Accept H1 Accept H1 or above the posted speed limit on the same section is about
With (94%), as illustrated in Figure 5. This indicates that the
Respect to
ZSample= -13.76 ZSample= 5.23
ZSample= majority of drivers commit speeding violations on a
Enforcement -21.53 roadway without radars.
Speed Limit

Result of ZSample < Z95%


ZSample > Z95%
ZSample < Z95% 30
Reject Ho, % of Observations 25
Test Accept Ho Accept Ho
Accept H1 20
15
The second case of the two tests of hypotheses involves
10
comparing the average speed of the case between radars on
5
the left lane to the case without radar control on the left lane
0
as well. For the third case, the location at radar on the right
lane, cannot be compared with the other two because of the
significant difference in speed between the lane groups,
proven by the analysis of the pilot study data, and the (χ2)
tests that prevented the use of an adjustment factor to get % of Enforcement Speed Limit
equivalent left-lane data for this case.
Figure 4. Distribution of Vehicle Speeds in Relation to the Enforcement
The test concerns the difference between two means
Speed - Location with no Radars- Left Lane
which come from independent samples. In ideal cases, the
difference in the speed averages of these two cases should
25
be 20 km/h or more since the difference between the posted
% of Observations

speed limits of the 2 roadways is 20 km/h. If, however, the 20


difference in averages is less than 20 km/h, this is an 15
indicator that on average drivers on the road without radar 10
enforcement are driving at speeds close to the average of
drivers driving between two radars, which is to say that the 5
drivers on the road without radar enforcement are violating 0
both the posted and enforcement speeds. As in the previous
case, the null and alternative hypotheses were first defined
and set up as follows:
H0: µ1 - µ2 ≥ 20
% of Posted Speed Limit
H1: µ1 - µ2 < 20
Figure 5. Distribution of Vehicle Speeds in Relation to the Posted Speed -
Also as in the previous test of hypothesis, the z-value for Location with no Radars-Left Lane
the sample is calculated and is compared to the z-value at
95% level of confidence; if the sample z-value was larger Considering the roadway section with radar control, it is
than the 95% z-value, Ho is rejected and H1 is accepted. observed that the percentage of speed violations drop
The z-value for the sample is calculated using Equation 2. significantly in comparison to a roadway without a radar
control. As shown in Figure 6, the majority of drivers (90%)
𝑠𝑠 2 𝑠𝑠 2 drive at speeds that are lower than the enforcement speed,
zsample = (µ1 − µ2 − 20)/�� 1 � + � 2 � (2)
𝑛𝑛 1 𝑛𝑛 2 and only (10%) of the drivers violate that speed. It can also
International Journal of Traffic and Transportation Engineering 2015, 4(5): 123-130 129

be noted that drivers overall drive very close to the enforcement speed, as the majority of drivers drive within
enforcement speed, as observed from the distribution where very close proximity to that enforcement speed. However,
all the observations are packed around the enforcement the data also shows that radars do not necessarily succeed in
speed. keeping drivers within the posted speed limits, as the
majority of drivers tend to drive above that speed.
100 From the road section without radar control, it is
% of Observations

80 observed that the majority of drivers commit serious traffic


60 offences, exceeding the enforcement speed by 30-40% in
40 some cases. In comparison to the road section with radar
control, speeding violations are much greater.
20
It was not possible to compare the case that is at the exact
0 location of a fixed radar control with the other two cases as
only right lane measurements were obtained for that case
due to safety regulations enforced by the Dubai Police.
Trials to convert the right lane measurements to equivalent
left lane measurements using adjustment factors or
% of Enforcement Speed Limit normalization techniques were proven unsuccessful due to
Figure 6. Distribution of Vehicle Speeds in Relation to the Enforcement the difference in distribution between the lane groups.
Speed - Location between Radars-Left Lane

As for the compliance of drivers to the posted speed limit, 6. Conclusions


none of the speeds recorded at the location between the
radar were below the posted speed limit, as depicted in After the analyses conducted for the spot speed study, it
Figure 7. was proven that radars help reduce the number of speeding
violations. However, legally, speed violations describe only
60 the speeds above the enforcement speed, not the posted
% of Observations

50 speed. So, if the effectiveness of radars is to be measured


40 against keeping drivers within the enforcement speed limit,
30
20 then the radar is deemed as effective. However, the
10 effectiveness of radars quickly diminishes if it is measured
0 against the posted speed limit, as overall, radars do not
succeed to keep drivers within the posted speed limit.
In conclusion, radars are effective in reducing the
average speed of vehicles travelling on a roadway. However,
due to the difference between the posted and enforcement
speeds, the majority of drivers tend to drive closer to the
% of Posted Speed Limit
enforcement speed, rather than the posted speed. Therefore,
Figure 7. Distribution of Vehicle Speeds in Relation to the Posted Speed - the Dubai Police is recommended to reduce the difference
Location between Radars-Left Lane between the posted speed limit and the enforcement speed
in order to reduce the actual speed on Dubai’s road
Between radars, on the left lane, drivers, on average,
network.
violate the posted speed limit but are contained within the
enforcement speed limit. However, in the road segment
without radar control, drivers violate both the posted speed
limit and the enforcement speed limit. Finally, in the
location at the radar on the right lane, drivers comply with REFERENCES
both the posted and enforcement speed limits, which is
[1] Ross, R.; Prassas, E.; McShane, W. (2011) “Traffic Engi
expected as the right lane is the slow lane and cannot be neering”, 4th edition, Pearson Printice Hall.
compared with the other two, left-lane cases.
The result of the second test of hypothesis indicates that [2] Abdalla, I. (2002). “Fatality Risk Assessment and Modeling
the average speed on a road segment without radar is higher of Drivers Responsibility for Causing Traffic Accidents in
Dubai.” Journal of safety research, 33(4), 483-496.
than both the speed limit and the enforcement speed as well.
For the location between radars, the average speed is higher [3] Abdelfatah, A.; Al-Zaffin, M. S.; Hijazi, W. (2015). “Trends
than the speed limit but less than the enforcement speed. and Causes of Traffic Accidents in Dubai.” Journal of Civil
From the distribution of data obtained from the location Engineering and Architecture, 9(2), 225-231.
between radars, it can be recognized that the use of radars [4] WHO (2004) “The Global Burden of Disease.” Geneva,
on a roadway can help in keeping drivers within the Switzerland.
130 Akmal Abdelfatah et al.: The Impact of Speed Radars on Drivers’ Behavior: A Case Study in Dubai

[5] WHO (2013) “Global Status Report on Road Safety 2013.” on Traffic Speed: A Quasi-Experimental Evaluation." The
Geneva, Switzerland. Canadian Journal of Program Evaluation 9(1): 51.
[6] Dubai Statistics Center: (2014). Government of Dubai [13] Chen, G.; Warburton, R. (2006). “Do Speed Cameras Produce
website. Retrieved from http://www.dsc.gov.ae/EN/NewsMe Net Benefits? Evidence from British Columbia, Canada.”
dia/NewsReleases/Pages/Detail.aspx?NewsId=85. Journal of Policy Analysis and Management, 25(3), 661-678.
doi:10.1002/pam.20196.
[7] Perez, K.; Mari-Dell'Olmo, M; Tobias, A.; Borrell, C. (2007).
"Reducing Road Traffic Injuries: Effectiveness of Speed [14] Carroll McCormick. (2010). “Is Photo Radar Working?”
Cameras in an Urban Setting." American Journal of Public Truck News, 30(12), 20.
Health 97(9): 1632-1637.
[15] Hajbabaie, A.; Ramezani, H.; Benekohal, R. (2011) “Speed
[8] Oliver, S.; Gerno, L. (2013). “A Smart Camera Based Traffic Photo Enforcement Effects on Headways in Work Zones.”
Enforcement System - Experiences from the Field.” Proceedings of the 1st Congress of the Transportation and
Proceedings of SPIE-IS and T Electronic Imaging - Video Development Institute of ASCE, p 1226-1234.
Surveillance and Transportation Imaging.
[16] Liu, S.; Sun, X.; He, Y. (2010) “Analysis of Speed Limit
[9] Walter, L; Broughton, J.; and Knowles, J. (2011). “The Effect of Automatic Enforcement.” Proceedings of the 2010
Effects of Increased Police Enforcement Along a Route in International Conference of Logistics Engineering and
London.” Accident Analysis and Prevention, 43(3), Management, v 387, 1836-1843.
1219-1227.
[17] Zhang, X.; Liu, P.; Huang, F.; Yu, H. (2011) “Evaluation of
[10] Dowling, K. W.; Holloman, E. (2008). "The Effects of the Speed Reduction Effects of Automated Speed
Conspicuous Traffic Enforcement on Speeding Behaviors: a Enforcements on Freeways.” Proceedings of the 11th
Study of Speed Reduction Response." International Social International Conference of Chinese Transportation
Science Review 83(3-4): 181. Professionals, 2061-2071.
[11] Goldenbeld, C.; Van Schagen I. (2005). "The Effects of [18] Roess, R.; Prassas, E.; McShane, W. (2011) “Traffic
Speed Enforcement with Mobile Radar on Speed and Engineering.” Upper Saddle River, NJ: Pearson.
Accidents. An Evaluation Study on Rural Roads in the Dutch
Province Friesland." Accident Analysis and Prevention 37(6): [19] Devore, J.; Farnum, N. (1999). “Applied Statistics for
1135. Engineers and Scientists. Pacific Grove, CA: Brooks/Cole
Publishing Company.
[12] Kirsten, S.; James, C. (1994). "The Impact of Radar Cameras

You might also like