- Email: [email protected]
Experimental research on the effectiveness of speed reduction markings based on driving simulation: A case study
Accident Analysis and Prevention 60 (2013) 211–218
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Experimental research on the effectiveness of speed reduction markings based on driving simulation: A case study Han Ding a,∗ , Xiaohua Zhao a , Jian Rong a , Jianming Ma b a b
Key Lab of Traffic Engineering, Beijing University of Technology, 100 Pingleyuan, Chaoyang District, Beijing 100124, China Texas Department of Transportation, 10016 Liriope Cove, Austin, TX 78750, USA
a r t i c l e
i n f o
Article history: Received 21 March 2013 Received in revised form 1 July 2013 Accepted 10 August 2013 Keywords: Speed reduction markings Driving simulator Relative speed difference Decelerating maneuver
a b s t r a c t The objective of this paper is to test the effectiveness of speed reduction markings (SRMs). The empirical data including vehicle operations and drivers’ psychological and physical reactions were collected in a driving simulator. An index—the relative speed difference—was developed to evaluate the effectiveness of speed reduction markings. Researchers analyzed data to test the effectiveness of speed reduction markings and quantitatively study their effects and the deceleration process. Out of 15 subjects, there were 1 (7%), 2 (13%), and 7 (47%) subjects who decelerated while driving through the downhill section in the No SRMs scenario, Longitudinal SRMs scenario and Transverse SRMs scenario, respectively. In terms of speed and subjects’ decelerating maneuver, the results of the analysis of variance (ANOVA) and the contrast analysis (S–N–K method) showed that transverse speed reduction markings (TSRMs) could significantly influence subjects’ speed choice and maneuver (P < 0.05). © 2013 Elsevier Ltd. All rights reserved.
1. Introduction With the growth of the national economy and the improvement of the Chinese people’s standard of living, the number of vehicles has increased rapidly in recent years. Data from official reports indicated that there were 125 million registered vehicles in China in 2006, and this number has reached 191 million in 2010 (National Bureau of Statistics of China, 2011). In the meantime, traffic crashes in China have contributed to a staggering amount of property damage as well as numerous deaths. According to an official report in China, in 2010 police reported 39,060,164 traffic crashes, resulting in 65,225 fatalities and 254,075 injuries and causing a direct economic loss of about ¥930 million (Road Traffic Crashes of China, 2011). Speeding, drunk driving and driver fatigue were the three main factors leading to traffic crashes; speeding alone was responsible for 14.2 percent of all traffic fatalities in 2010 (Zheng, 2007). Therefore, to better address this issue, installing speed control devices at sites with greater possibility of speedrelated traffic crashes is an important approach to reducing crashes and improving road safety. In China, speed control devices are classified into three categories: non-intrusive speed control devices, including automatic speed enforcement, speed limit signs and speed reduction
∗ Corresponding author. Tel.: +86 13717784903. E-mail addresses: [email protected] (H. Ding), [email protected] (X. Zhao). 0001-4575/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.aap.2013.08.007
marking; less intrusive speed control devices, such as thermoplastic vibration speed reduction markings; and intrusive speed control devices, including spikes, cement speed bumps, pavement groove strips and rubber speed bumps. Non-intrusive speed control devices, especially speed reduction markings (SRMs), have been used extensively on highways and urban expressways in China. According to the Chinese national standard Road Traffic Signs and Markings, SRMs are used to warn drivers of the need to reduce their speed (Standardization Administration of the People’s Republic of China, 2009). SRMs are placed on or in advance of horizontal or vertical curves, tunnels, or other featured roads where drivers need to slow down in advance. SRMs include longitudinal speed reduction markings (LSRMs) and transverse speed reduction markings (TSRMs); both can be audible and vibratory (Standardization Administration of the People’s Republic of China, 2009). The patterns of SRMs were shown in Fig. 1(a) and (b). The national standard describes detailed design requirements, including the size, length, width, spacing, as well as the placement location. However, there is relatively little research on the effectiveness of speed reduction markings on speed control in China and the current practices in their design and management operations may not be economical. In addition, SRMs used in other countries often differ in their design patterns as well as placement, and research results from other countries may not be applicable in China. Thus, this research project investigated the effectiveness of SRMs in order to better improve their application in China.
212
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
Fig. 1. Speed reduction markings in China and the U.S.
2. Literature review The detailed design, application, placement, guidance, options and support provisions for SRMs can be found in the U.S. 2009 Manual on Uniform Traffic Control Devices (the 2009 MUTCD). According to the 2009 MUTCD, SRMs (see Fig. 1(c)) are transverse markings that are placed on the roadway within a lane (along both edges of the lane) in a pattern of progressively reduced spacing to create the illusion that drivers are driving faster than they really are, thus persuading them to slow down (Federal Highway Administration, 2009). Compared Fig. 1(c) to Fig. 1(a) and (b), it is obvious that the
pattern of SRMs (which are classified as transverse markings) in the 2009 MUTCD is similar only in being transverse, otherwise not to LSRMs in China. Unlike speed reduction markings in the 2009 MUTCD, TSRMs in China traverse the lanes, so that drivers may feel slight vibration due to the elevated transversal bars created by the thermoplastic paint. Therefore, it is unclear whether the speed reduction is attributable to the distorted perception or the vibration they create. Meyer (1999) examined the use of optical speed bars to reduce speed and speed variation in highway work zones. The optical speed bars are innovative traffic control devices which consist of
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
a series of transverse stripes. The pattern and function of optical speed bars is similar to the SRMs regulated in the 2009 MUTCD. The research investigated the effect of optical speed bars on drivers’ speed, the underlying mechanism, and the usefulness of the technique to work zones. Experimenters obtained subjective evaluations of the relative effectiveness by developing simulations of various designs of optical speed bars. It was found that optical speed bars cause a small reduction in both speed and speed variation (Meyer, 1999) through their warning and perceptual effects. Godley et al. (2002) used a driving simulator as a validation tool for speed research. In their study, among mature drivers, 24 subjects drove an instrumented car and 20 drove in the simulator in two separate experiments. The experimental roads in the simulator and in the real world contained transverse rumble strips at three sites, as well as three equivalent control sites, and transverse rumble strips were placed in advance of stop sign-controlled intersections, right curves, and left curves. It was concluded that the simulator was a valid tool in evaluating speed related countermeasures. Gates et al. (2008) studied the effectiveness of experimental transverse-bar pavement marking as a speed-reduction treatment on freeway curves. A before-and-after analysis of speed was performed to determine the short- and long-term effectiveness of an experimental transverse-bar pavement marking treatment. The experimental transverse pavement marking was installed. In the research, researchers performed an analysis of variance (ANOVA) to examine the mean speeds; they found that curve speeds were effectively reduced, especially shortly after installation of the experimental transverse pavement marking treatment. Jiang et al. (2010) analyzed the effectiveness and adaptability of typical speed control devices, including speed limit signs, rumbles strips and speed bumpers, according to the experiment data collected on real roads. The results showed that speed bumpers were most effective in reducing speeds. The speed limit signs were the least effective but had the most long-lasting effect. In terms of adaptability, speed limit signs rendered the lowest level of tension and discomfort, and speed bumpers resulted in the highest level of tension and discomfort. Wang et al. (2009) observed the effectiveness of vibratory speed reduction bars in controlling speed by recording the speeds of vehicles traveling through vibratory speed reduction bars. The results showed that passenger cars had a larger speed reduction (5 times) than heavy vehicles. Bao et al. (2010) examined the practical application of speed reduction markings from both drivers and research data perspectives. They concluded that speed reduction markings significantly controlled speed in practical applications, which may be an effective way of preventing speed-related traffic crashes. As can be seen from the previous research, there are differences in definitions, design, and installation methods between China and other countries. There is relatively little research about the effectiveness of speed reduction markings that have been implemented in China. Even though some research using data collected on real roadways has been used to examine drivers’ deceleration behavior due to speed reduction measures, the results are inconclusive, as isolating the speed reduction effect of SRMs is difficult. Therefore, this paper will report research on the effectiveness of SRMs based on the driving simulator study.
3. Methods 3.1. Subjects To ensure the validity of data and avoid effects of character difference among the subjects, the current study focuses on
213
Fig. 2. Driving simulator.
relatively young male drivers. A total of fifteen male subjects, ranging in age from 21 to 41 years, with an average age of 25.9 years and an average of 4 years driving experience were recruited through advertisement. The experimental procedures and potential risks associated with the study were explained to each subject, who provided a written informed consent before participating in the driving simulation experiment.
3.2. Apparatus A fixed-base driving simulator (Fig. 2) was used in the current study, through which real-time data were collected, including operating data of vehicles (speed, acceleration, etc.) and maneuver data of subjects (gears, clutch, accelerator, brake, etc.). The data acquisition frequency of the driving simulator is 30 Hz. In addition, the virtual scenario was projected onto three large screens, providing a 130◦ field of view.
3.3. Scenarios Although the national standard has explicitly specified the SRMs should be installed at sites where the speed reduction is necessary, the crash history plays an important role in deciding where to deploy the SRMs. There are 231 interchanges in the urban area of Beijing. The SRMs have been implemented on the downhill section of interchanges, which are designed at a grade ranging from 1.5% to 2%. The Majialou Interchange on the South 4th Ring Road in Beijing was selected as the study site because this typical interchange has a crest vertical curve at a 1.7-percent grade, with SRMs installed on the downhill section. The design speed of the roadway is 120 km/h, and the posted speed limit is 80 km/h. Three virtual scenarios based on that interchange were developed, and referred to No SRMs scenario, Longitudinal SRMs scenario, and Transverse SRMs scenario, respectively (Fig. 3(a)). All other factors were identical in the three scenarios and no other simulated vehicles were included in the scenarios to eliminate the disturbance to subjects. To better represent the real roadway environment, the roadway geometric design parameters of the interchange were collected using GIPSI-Trac Road Geometry System. Based on these parameters, virtual scenarios were developed. The scenario length is 3.72 km; the interchange height is 8.6 m; the length of the downhill section is 521 m; and the lane width is 3.75 m. The elevation profile of the interchange is shown in Fig. 3(b) and the detailed designs of two types of SRMs are shown in Fig. 3(c–e), and the values of design parameters are measured in centimeters.
214
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
3.4. Procedures The drivers were given a tutorial on using the driving simulator, including the steering wheel, pedals and manual gears, and they were aware that they could drive at any speed that they preferred. The speed readings were shown on the speedometer. After the tutorial, each subject was required to fill out a questionnaire at pre-test, which recorded the basic information (such as age, gender, driving experiences, and so on) as well as the subject’s physiological and psychological status, including how long he/she slept, whether
he/she was exhausted now, whether he/she had drunk tea or coffee, and so on. Then subjects were required to perform a test drive for 5–10 min on a specific alignment to familiarize them with the driving simulator. To avoid driving fatigue, the experiment was divided into three sessions: the morning session (8:30–11:30 am), the afternoon session (3:00–5:00 pm), and the evening session (6:30–9:00 pm). Each test in one session lasted about 40 min per driver, and each driver drove in three scenarios characterized by speed reduction markings in one test. Besides, each scenario will be driven on for three
Fig. 3. Scenarios and designs of SRMs.
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
215
Fig. 3. (Continued)
times in a full experiment. The order of the scenarios was randomly selected among subjects to avoid the influence of the order. The data of the downhill section in each scenario would be analyzed. Data was saved each time when a subject completed the driving task in each scenario. The simulator records the speed, acceleration, as well as the number of times and the pressure of the driver applying the accelerator and brake pedal, therefore allowing identification of the beginning of the deceleration maneuver as well. When each full experiment was finished, every subject was required to fill out a questionnaire to report his subjective evaluation of the driving simulator and speed reduction markings, together with the physiological and psychological status at posttest. 4. Analysis and results 4.1. Subjective questionnaires In this experiment, a part of the questionnaire asked subjects to subjectively evaluate the performance of the driving simulator. The evaluated items included the brake, accelerator, scenarios, SRMs, and so on. The ratings ranged from 0 (stands for “not at all similar to the real world”) to 10 (stands for “extremely similar to the real world”). The results are shown in Table 1. Another part of the questionnaire was designed for participants’ subjective evaluation of the effectiveness of SRMs. Eighty-seven percent of subjects thought they were influenced by SRMs and 86 percent of subjects (6 percent totally; 53 percent strongly; 27 percent partially) agreed that SRMs can significantly reduce their vehicle speed (Fig. 4). Observing the speed and maneuver data, we found that all of those who totally or strongly agreed that SRMs can reduce speed had braked or released accelerator much more in SRMs scenarios, although the change of speed was not obvious. This is because a vehicle’s speed when going downhill tends to increase. SRMs seem to make drivers let up on the accelerator and brake more to counter against the force of the gravity. 4.2. Vehicle operating data The average acceleration in the downhill section was used to compare the three scenarios. There were 1 (7%), 2 (13%), and 7 (47%) subjects who decelerated while driving through the downhill
section in the No SRMs, Longitudinal SRMs and Transverse SRMs scenarios, respectively (Table 2). These results demonstrate that SRMs, especially TSRMs in the Transverse SRMs scenario, had a greater influence on subjects’ choice of reducing speeds. To quantitatively evaluate the effect of SRMs, the difference between the entering speed and the exiting speed was used to evaluate the effectiveness of SRMs. As such, the relative speed difference was defined: v2 − v1 = (1)
v1
where v1 is the entering speed, denoting the vehicle speed upon entering SRMs; v2 is the exiting speed, denoting the vehicle speed upon leaving SRMs; and is the relative speed difference. The relative speed difference can reflect the degree of change of vehicle’s speed relative to the entering speed; this index is used to represent the impact of SRMs. When the value of the ratio is negative, the entering speed is higher than the exiting speed, indicating that the speed declines, and vice versa. The average relative speed differences in both the No SRMs (MEAN = 0.10, SD = 0.10) and Longitudinal SRMs scenarios (MEAN = 0.09, SD = 0.14) were positive, which means that there was an increase in speed on the downhill section for both scenarios. The rate of speed increase in the Longitudinal SRMs scenario dropped nearly 10% lower compared to the one in the No SRMs scenario. However, the average relative speed difference was −0.02 in the Transverse SRMs scenario (MEAN = −0.02, SD = 0.16). This suggests that TSRMs tend to reduce speeds to some degree. The ANOVA was used to evaluate the effectiveness of SRMs in terms of speed reduction. Results of the ANOVA showed a significant main effect of the scenario on the relative speed difference (F(2,28) = 6.889; P = 0.004). A contrast analysis (S–N–K method) revealed that the mean relative speed difference in the Transverse SRMs scenario was significantly lower than that in the other two scenarios (P < 0.05 in all contrasts). The test result suggested that LSRMs did not have a significant effect on reducing speed, while TSRMs had a significant effect on reducing speed. Apart from evaluating the effectiveness of SRMs, it is interesting to further observe how the SRMs, especially TSRMs affect the subjects’ deceleration process. To address this issue, speed data were collected at seven equally spaced locations from the top to the bottom of the downhill section. The data were plotted on seven boxplot diagrams respectively in Fig. 5. The median speed in the No SRMs
216
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
Table 1 Ratings of subjective evaluation. Evaluated item
General level
Steering wheel
Accelerator
Brake
Clutch
Gears
Scenarios
Speed perception
SRMs
Average ratings Standard deviation
8 0.8
8 1.2
8 1.8
8 1.5
9 0.7
7 1.6
8 1.2
7 1.8
8 0.7
Fig. 4. Results of questionnaires.
scenario kept increasing on the downhill section; the median speed in the LSRMs scenario kept increasing until reaching the middle point of the downhill section, and then maintained that speed for the rest of the downhill section. In the Transverse SRMs scenario, the median speed dropped significantly on the first half of the section and maintained the lower speed until reaching the bottom of the downhill section. The different styles of driving speed changes in the three scenarios suggest that TSRMs were most effective in encouraging drivers to reduce speed. It was concluded that both LSRMs and TSRMs could influence subjects’ speed choice to different degrees, and the TSRMs had a statistically significantly better impact on reducing driver’s speed and encouraging more drivers to decelerate than the LSRMs. 4.3. Subjects’ maneuver data The perception of the road and traffic environment and the processing of information enable drivers to judge and make
decisions, resulting in the manipulation of vehicles (the perceptionjudgment-manipulation mechanism) (Ren et al., 2008). Based on the literature review, LSRMs give drivers the impression that the lane is narrowing, and TSRMs make them feel their speed is increasing, which potentially causes them to slow down. Therefore, it is necessary for researchers to evaluate the effectiveness in terms of subjects’ maneuvers. In this paper, the subjects’ decelerating maneuver is denoted as the accelerator and brake pedal responses. The maneuver data includes the number of times each driver presses the brake pedal and releases the accelerator and the total number of times for both maneuvers. The calculated results are presented in Fig. 6. It was notable that the effect of decelerating by pressing the brake pedal was different from decelerating by releasing the accelerator in the driving simulator. According to a research report (Wu et al., 2012), pressing the brake pedal one time is equivalent to releasing the accelerator 2.5 times to achieve same degree of deceleration, and this factor was used in this experiment (Wu et al., 2012). Among all the groups, the
Table 2 Average accelerations in different scenarios (m/s2 ).
Subjects No.
No SRMs scenario
Longitudinal SRMs scenario
Transverse SRMs scenario
J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 J11 J12 J13 J14 J15 Average Standard Deviation
0.53 0.07 0.02 0.02 0.11 0.24 0.24 0.08 0.20 0.13 0.16 0.23 0.30 0.20 -0.11
0.49 0.05 -0.37 0.47 0.02 0.22 0.21 0.07 0.07 0.04 0.08 -0.27 0.45 0.13 0.23 0.13 0.24
0.24 -0.15 -0.26 -0.35 -0.53 0.38 0.18 0.25 0.04 -0.11 0.10 0.03 -0.27 -0.10 0.07 -0.03 0.25
0.16 0.14
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
217
200 180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
2466
Distance (m) Fig. 6. Numbers of times of maneuver in different scenarios.
(a) No SRMs scenario 200 180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
2466
Distance (m)
(b) Longitudinal SRMs scenario
5. Discussion
180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
Distance (m)
(c) Transverse SRMs scenario Fig. 5. Speed distributions in three scenarios.
Transverse SRMs scenario showed the lowest number of times of releasing the accelerator but the highest number of times of pressing the brake pedal, which was 6 times that in the No SRMs scenario and 3 times that in the Longitudinal SRMs scenario. The total number of maneuvers in the Transverse SRMs scenario was the highest as well, which was 1.9 times that in the No SRMs scenario, and 1.7 times that in the Longitudinal SRMs scenario. This implies that TSRMs had a better effect on subjects’ willingness to decelerate and deceleration behavior than LSRMs. ANOVA was used to evaluate the effectiveness of SRMs in terms of subjects’ maneuvers. Results of ANOVA showed a significant main effect of the scenario on the total number of maneuvers (F(2,28) = 3.767; P = 0.036). A contrast analysis (S–N–K method) revealed that the total number of maneuvers in the Transverse SRMs scenario was significantly higher than the total number of maneuvers in the other two scenarios (P < 0.05 in all contrasts). The test result suggested that LSRMs had an effect on subjects’ deceleration willingness and maneuvering to a limited degree, while TSRMs had a significant effect on subjects’ deceleration willingness and maneuvering.
2466
The objective of this paper is to evaluate the effectiveness of SRMs defined in the China Road Traffic Signs and Markings Standard (GB5768-2009), with respect to vehicle operation and subjects’ maneuvering. The experiment was conducted in a driving simulator. Regarded as a valid tool in evaluating speed-related countermeasures, the driving simulator allowed us to better control the various impacting factors so as to minimize the confounding effect. Moreover, an experiment based on driving simulators is usually safer and more cost-effective than naturalistic driving research, and data collection is relatively easy. However, there are also some limitations with driving simulators or experimental procedures, which may influence the results, as well as the conclusions. The current study investigated the effectiveness of SRMs on a downhill section based on an urban expressway (an interchange on the South 4th Ring Road in Beijing). Since there are few sharp horizontal curves on the Beijing expressway system, this paper does not evaluate the effects of SRM on horizontal curves. But evaluating the speed reduction effects of SRM installed on horizontal curves is definitely an interesting topic for future research. Numerous studies state that driving performance could be affected by age and gender. In this study, for the purpose of assessing the effectiveness of speed reduction markings, a homogeneous group of subjects was recruited to avoid the bias resulting from sample heterogeneity. However, this may limit our results and
218
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
conclusions only to special populations. Female drivers in China account for about 19 percent of all licensed drivers (China, 2012) and there are few older drivers in China. However, testing other categories of drivers (female and older subjects) in the future will be able to explore the impact of gender and age on driving behavior in China. What’s more, we will increase sample sizes in our future study, which may lead to better results or conclusions. Data in this paper were collected in a driving simulator based on virtual scenarios. No speed limits were set in the virtual experiment so that the speed reduction effect of the SRMs could be evaluated. Therefore, the conclusions of the effectiveness are relative. The effectiveness of SRMs in the real world must be tested and the performance of driving simulators should be validated. The validation of the driving simulator should be performed in future research through field tests. As shown in Table 1, the subjective ratings show that the majority of subjects think that the simulator is effective. In this paper, the number of times each driver pressed the brake pedal and released the accelerator was used to evaluate the effectiveness of SRMs. However, some subjects pressed the brake pedal for a long time when they were decelerating, while others pressed the brake pedal for a short time; some subjects pressed the brake pedal gently, while others pressed the brake pedal heavily. Similar situations happened for the maneuver of releasing the accelerator. Therefore, future research should consider the efficiency of maneuvers. 6. Conclusions This paper evaluated the effectiveness of SRMs with respect to vehicle operation and subjects’ maneuvering. By analyzing the data collected in a driving simulator, the following conclusions can be reached: • According to the questionnaires, almost all subjects were influenced by SRMs, and the majority of subjects agreed that SRMs made them slow down. • SRMs had an effect on subjects’ speed choice, and TSRMs made more subjects decelerate than LSRMs; the magnitude of speed reduction due to TSRMs is significantly greater than that due to LSRMs. • Under this experimental environment, TSRMs reduces vehicles’ speed primarily on the first half of the downhill section; the speed drop reached 10 km/h at the location of 300 m from the beginning of the 521 m downhill section. • The analytical results indicated that SRMs, particularly TSRMs, had a significant effect on (1) the number of subjects decelerating, (2) the deceleration rate, and (3) accelerator/brake pedal response. • Under this experimental environment, subjects accelerated while driving through the downhill section in the No SRMs scenario, in which there were no SRMs or other speed control
devices at all. Compared to this, LSRMs could slightly encourage drivers to decelerate and manipulate their vehicles accordingly, yet the effect was not statistically significant in the current research. TSRMs had a significantly greater impact than LSRMs in terms of making drivers slow down. Therefore, speed reduction markings, especially transverse speed reduction markings, were recommended for implementation on roadways similar to this experimental environment in order to reduce speeding related traffic crashes and improve safety. Acknowledgments This study is supported by the NNSFC project: The Study of the Mechanism for Traffic Signs Influence on Driving Behavior and Its Cognitive Model, No. 51108011; and the BJNSF project: The Study of Identification Method of Drunk Driving Based on Driving Behavior of Driving Personal Character, No. 8112004. The authors would like to thank the anonymous reviewers for their valuable and constructive comments that greatly contributed to improving the final version of the paper. References Bao, X.J., Sun, X.D., He., Y.L., 2010. Application research on freeway speed reduction markings. Journal of Transport Standardization (215/217), 86–88. 2012. China has 0.236 Billion Licensed Drivers and 0.225 Billion Registered Vehicles, Retrieved from http://news.mycar168.com/2012/02/263721.html (retrieved 30.05.12). Federal Highway Administration, 2009. Manual on Uniform Traffic Control Devices for Streets and Highways (2009 Edition). U.S. Department of Transportation. Gates, T.J.X., Qin, D.A., Noyce, 2008. Effectiveness of experimental transverse-bar pavement marking as speed-reduction treatment on freeway curves. In: Transportation Research Record: Journal of the Transportation Research Board, No. 2056. Transportation Research Board of the National Academies, Washington, DC, pp. 95–103. Godley, S.T., Triggs, T.J., Fildes, B.N., 2002. Driving simulator validation for speed research. Journal of Accident Analysis and Prevention 34 (5), 589–600. Jiang, J.Z., Wang, J., Wu, J., Lu., J., 2010. Effectiveness and adaptability analysis of typical speed control measures. Journal of Transport Information and Safety 28 (155), 96–99. Meyer, E., 1999. Application of Optical Speed Bars to Highway Work Zones. In: Transportation Research Record: Journal of the Transportation Research Board, No. 1657. Transportation Research Board of the National Academies, Washington, DC, pp. 48–54. National Bureau of Statistics of China, 2011. China Statistical Yearbook. China Statistics Press, Beijing. Ren, F.T., Liu, X.M., Rong., J., 2008. Traffic Engineering, 2nd ed. China Communications Press, Beijing. 2010. Road Traffic Crashes of China in 2010. The Ministry of Public Security of the People’s Republic of China, Beijing, http://www.mps.gov.cn/ n16/n85753/n85870/2758752.html (accessed 25.01.11). Standardization Administration of the People’s Republic of China, 2009. Road Traffic Signs and Markings (GB5768-2009). Beijing. Wang, C., Sun, X.D., Shi.F Y., 2009. Research on vibrate decelerate belt and its speeding controlling effectiveness. Journal of Western China (19), 10–13. Wu, Y.P., Zhao, X.H., Rong, J., Ma, J.M. 2012. The Effect of Chevron Alignment Signs on Driver Performance. Presented at the 3rd International Conference on Green Intelligent Transportation System and Safety, Lanzhou, China. Zheng, F., 2007. Research on Design and Application of Road Speed Control Facilities. Jilin University, Changchun (Master Thesis of Jilin University).
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Experimental research on the effectiveness of speed reduction markings based on driving simulation: A case study Han Ding a,∗ , Xiaohua Zhao a , Jian Rong a , Jianming Ma b a b
Key Lab of Traffic Engineering, Beijing University of Technology, 100 Pingleyuan, Chaoyang District, Beijing 100124, China Texas Department of Transportation, 10016 Liriope Cove, Austin, TX 78750, USA
a r t i c l e
i n f o
Article history: Received 21 March 2013 Received in revised form 1 July 2013 Accepted 10 August 2013 Keywords: Speed reduction markings Driving simulator Relative speed difference Decelerating maneuver
a b s t r a c t The objective of this paper is to test the effectiveness of speed reduction markings (SRMs). The empirical data including vehicle operations and drivers’ psychological and physical reactions were collected in a driving simulator. An index—the relative speed difference—was developed to evaluate the effectiveness of speed reduction markings. Researchers analyzed data to test the effectiveness of speed reduction markings and quantitatively study their effects and the deceleration process. Out of 15 subjects, there were 1 (7%), 2 (13%), and 7 (47%) subjects who decelerated while driving through the downhill section in the No SRMs scenario, Longitudinal SRMs scenario and Transverse SRMs scenario, respectively. In terms of speed and subjects’ decelerating maneuver, the results of the analysis of variance (ANOVA) and the contrast analysis (S–N–K method) showed that transverse speed reduction markings (TSRMs) could significantly influence subjects’ speed choice and maneuver (P < 0.05). © 2013 Elsevier Ltd. All rights reserved.
1. Introduction With the growth of the national economy and the improvement of the Chinese people’s standard of living, the number of vehicles has increased rapidly in recent years. Data from official reports indicated that there were 125 million registered vehicles in China in 2006, and this number has reached 191 million in 2010 (National Bureau of Statistics of China, 2011). In the meantime, traffic crashes in China have contributed to a staggering amount of property damage as well as numerous deaths. According to an official report in China, in 2010 police reported 39,060,164 traffic crashes, resulting in 65,225 fatalities and 254,075 injuries and causing a direct economic loss of about ¥930 million (Road Traffic Crashes of China, 2011). Speeding, drunk driving and driver fatigue were the three main factors leading to traffic crashes; speeding alone was responsible for 14.2 percent of all traffic fatalities in 2010 (Zheng, 2007). Therefore, to better address this issue, installing speed control devices at sites with greater possibility of speedrelated traffic crashes is an important approach to reducing crashes and improving road safety. In China, speed control devices are classified into three categories: non-intrusive speed control devices, including automatic speed enforcement, speed limit signs and speed reduction
∗ Corresponding author. Tel.: +86 13717784903. E-mail addresses: [email protected] (H. Ding), [email protected] (X. Zhao). 0001-4575/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.aap.2013.08.007
marking; less intrusive speed control devices, such as thermoplastic vibration speed reduction markings; and intrusive speed control devices, including spikes, cement speed bumps, pavement groove strips and rubber speed bumps. Non-intrusive speed control devices, especially speed reduction markings (SRMs), have been used extensively on highways and urban expressways in China. According to the Chinese national standard Road Traffic Signs and Markings, SRMs are used to warn drivers of the need to reduce their speed (Standardization Administration of the People’s Republic of China, 2009). SRMs are placed on or in advance of horizontal or vertical curves, tunnels, or other featured roads where drivers need to slow down in advance. SRMs include longitudinal speed reduction markings (LSRMs) and transverse speed reduction markings (TSRMs); both can be audible and vibratory (Standardization Administration of the People’s Republic of China, 2009). The patterns of SRMs were shown in Fig. 1(a) and (b). The national standard describes detailed design requirements, including the size, length, width, spacing, as well as the placement location. However, there is relatively little research on the effectiveness of speed reduction markings on speed control in China and the current practices in their design and management operations may not be economical. In addition, SRMs used in other countries often differ in their design patterns as well as placement, and research results from other countries may not be applicable in China. Thus, this research project investigated the effectiveness of SRMs in order to better improve their application in China.
212
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
Fig. 1. Speed reduction markings in China and the U.S.
2. Literature review The detailed design, application, placement, guidance, options and support provisions for SRMs can be found in the U.S. 2009 Manual on Uniform Traffic Control Devices (the 2009 MUTCD). According to the 2009 MUTCD, SRMs (see Fig. 1(c)) are transverse markings that are placed on the roadway within a lane (along both edges of the lane) in a pattern of progressively reduced spacing to create the illusion that drivers are driving faster than they really are, thus persuading them to slow down (Federal Highway Administration, 2009). Compared Fig. 1(c) to Fig. 1(a) and (b), it is obvious that the
pattern of SRMs (which are classified as transverse markings) in the 2009 MUTCD is similar only in being transverse, otherwise not to LSRMs in China. Unlike speed reduction markings in the 2009 MUTCD, TSRMs in China traverse the lanes, so that drivers may feel slight vibration due to the elevated transversal bars created by the thermoplastic paint. Therefore, it is unclear whether the speed reduction is attributable to the distorted perception or the vibration they create. Meyer (1999) examined the use of optical speed bars to reduce speed and speed variation in highway work zones. The optical speed bars are innovative traffic control devices which consist of
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
a series of transverse stripes. The pattern and function of optical speed bars is similar to the SRMs regulated in the 2009 MUTCD. The research investigated the effect of optical speed bars on drivers’ speed, the underlying mechanism, and the usefulness of the technique to work zones. Experimenters obtained subjective evaluations of the relative effectiveness by developing simulations of various designs of optical speed bars. It was found that optical speed bars cause a small reduction in both speed and speed variation (Meyer, 1999) through their warning and perceptual effects. Godley et al. (2002) used a driving simulator as a validation tool for speed research. In their study, among mature drivers, 24 subjects drove an instrumented car and 20 drove in the simulator in two separate experiments. The experimental roads in the simulator and in the real world contained transverse rumble strips at three sites, as well as three equivalent control sites, and transverse rumble strips were placed in advance of stop sign-controlled intersections, right curves, and left curves. It was concluded that the simulator was a valid tool in evaluating speed related countermeasures. Gates et al. (2008) studied the effectiveness of experimental transverse-bar pavement marking as a speed-reduction treatment on freeway curves. A before-and-after analysis of speed was performed to determine the short- and long-term effectiveness of an experimental transverse-bar pavement marking treatment. The experimental transverse pavement marking was installed. In the research, researchers performed an analysis of variance (ANOVA) to examine the mean speeds; they found that curve speeds were effectively reduced, especially shortly after installation of the experimental transverse pavement marking treatment. Jiang et al. (2010) analyzed the effectiveness and adaptability of typical speed control devices, including speed limit signs, rumbles strips and speed bumpers, according to the experiment data collected on real roads. The results showed that speed bumpers were most effective in reducing speeds. The speed limit signs were the least effective but had the most long-lasting effect. In terms of adaptability, speed limit signs rendered the lowest level of tension and discomfort, and speed bumpers resulted in the highest level of tension and discomfort. Wang et al. (2009) observed the effectiveness of vibratory speed reduction bars in controlling speed by recording the speeds of vehicles traveling through vibratory speed reduction bars. The results showed that passenger cars had a larger speed reduction (5 times) than heavy vehicles. Bao et al. (2010) examined the practical application of speed reduction markings from both drivers and research data perspectives. They concluded that speed reduction markings significantly controlled speed in practical applications, which may be an effective way of preventing speed-related traffic crashes. As can be seen from the previous research, there are differences in definitions, design, and installation methods between China and other countries. There is relatively little research about the effectiveness of speed reduction markings that have been implemented in China. Even though some research using data collected on real roadways has been used to examine drivers’ deceleration behavior due to speed reduction measures, the results are inconclusive, as isolating the speed reduction effect of SRMs is difficult. Therefore, this paper will report research on the effectiveness of SRMs based on the driving simulator study.
3. Methods 3.1. Subjects To ensure the validity of data and avoid effects of character difference among the subjects, the current study focuses on
213
Fig. 2. Driving simulator.
relatively young male drivers. A total of fifteen male subjects, ranging in age from 21 to 41 years, with an average age of 25.9 years and an average of 4 years driving experience were recruited through advertisement. The experimental procedures and potential risks associated with the study were explained to each subject, who provided a written informed consent before participating in the driving simulation experiment.
3.2. Apparatus A fixed-base driving simulator (Fig. 2) was used in the current study, through which real-time data were collected, including operating data of vehicles (speed, acceleration, etc.) and maneuver data of subjects (gears, clutch, accelerator, brake, etc.). The data acquisition frequency of the driving simulator is 30 Hz. In addition, the virtual scenario was projected onto three large screens, providing a 130◦ field of view.
3.3. Scenarios Although the national standard has explicitly specified the SRMs should be installed at sites where the speed reduction is necessary, the crash history plays an important role in deciding where to deploy the SRMs. There are 231 interchanges in the urban area of Beijing. The SRMs have been implemented on the downhill section of interchanges, which are designed at a grade ranging from 1.5% to 2%. The Majialou Interchange on the South 4th Ring Road in Beijing was selected as the study site because this typical interchange has a crest vertical curve at a 1.7-percent grade, with SRMs installed on the downhill section. The design speed of the roadway is 120 km/h, and the posted speed limit is 80 km/h. Three virtual scenarios based on that interchange were developed, and referred to No SRMs scenario, Longitudinal SRMs scenario, and Transverse SRMs scenario, respectively (Fig. 3(a)). All other factors were identical in the three scenarios and no other simulated vehicles were included in the scenarios to eliminate the disturbance to subjects. To better represent the real roadway environment, the roadway geometric design parameters of the interchange were collected using GIPSI-Trac Road Geometry System. Based on these parameters, virtual scenarios were developed. The scenario length is 3.72 km; the interchange height is 8.6 m; the length of the downhill section is 521 m; and the lane width is 3.75 m. The elevation profile of the interchange is shown in Fig. 3(b) and the detailed designs of two types of SRMs are shown in Fig. 3(c–e), and the values of design parameters are measured in centimeters.
214
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
3.4. Procedures The drivers were given a tutorial on using the driving simulator, including the steering wheel, pedals and manual gears, and they were aware that they could drive at any speed that they preferred. The speed readings were shown on the speedometer. After the tutorial, each subject was required to fill out a questionnaire at pre-test, which recorded the basic information (such as age, gender, driving experiences, and so on) as well as the subject’s physiological and psychological status, including how long he/she slept, whether
he/she was exhausted now, whether he/she had drunk tea or coffee, and so on. Then subjects were required to perform a test drive for 5–10 min on a specific alignment to familiarize them with the driving simulator. To avoid driving fatigue, the experiment was divided into three sessions: the morning session (8:30–11:30 am), the afternoon session (3:00–5:00 pm), and the evening session (6:30–9:00 pm). Each test in one session lasted about 40 min per driver, and each driver drove in three scenarios characterized by speed reduction markings in one test. Besides, each scenario will be driven on for three
Fig. 3. Scenarios and designs of SRMs.
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
215
Fig. 3. (Continued)
times in a full experiment. The order of the scenarios was randomly selected among subjects to avoid the influence of the order. The data of the downhill section in each scenario would be analyzed. Data was saved each time when a subject completed the driving task in each scenario. The simulator records the speed, acceleration, as well as the number of times and the pressure of the driver applying the accelerator and brake pedal, therefore allowing identification of the beginning of the deceleration maneuver as well. When each full experiment was finished, every subject was required to fill out a questionnaire to report his subjective evaluation of the driving simulator and speed reduction markings, together with the physiological and psychological status at posttest. 4. Analysis and results 4.1. Subjective questionnaires In this experiment, a part of the questionnaire asked subjects to subjectively evaluate the performance of the driving simulator. The evaluated items included the brake, accelerator, scenarios, SRMs, and so on. The ratings ranged from 0 (stands for “not at all similar to the real world”) to 10 (stands for “extremely similar to the real world”). The results are shown in Table 1. Another part of the questionnaire was designed for participants’ subjective evaluation of the effectiveness of SRMs. Eighty-seven percent of subjects thought they were influenced by SRMs and 86 percent of subjects (6 percent totally; 53 percent strongly; 27 percent partially) agreed that SRMs can significantly reduce their vehicle speed (Fig. 4). Observing the speed and maneuver data, we found that all of those who totally or strongly agreed that SRMs can reduce speed had braked or released accelerator much more in SRMs scenarios, although the change of speed was not obvious. This is because a vehicle’s speed when going downhill tends to increase. SRMs seem to make drivers let up on the accelerator and brake more to counter against the force of the gravity. 4.2. Vehicle operating data The average acceleration in the downhill section was used to compare the three scenarios. There were 1 (7%), 2 (13%), and 7 (47%) subjects who decelerated while driving through the downhill
section in the No SRMs, Longitudinal SRMs and Transverse SRMs scenarios, respectively (Table 2). These results demonstrate that SRMs, especially TSRMs in the Transverse SRMs scenario, had a greater influence on subjects’ choice of reducing speeds. To quantitatively evaluate the effect of SRMs, the difference between the entering speed and the exiting speed was used to evaluate the effectiveness of SRMs. As such, the relative speed difference was defined: v2 − v1 = (1)
v1
where v1 is the entering speed, denoting the vehicle speed upon entering SRMs; v2 is the exiting speed, denoting the vehicle speed upon leaving SRMs; and is the relative speed difference. The relative speed difference can reflect the degree of change of vehicle’s speed relative to the entering speed; this index is used to represent the impact of SRMs. When the value of the ratio is negative, the entering speed is higher than the exiting speed, indicating that the speed declines, and vice versa. The average relative speed differences in both the No SRMs (MEAN = 0.10, SD = 0.10) and Longitudinal SRMs scenarios (MEAN = 0.09, SD = 0.14) were positive, which means that there was an increase in speed on the downhill section for both scenarios. The rate of speed increase in the Longitudinal SRMs scenario dropped nearly 10% lower compared to the one in the No SRMs scenario. However, the average relative speed difference was −0.02 in the Transverse SRMs scenario (MEAN = −0.02, SD = 0.16). This suggests that TSRMs tend to reduce speeds to some degree. The ANOVA was used to evaluate the effectiveness of SRMs in terms of speed reduction. Results of the ANOVA showed a significant main effect of the scenario on the relative speed difference (F(2,28) = 6.889; P = 0.004). A contrast analysis (S–N–K method) revealed that the mean relative speed difference in the Transverse SRMs scenario was significantly lower than that in the other two scenarios (P < 0.05 in all contrasts). The test result suggested that LSRMs did not have a significant effect on reducing speed, while TSRMs had a significant effect on reducing speed. Apart from evaluating the effectiveness of SRMs, it is interesting to further observe how the SRMs, especially TSRMs affect the subjects’ deceleration process. To address this issue, speed data were collected at seven equally spaced locations from the top to the bottom of the downhill section. The data were plotted on seven boxplot diagrams respectively in Fig. 5. The median speed in the No SRMs
216
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
Table 1 Ratings of subjective evaluation. Evaluated item
General level
Steering wheel
Accelerator
Brake
Clutch
Gears
Scenarios
Speed perception
SRMs
Average ratings Standard deviation
8 0.8
8 1.2
8 1.8
8 1.5
9 0.7
7 1.6
8 1.2
7 1.8
8 0.7
Fig. 4. Results of questionnaires.
scenario kept increasing on the downhill section; the median speed in the LSRMs scenario kept increasing until reaching the middle point of the downhill section, and then maintained that speed for the rest of the downhill section. In the Transverse SRMs scenario, the median speed dropped significantly on the first half of the section and maintained the lower speed until reaching the bottom of the downhill section. The different styles of driving speed changes in the three scenarios suggest that TSRMs were most effective in encouraging drivers to reduce speed. It was concluded that both LSRMs and TSRMs could influence subjects’ speed choice to different degrees, and the TSRMs had a statistically significantly better impact on reducing driver’s speed and encouraging more drivers to decelerate than the LSRMs. 4.3. Subjects’ maneuver data The perception of the road and traffic environment and the processing of information enable drivers to judge and make
decisions, resulting in the manipulation of vehicles (the perceptionjudgment-manipulation mechanism) (Ren et al., 2008). Based on the literature review, LSRMs give drivers the impression that the lane is narrowing, and TSRMs make them feel their speed is increasing, which potentially causes them to slow down. Therefore, it is necessary for researchers to evaluate the effectiveness in terms of subjects’ maneuvers. In this paper, the subjects’ decelerating maneuver is denoted as the accelerator and brake pedal responses. The maneuver data includes the number of times each driver presses the brake pedal and releases the accelerator and the total number of times for both maneuvers. The calculated results are presented in Fig. 6. It was notable that the effect of decelerating by pressing the brake pedal was different from decelerating by releasing the accelerator in the driving simulator. According to a research report (Wu et al., 2012), pressing the brake pedal one time is equivalent to releasing the accelerator 2.5 times to achieve same degree of deceleration, and this factor was used in this experiment (Wu et al., 2012). Among all the groups, the
Table 2 Average accelerations in different scenarios (m/s2 ).
Subjects No.
No SRMs scenario
Longitudinal SRMs scenario
Transverse SRMs scenario
J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 J11 J12 J13 J14 J15 Average Standard Deviation
0.53 0.07 0.02 0.02 0.11 0.24 0.24 0.08 0.20 0.13 0.16 0.23 0.30 0.20 -0.11
0.49 0.05 -0.37 0.47 0.02 0.22 0.21 0.07 0.07 0.04 0.08 -0.27 0.45 0.13 0.23 0.13 0.24
0.24 -0.15 -0.26 -0.35 -0.53 0.38 0.18 0.25 0.04 -0.11 0.10 0.03 -0.27 -0.10 0.07 -0.03 0.25
0.16 0.14
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
217
200 180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
2466
Distance (m) Fig. 6. Numbers of times of maneuver in different scenarios.
(a) No SRMs scenario 200 180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
2466
Distance (m)
(b) Longitudinal SRMs scenario
5. Discussion
180 160
Speed (km/h)
140 120 100 80 60 40 20 0
1945
2045
2145
2245
2345
2445
Distance (m)
(c) Transverse SRMs scenario Fig. 5. Speed distributions in three scenarios.
Transverse SRMs scenario showed the lowest number of times of releasing the accelerator but the highest number of times of pressing the brake pedal, which was 6 times that in the No SRMs scenario and 3 times that in the Longitudinal SRMs scenario. The total number of maneuvers in the Transverse SRMs scenario was the highest as well, which was 1.9 times that in the No SRMs scenario, and 1.7 times that in the Longitudinal SRMs scenario. This implies that TSRMs had a better effect on subjects’ willingness to decelerate and deceleration behavior than LSRMs. ANOVA was used to evaluate the effectiveness of SRMs in terms of subjects’ maneuvers. Results of ANOVA showed a significant main effect of the scenario on the total number of maneuvers (F(2,28) = 3.767; P = 0.036). A contrast analysis (S–N–K method) revealed that the total number of maneuvers in the Transverse SRMs scenario was significantly higher than the total number of maneuvers in the other two scenarios (P < 0.05 in all contrasts). The test result suggested that LSRMs had an effect on subjects’ deceleration willingness and maneuvering to a limited degree, while TSRMs had a significant effect on subjects’ deceleration willingness and maneuvering.
2466
The objective of this paper is to evaluate the effectiveness of SRMs defined in the China Road Traffic Signs and Markings Standard (GB5768-2009), with respect to vehicle operation and subjects’ maneuvering. The experiment was conducted in a driving simulator. Regarded as a valid tool in evaluating speed-related countermeasures, the driving simulator allowed us to better control the various impacting factors so as to minimize the confounding effect. Moreover, an experiment based on driving simulators is usually safer and more cost-effective than naturalistic driving research, and data collection is relatively easy. However, there are also some limitations with driving simulators or experimental procedures, which may influence the results, as well as the conclusions. The current study investigated the effectiveness of SRMs on a downhill section based on an urban expressway (an interchange on the South 4th Ring Road in Beijing). Since there are few sharp horizontal curves on the Beijing expressway system, this paper does not evaluate the effects of SRM on horizontal curves. But evaluating the speed reduction effects of SRM installed on horizontal curves is definitely an interesting topic for future research. Numerous studies state that driving performance could be affected by age and gender. In this study, for the purpose of assessing the effectiveness of speed reduction markings, a homogeneous group of subjects was recruited to avoid the bias resulting from sample heterogeneity. However, this may limit our results and
218
H. Ding et al. / Accident Analysis and Prevention 60 (2013) 211–218
conclusions only to special populations. Female drivers in China account for about 19 percent of all licensed drivers (China, 2012) and there are few older drivers in China. However, testing other categories of drivers (female and older subjects) in the future will be able to explore the impact of gender and age on driving behavior in China. What’s more, we will increase sample sizes in our future study, which may lead to better results or conclusions. Data in this paper were collected in a driving simulator based on virtual scenarios. No speed limits were set in the virtual experiment so that the speed reduction effect of the SRMs could be evaluated. Therefore, the conclusions of the effectiveness are relative. The effectiveness of SRMs in the real world must be tested and the performance of driving simulators should be validated. The validation of the driving simulator should be performed in future research through field tests. As shown in Table 1, the subjective ratings show that the majority of subjects think that the simulator is effective. In this paper, the number of times each driver pressed the brake pedal and released the accelerator was used to evaluate the effectiveness of SRMs. However, some subjects pressed the brake pedal for a long time when they were decelerating, while others pressed the brake pedal for a short time; some subjects pressed the brake pedal gently, while others pressed the brake pedal heavily. Similar situations happened for the maneuver of releasing the accelerator. Therefore, future research should consider the efficiency of maneuvers. 6. Conclusions This paper evaluated the effectiveness of SRMs with respect to vehicle operation and subjects’ maneuvering. By analyzing the data collected in a driving simulator, the following conclusions can be reached: • According to the questionnaires, almost all subjects were influenced by SRMs, and the majority of subjects agreed that SRMs made them slow down. • SRMs had an effect on subjects’ speed choice, and TSRMs made more subjects decelerate than LSRMs; the magnitude of speed reduction due to TSRMs is significantly greater than that due to LSRMs. • Under this experimental environment, TSRMs reduces vehicles’ speed primarily on the first half of the downhill section; the speed drop reached 10 km/h at the location of 300 m from the beginning of the 521 m downhill section. • The analytical results indicated that SRMs, particularly TSRMs, had a significant effect on (1) the number of subjects decelerating, (2) the deceleration rate, and (3) accelerator/brake pedal response. • Under this experimental environment, subjects accelerated while driving through the downhill section in the No SRMs scenario, in which there were no SRMs or other speed control
devices at all. Compared to this, LSRMs could slightly encourage drivers to decelerate and manipulate their vehicles accordingly, yet the effect was not statistically significant in the current research. TSRMs had a significantly greater impact than LSRMs in terms of making drivers slow down. Therefore, speed reduction markings, especially transverse speed reduction markings, were recommended for implementation on roadways similar to this experimental environment in order to reduce speeding related traffic crashes and improve safety. Acknowledgments This study is supported by the NNSFC project: The Study of the Mechanism for Traffic Signs Influence on Driving Behavior and Its Cognitive Model, No. 51108011; and the BJNSF project: The Study of Identification Method of Drunk Driving Based on Driving Behavior of Driving Personal Character, No. 8112004. The authors would like to thank the anonymous reviewers for their valuable and constructive comments that greatly contributed to improving the final version of the paper. References Bao, X.J., Sun, X.D., He., Y.L., 2010. Application research on freeway speed reduction markings. Journal of Transport Standardization (215/217), 86–88. 2012. China has 0.236 Billion Licensed Drivers and 0.225 Billion Registered Vehicles, Retrieved from http://news.mycar168.com/2012/02/263721.html (retrieved 30.05.12). Federal Highway Administration, 2009. Manual on Uniform Traffic Control Devices for Streets and Highways (2009 Edition). U.S. Department of Transportation. Gates, T.J.X., Qin, D.A., Noyce, 2008. Effectiveness of experimental transverse-bar pavement marking as speed-reduction treatment on freeway curves. In: Transportation Research Record: Journal of the Transportation Research Board, No. 2056. Transportation Research Board of the National Academies, Washington, DC, pp. 95–103. Godley, S.T., Triggs, T.J., Fildes, B.N., 2002. Driving simulator validation for speed research. Journal of Accident Analysis and Prevention 34 (5), 589–600. Jiang, J.Z., Wang, J., Wu, J., Lu., J., 2010. Effectiveness and adaptability analysis of typical speed control measures. Journal of Transport Information and Safety 28 (155), 96–99. Meyer, E., 1999. Application of Optical Speed Bars to Highway Work Zones. In: Transportation Research Record: Journal of the Transportation Research Board, No. 1657. Transportation Research Board of the National Academies, Washington, DC, pp. 48–54. National Bureau of Statistics of China, 2011. China Statistical Yearbook. China Statistics Press, Beijing. Ren, F.T., Liu, X.M., Rong., J., 2008. Traffic Engineering, 2nd ed. China Communications Press, Beijing. 2010. Road Traffic Crashes of China in 2010. The Ministry of Public Security of the People’s Republic of China, Beijing, http://www.mps.gov.cn/ n16/n85753/n85870/2758752.html (accessed 25.01.11). Standardization Administration of the People’s Republic of China, 2009. Road Traffic Signs and Markings (GB5768-2009). Beijing. Wang, C., Sun, X.D., Shi.F Y., 2009. Research on vibrate decelerate belt and its speeding controlling effectiveness. Journal of Western China (19), 10–13. Wu, Y.P., Zhao, X.H., Rong, J., Ma, J.M. 2012. The Effect of Chevron Alignment Signs on Driver Performance. Presented at the 3rd International Conference on Green Intelligent Transportation System and Safety, Lanzhou, China. Zheng, F., 2007. Research on Design and Application of Road Speed Control Facilities. Jilin University, Changchun (Master Thesis of Jilin University).