Driver Response Times to Side Road Path Intrusions from SHRP-2 Naturalistic Database

Introduction

In the current environment, where major manufacturers are seeking a better understanding of when to trigger an emergency response, it is crucial to analyze how humans react to emergency events. Understanding driver behavior allows for the design of systems that can assist drivers where they perform poorly or completely eliminate the need for human accident avoidance maneuvers. Research has consistently shown that driver response times are not fixed across all scenarios, but rather change based on the information available to the driver.

The focus of this research is angular two-vehicle crashes that occur at intersections. A review of the FARS dataset showed that, between 2014 and 2018, there was an annual average of 5,890 fatal crashes involving two vehicles in an angled collision, with approximately 68% of these occurring at an intersection. The study specifically concentrated on two path-intrusion crash types: Crash Type 1, where the Subject Vehicle (SV) is traveling straight and its path is intruded upon by a Principal Other Vehicle (POV) turning left from 90 degrees; and Crash Type 2, where the SV is traveling straight and its path is intruded upon by a POV crossing straight through from 90 degrees.

The goal of this research was to identify the onset trigger for an emergency response and the corresponding driver response times for these path intrusion events. Researchers aimed to identify the factors that most influence a change in drivers’ response times and to refine the accepted practice for measuring these times. The study hypothesized that driver response times would be different based on the onset methodology used.

Methodology

The study utilized a subset of the Strategic Highway Research Program 2 (SHRP-2) naturalistic driving dataset. The research requested crash and near-crash events where a driver moving straight through an intersection faced a hazardous situation.

Participant Characteristics and Data Collection: The initial query yielded 196 events, which were filtered to remove events where the SV was moving less than 20 km/h prior to the avoidance maneuver, where the SV was braking prior to the hazard, or which occurred in parking lots. The final study pool consisted of 169 crash and near-crash events, which included events from both Crash Type 1 and Crash Type 2. A total of 158 drivers were represented, with 79 female drivers and 77 male drivers. The sample size for the detailed analysis of response times was reduced to 101 events because the stop bar had to be visible in the SV camera for the specific measurement of Perception-Response Time (PRT). Time series data, including speed and acceleration, were obtained from the SV’s onboard data acquisition system and synchronized with the video.

Research Design and Analytical Methods: The research focused on measuring Perception-Response Times (PRTs), which began when the POV was at a specific onset position and ended when the driver decelerated at 0.4g or more (hard braking). This definition of PRT included the time from brake application until the required deceleration was achieved. The PRTs were measured starting from two primary onset events:

1. PRTS (Stop Bar): Measured when the POV was at the stop bar, or its equivalent position.
2. PRTV (First Visible): Measured from when the intruding vehicle was first visible within camera range.

Independent Variables: Independent variables tested for their influence on response times were categorized as: SV Driver (e.g., age, gender, response type), POV/Intruding Driver (e.g., starting from a stop, obscured status, speed), and Other External Factors (e.g., lighting, weather, traffic signal phase). POV speed was specifically categorized as fast (traveling from the stop line to the road edge in $\le$ 0.5 seconds), slow (traveling that distance in > 0.5 seconds), or stop. Statistical comparisons utilized t-tests for two treatments and ANOVA for multiple levels.

Results

The analysis of 101 events showed that driver response times varied significantly based on the characteristics and behavior of the intruding vehicle.

Impact of Onset Methodology: Measuring PRT from when the POV was first visible (PRTV) resulted in longer response times, some exceeding four seconds, because the duration of visibility did not necessarily indicate an immediate hazard. In contrast, PRTS (stop bar onset) measurements showed a more normal distribution of response times.

Influence of Intruder Speed and Stopping Behavior:

• Speed: PRTs were significantly faster when the POV entered the road at a fast speed (mean = 0.82s, SD= 0.76s) compared to when the POV traveled at a slow speed (mean=1.3s, SD=0.72s) or started from a stop [F(2,95)=9.391,p= 0.0001].
• Stopping Behavior: Driver response times were significantly slower when the POV started from a stop (M = 1.73s, SD =0.85 s) compared to when the POV did not start from a stop (M = 1.13s, SD = 0.77 s). This finding is consistent with previous research.

Influence of Visibility: Response times were significantly faster in events where the POV was initially obscured (PRT= 1.17s, SD=0.8s) compared to events where the view to the POV was not obscured (PRT = 1.58s, SD= 0.86s). This suggests that the sudden appearance of an intruder may have a startling effect on the SV driver, a trend observed in previous studies.

Non-Significant Factors: The analysis found that response times did not significantly differ between Crash Type 1 (turning POV) and Crash Type 2 (straight-through POV). Furthermore, driver responses were not significantly different when compared across SV driver age [F(2,94)=1.477, p=0.234], driver gender [t(96)=-0.68, p=0.49], driver response type [F(2,95)=1.22, p=0.297], lighting at the scene [F(2,95)=0.081, p=0.992], or weather [F(2,95)=1.302, p= 0.227].

References

1. Attalla, S., Toxopeus, R., Kodsi, S., & Oliver, M. (2018). Driver Response Time to Left-Turning Vehicles at Traffic Signal Controlled Intersections (No. 2018-01-0521). SAE Technical Paper.
2. Bärgman, J., Smith, K., & Werneke, J. (2015). Quantifying Drivers’ Comfort-Zone and Dread-Zone Boundaries in Left Turn across Path/Opposite Direction (LTAP/OD) Scenarios, Transportation Research Part F: Traffic Psychology and Behaviour 35.
3. Barrett, G., Kobayashi, M., & Fox, B. H. (1968). Feasibility of studying driver reaction to sudden pedestrian emergencies in an automobile simulator, Human Factors, 10, 19-26.
4. Broen, N. L., & Chiang, D. P. (1996). Braking response times for 100 drivers in the avoidance of an unexpected obstacle as measured in a driving simulator, Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting, 900.
5. Chang, J. S., Wu, Y. J., Liu, C. Y., Huang, H. S., & Hsieh, S. K. (2008). Driver reaction time and kinematics in simulated brake response to an unexpected cut-in or cut-out in curves, SAE Technical Paper 2008-01-0311.
6. D’Addario, P.M., & Donmez, B. (2019). The effect of cognitive distraction on perception-response time to unexpected abrupt and gradual onset roadway hazards, Accident Analysis and Prevention, 127, 177-185.
7. Dinakar, S., & Muttart, J. (2019). Driver Behavior in Left Turn across Path from Opposite Direction Crash and near Crash Events from SHRP2 Naturalistic Driving (No. 2019-01- 0414). SAE Technical Paper.
8. Dinakar, S., Muttart, J., Suway, J., Marr, J., Edewaard, D., & Enes, A. (2020). Driver Response Times to Side Road Path Intrusions from SHRP-2 Naturalistic Database. Crash Safety Research Center, LLC.
9. Filio, D., Dony, L., Gonzalez, & D., Olive, M. (2017). Comparison of Wrap Around Screens and HMDs on A Driver’s Response to An Unexpected Pedestrian Crossing Using Simulator Vehicle Parameters, Ninth International Driving Symposium on Human Factors in Driver Assessment, Training, and Vehicle Design.
10. Gibson, J. J., & Crooks, L. E. (1938). A theoretical field-analysis of automobile-driving. The American journal of psychology, 51(3), 453-471.
11. Hankey, J. M., McGehee D. V., Dingus, T. A., Mazzae, E. N., & Garrott, W. R. (1996). Initial driver avoidance behavior and reaction time to an unalerted intersection incursion, Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting.
12. Jurecki, R., & Stanczyk, T.L., (2011). The test methods and the reaction time of drivers, Eksploatacja I Niezawodnosc, 3, 16-23.
13. Lechner, D., & Malaterre, G. (1991). Emergency maneuver experimentation using a driving simulator. (SAE Paper No. 910016) Warrendale, PA: Society of Automotive Engineers.
14. Lerner, N. D., Huey, R. W., McGee, H., & Sullivan, A. (1996). Older driver PRT for intersection sight distance and object detection. Accident Investigation Quarterly, 9, 20-36.
15. Lubbe, N., & Davidson, J. (2015). Drivers’ comfort boundaries in pedestrian crossing: a study in driver braking characteristics as a function of pedestrian walking speed, Safety Science, 75, 100-106.
16. Lubbe, N., & Rosen, E. (2014). Pedestrian crossing situations: Quantification of comfort boundaries to guide intervention training, Accident Analysis and Prevention, 71, 261-266.
17. Maeda, Teruo, Namio Irie, Kunihiko Hidaka, and Hiroaki Nishimura. “Performance of Driver-Vehicle System in Emergency Avoidance.” SAE Transactions 86 (1977): 518-41.
18. Marshall, D., Brown, T., Boyle, L.N., & Wu, X. (2016). Connected vehicle alerts: One size doesn’t fit all scenarios, Transportation Research Board Annual Meeting
19. Mazzae, E. N., Baldwin, G. H. S., & McGehee, D. V. (1999). Driver crash avoidance behavior with ABS in an intersection incursion scenario on the Iowa driving simulator. (SAE paper No. 1999-01-1290) Warrendale, PA; SAE
    Mazzae, E. N., Barickman, F., Baldwin, G. H. S., & Forkenbrock, G. (1999). Driver crash avoidance behavior with ABS in an intersection incursion scenario on dry versus wet pavement. (SAE Paper No. 1999-01-1288) Warrendale, PA; Society of Automotive Engineers.
20. Muttart, J. W. (2003). Development and evaluation of driver response time predictors based upon meta-analysis. SAE transactions, 876-896.
21. Muttart, J. W. (2005). Quantifying driver response times based upon research and real-life data.
22. Muttart, J. W., Fisher, D. L., & Pollatsek, A. P. (2014). Hazard mitigation when approaching Obstructed Intersections in a Simulator, Washington, DC: Transportation Research Board of the National Academies.
23. Perron T., Chevennement J., Damville A., Mautuit C., Thomas C., & Le Coz J.Y. (1998). Pilot Study of Accident Scenarios on a Driving Simulator, Proceedings of the 16th ESV Conference, June 1-4, 1998, Paper no 98-S2-0-02, pp. 374-385
24. Smiley, A, Smahel, T., Donderi, D.C., Caird, J.K., Chisholm, S.L., Caird, J.K., Lockhart, J., & Teteris, E. (2007).The effects of cellphone and CD use on novice and experienced driver performance, Insurance Bureau of Canada.
25. Stanczyk, T.L., Jurecki, R., Jaskiewicz, M., Walczak, S., & Janczur, R. (2011). Researches on the Reaction of a Pedestrian Stepping into the Road from the3 Right Side from Behind an Obstacle Realized on the Track, Journal of KONES Powertrain and Transport, 18 (1), 615-622.
26. Toxopeus, R., Attalla, S., Kodsi, S., & Oliver, M. (2018). Driver Response Time to Cyclist Path Intrusions, SAE Technical Paper 2018-01-0531, 2018.
27. Toxopeus, R., Attalla, S., Kodsi, S., & Oliver, M. (2018) Driver Response Time to Midblock Crossing Pedestrians, SAE Technical Paper 2018-01-0514, 2018.