Modeling Vehicle Acceleration at Stop-Controlled Intersections

Introduction

A primary challenge in crash reconstruction and the development of collision avoidance systems is determining whether a specific collision was avoidable. To make this determination, investigators must understand the “time available”—the duration a driver has to respond to a hazard. In scenarios involving path intrusions at intersections, this requires precise knowledge of how long a vehicle on a minor road has been accelerating prior to impact. The purpose of this research was to gather naturalistic and experimental acceleration data to develop mathematical models that predict acceleration duration based on distance traveled. Historically, many models assumed a constant or linear acceleration, often neglecting the initial slow phase where a driver moves their foot from the brake to the accelerator. This study sought to address these gaps by modeling the non-linear nature of passenger vehicle acceleration specifically at two-way-stop controlled intersections.

Methodology

The researchers employed a two-part study design to capture both naturalistic behavior and high-precision instrumented data. In Part 1, a naturalistic study observed 244 unaware drivers at eleven rural, two-way-stop controlled intersections in Ontario, Canada. These intersections featured flat terrain and major road speed limits of 80 km/h. Data was collected via digital video recordings at 30 frames per second and analyzed using videogrammetry to track vehicle positions over the first three seconds of movement. In Part 2, 10 participants drove instrumented vehicles along a route that included four acceleration runs at similar intersections. The instrumented vehicles utilized 100 Hz accelerometers and GPS data loggers to provide high-resolution profiles for up to five seconds of acceleration. To ensure scientific rigor, the researchers validated the videogrammetry method against instrumented vehicle data, finding a high degree of accuracy (within 5% for durations of 3 seconds or more). Analytical methods involved trend analysis to establish power, quadratic, and cubic functions for time-distance, distance-time, and speed-time relationships.

Results

The study concluded that driver acceleration is a non-linear, three-phase process. Phase one is an initial slow period (averaging 0.9 seconds) as the driver transitions from the brake to the accelerator. Phase two involves more aggressive, near-linear acceleration reaching a peak at approximately 1.6 seconds. Phase three is characterized by a gradual tapering off as the driver nears their desired speed. Key findings indicated that driver acceleration is a matter of personal preference and intersection type rather than vehicle capability, as most drivers utilized only a fraction of their vehicle’s potential power. Specifically, the overall average acceleration for both studied phases was found to be 0.21 g. The researchers developed a power function as the most effective model for determining acceleration duration from a known distance. This is critical for reconstructionists, as the sources note that assuming a constant acceleration of 0.15 g (a common industry standard) can lead to a 13% overestimate of the time taken to travel 15 meters, potentially leading to incorrect conclusions regarding crash avoidability.

References

1. Long, G. (2000). Acceleration characteristics of starting vehicles (TRR 1737). Washington, DC: Transportation Research Board of the National Academies.
2. Bham, G. H., & Benekohal, R. F. (2002). Development, evaluation, and comparison of acceleration models. Washington, DC: Transportation Research Board, 81st Annual Conference.
3. Akcelik, R., & Beasley, M. (2001, December). Acceleration and deceleration models. Melbourne, AU: 23rd Conference on Australian Transportation Research.
4. Hong, Z. (2007). Normal acceleration characteristics of the leading vehicle in a queue at signalized intersections on arterial streets (Master’s thesis). Atlanta, GA: Georgia Institute of Technology.
5. American Association of State Highways and Transportation Officials [AASHTO] (2001). A policy of geometric designs for highways and streets. Washington, DC.
6. Harwood, D. W., et al. (1996). Intersection sight distance (NCHRP Report 383). Washington, DC: TRB National Research Council.
7. Wang, J., et al. (2004). Normal acceleration behavior of passenger vehicles starting from rest at all-way-stop-controlled intersections (Transportation Research Record 1883, pp. 158-166). Washington, DC: Transportation Research Board of the National Academies.
8. Bonneson, J. A. (1992). Modeling queued driver behavior at signalized junctions (Transportation Research Record 1365, pp. 99-107). Washington, DC: Transportation Research Board.
9. Happer, A. J., Peck, M. D., & Hughes, M. C. (2009). Analysis of left-turning vehicles at a 4-way medium-sized signalized intersection. SAE International Journal of Passenger Cars – Mechanical Systems, 2(1), 359-370.
10. Fugger, T. F., Jr., Wobrock, J. L., Randles, B. C., Stein, A. C., et al. (2001). Driver characteristics at signal-controlled intersections (SAE Technical Paper 2001-01-0045). SAE International.
11. Proctor, C. L., II, Grimes, W. D., Fournier, D. J., Jr., Rigol, J., Jr., et al. (1995). Analysis of acceleration in passenger cars and heavy trucks (SAE Technical Paper 950136). SAE International.
12. Kosaka, H., Hashikawa, T., Higashikawa, N., Noda, M., et al. (2007). On-the-spot investigation of negotiation patterns of passing cars without right of way at a non-signalized intersection (SAE Technical Paper 2007-01-3599). SAE International.
13. Yan, X., Radwan, E., & Guo, D. (2007). Effects of major-road vehicle speed and driver age and gender on left-turn gap acceptance. Accident Analysis & Prevention, 39, 843-852.
14. Shechtman, O., Classen, S., Stephens, B., Davis, E., Bendixen, R., Belchior, P., Sandhu, M., McCarthy, D., & Mann, W. (2006). The impact of intersection design on simulated driving performance of young and senior adults: Preliminary results. Topics in Geriatric Rehabilitation, 22(1), 27-35.
15. Muttart, J. W. (1996). Vehicle acceleration: Observations and test results. Accident Investigation Quarterly, 10.
16. Fricke, L. B. (1990). Traffic accident reconstruction. Evansville, IL: Northwestern University Traffic Institute.
17. Bartlett, W. (2009). Essay: Vehicle acceleration. http://mfes.com/accel.html.
18. Muttart, J. W. (2004). Estimating driver response times. In Noy & Karkowski (Eds.), Handbook of human factors in litigation (Ch. 14, pp. 14-1 – 14-24). Boca Raton, FL: CRC Press.
19. Bartlett, W. (2008). Monte Carlo analysis for accident reconstruction. Accident Investigation Quarterly, Winter, 18-22.

To better understand this non-linear movement, imagine a runner starting a sprint from the blocks: they don’t reach their maximum speed in the first millisecond; there is a distinct moment of preparation, followed by a powerful burst of energy, and finally a transition into a steady, sustainable pace. The sources demonstrate that drivers follow this same “wind-up and go” pattern when entering an intersection.