Estimating Crash Risk Using a Microscopic Trac Model


Department of Civil Engineering,Sharif University of Technology


Abstract. In this research, a microscopic model is developed that combines car following and lane
changing models, describing driver behavior as a utility maximization process of drivers for reducing crash
risk and increasing speed. This model is simulated by a cellular automata simulator and compared with
the real data. It is shown that there is no reason to consider the model invalid for driver behavior in
basic segments of the freeways in Iran, under non-congested conditions. Considering that the uncertain
position of vehicles is caused by their acceleration or deceleration, a probability function is calibrated
for calculating the presence probability of vehicles in their feasible cells. By multiplying the presence
probability and impact of a crash, the crash risk of cells is calculated. An idea for estimating the crash
risk of vehicles is introduced, named total risk. Total risk is the sum of risks on the path of the considered
vehicles. It is shown that, when the di erence between vehicle characteristics such as brake deceleration
increases, crash risk also increases, and vice versa.