| The uncertainties of pedestrian mobility are important factors affecting the accuracy and robustness of an active pedestrian protection system.This study is to provide the means for probabilistic risk evaluation of pedestrian-vehicle collision by counting the uncertainties in pedestrian motion.The probability distribution of TTC(Time-To-Collision)is used to describe the risk of pedestrian-vehicle collision,considering the uncertainty of pedestrian mobility.The pedestrian is modeled by a first-order Markov model and White Gaussian Noise to characterize the stochastic properties in mobility according to field experiments of pedestrians crossing an uncontrolled road.Based on the pedestrian and vehicle movement models,a model of the pedestrian-vehicle distance is proposed,and two estimation algorithms are designed to calculate the minimal distance and TTC respectively.Numerous Monte Carlo(MC)simulations are carried out to study the mathematical characteristics of the minimal distance and TTC,the fitting functions derived from Gauss distribution are proposed,then the Unscented Transformation(UT)is employed to predict the probability density function of the minimal distance and TTC.5000 scenarios were selected randomly,and the unscented transformation algorithm was used to predict each condition.The results were compared with 10000 MC simulations.The experimental results show that the probability density distributions of TTC,minimal distance,and collision probability estimated by UT coincide with the MC simulations under various pedestrian-vehicle conflict scenarios,and the maximal deviation of collision probability is 1.61%.The UT method is about 600 times faster than the MC method(10,000 runs),the average calculation time of UT is 3.1 ms,which means that the proposed method has the potential for online application. |
