An intelligent pedestrian-vehicle collision avoidance strategy based on evolutionary game

Abstract

It is very important for intelligent vehicles to make reasonable collision avoidance strategy when facing crossing pedestrians. In this paper, an evolutionary dynamics model of the conflict between pedestrians and intelligent vehicles is proposed for the scenario that intelligent vehicles perceive pedestrian crossing street and predict that a collision may occur. Considering factors such as vehicle type and the radical degree of pedestrian behavior, values are assigned to the parameters of the game matrix of the conflict between pedestrians and intelligent vehicles. Then their evolutionary paths are analyzed to predict whether pedestrians will pass or not. On this basis, strategy is made on whether the intelligent vehicle will pass. This paper also simulates the evolutionary path according to the determined parameters. Results show that the method of making decisions on the behavior of intelligent vehicles according to the evolutionary process so as to avoid colliding with pedestrians crossing the street is reasonable and effective.