교차로 환경에서 확률 기반의 주변 차량 경로 예측 및 위험도 판단

Abstract

For ADAS technologies development, predicting the trajectory of surrounding vehicles and assessing their risk is important. In urban driving situations, especially at intersections, the behavior of surrounding vehicles is complicated, such as going straight, turning left, and turning right, which leads to a high degree of uncertainty. In this paper, we suggest a probabilistic vehicle trajectory prediction and risk assessment algorithm, which can reflect uncertainty of the vehicles. The proposed algorithm employed IMM filter, which is used to determine the suitability of multiple models by probability, to consider the uncertainty between straight and turn motion of the vehicles. The collision risk is determined based on the probability of the predicted trajectory and desired acceleration for collision avoidance is determined using the cost function. The proposed algorithm has been verified through DGPS measurements of two vehicles in an intersection environment.