Trajectory planning for autonomous vehicle based on window-constrained Pearl model

Abstract

Speed guidance is a significant application in driver assistance system or driverless automated system. Considering the traditional Pearl model-based trajectory planning, this paper proposes an improved method based on window constraints to enhance the rationality at intersection positions. To solve the problem, where the slopes of traditional time-distance Pearl curve at the two boundaries generate relatively low values and which is not in accordance with the actual situation, a new window constraint is used to guide the vehicle passing through the intersection at a high speed. Regarding the new window constraints, together with the maximum speed, tolerating accelerated speed constraints, and traffic signal length constraints, a multi-variable and single objective optimization improved nonlinear programing scheme is applied to obtain the most comfort trajectory of passengers for speed guidance. Simulation results show that, the proposed method presents better performance in traveling time and trajectory smoothness compared with the traditional method under different initial vehicle speeds at upstream intersections and signal offset cases.