Path planning of automatic valet parking under mixed traffic scenarios in parking lot

Abstract

As a key technology for intelligent vehicles, autonomous valet parking is becoming increasingly popular in the area of research. Autonomous valet parking technology is available for safe and quick parking operations without a driver, and improving the driving convenience while greatly reducing the probability of parking accidents. A path planning method of autonomous valet parking in parking lots mixed traffic scenarios is proposed in this paper to resolve the following issues presented in the existing autonomous valet parking systems, that is, low efficiency of global path planning and low success rates of local obstacle avoidance. First, the parking lot traffic scenario model is established and analyzed. Secondly, a global path planning method based on Floyd algorithm is proposed and optimized. Then, the obstacle avoidance path is decoupled laterally and longitudinally based on the fifth-order polynomial in the Frenet coordinate system, and the optimal local path is screened out through the coupling loss function and defined constraints. Finally, a simulation scenario was built to verify the path planning method studied in this article. The results show that the autonomous valet parking path planning method studied in this paper can achieve good results and has strong robustness.