A Path-Planning Algorithm for Automatic Parallel Parking

Path-planning is a key issue of automatic parking assist system due to the non-holonomic constraints. A shortest path algorithm for the parallel parking problem in a certain condition is proposed and proved. A feasible path-planning approach is presented to meet the requirement that the parking space is narrow and needs park iteratively by improving the shortest path. Considering several possibilities of collision with obstacles in the parking process, the parking region where the cars can park with no collision based on the proposed algorithm is designed. The proposed algorithm is verified combined with vehicle dynamics constraints under the limitation of the steering angle rotating speed in high-precision vehicle dynamics simulation software veDYNA. The presented simulation results clearly show that the proposed algorithm provides practical solutions for automatic parallel parking problems.