Path Planning for Autonomous Vehicles Based on Improved Artificial Potential Field
Dongsheng Hua, Tao He, Tao Li, Haibo Gu · 2024
Autonomous driving technology is developing rapidly. Path planning plays an important role in autonomous driving. Current APF (Artificial Potential Field) algorithms generate smooth and safe paths; however, they have drawbacks, including local optimal solutions and the possibility of unreachable goals. To address these issues, The APF algorithm will be improved. To tackle the issue of local optima, a tuning coefficient is introduced to optimize the potential field expression, allowing the car to escape from local optimum traps. For the issue of unreachable goals, a virtual vehicle method with bidirectional trajectories is put forward to avert the effect of barriers near the starting and target points. Finally, the modified APF algorithm in comparison with the traditional APF in a static obstacle environment, and simulation results are validated.