A Fast Stable Lane Change Path Planning Method Based On Hybrid Intelligent Algorithms

Abstract

To improve the solving speed and success rate of the lane change path planning problem for autonomous vehicles, a fast and stable planning method (FSDETS) is proposed in this paper. This lane change path planning scheme is formulated as an optimal control problem according to the three-segment lane change model. Then the B-spline curve is applied to smooth the path. The core contribution is that a method based on hybrid intelligent algorithms is proposed to solve the optimal control problem fast and stably. First, an extended differential evolution is proposed to quickly provide a tough initial guess for the subsequent search. Then two criteria are set to judge whether the tabu search can start. After satisfying them, tabu search seeks a better solution based on the initial guess. Finally, a 1000 cycles simulation experiment is run to verify its stability and real-time performance. Simulation results show that the average time consumption is 0.7417ms, the success rate is 99.97% and the average maximum curvature is 0.13947m−1. Besides, it is also compared with other intelligent optimization algorithms as well as the optimization solver MIDACO. The result shows that our method surpasses other algorithms in comprehensive performances.