Vehicle Motion Planning in Complex Environment via Decomposition and Convexification
Ruishuang Chen, Jie Cheng, Zhihui Liang, Shuan Ding, Zaiyue Yang · IEEE Internet of Things Journal · 2023
For autonomous driving, vehicle motion planning in complex environment is always a classic and intricate optimization problem which is highly nonlinear and nonconvex. The difficulty of problem lies in the complicated obstacle avoidance constraints and their coupling with dynamics constraints. In this article, an efficient algorithm framework is proposed based on alternating direction method of multipliers (ADMMs) and convex feasible set (CFS) algorithm. The ADMM is employed to decompose the original problem into two relatively simple subproblems containing only dynamics constraints and obstacle avoidance constraints, respectively, while the CFS is utilized to convexify the feasible region based on geometrical properties and transform the subproblem with only obstacle avoidance constraints into a series of quadratic programming (QP) problems. The combination of ADMM and CFS decouples the complicated constraints and enables the proposed algorithm to efficiently solve motion planning problem in complex environment with less dependence on the initial guess. The convergence of the proposed algorithm is proven, while the numerical evaluation results in different cases verify the feasibility and effectiveness of the proposed algorithm framework.