Convex Optimization Based Collision Avoidance Path Planning Method for Unmanned Ground Vehicles

Abstract

This paper investigates a collision avoidance path planning method for unmanned ground vehicles based on convex optimization. An optimal traveling trajectory is planned in real time based on known global environment information. Firstly, we modeled the unmanned ground vehicle particle dynamics. On this basis, the trajectory optimization problem model is established by combining obstacle constraints and endpoint constraints with minimum turning speed and turning angular speed as the performance index. Then, lossless and successive convexification of the original trajectory optimization problem is achieved using convex optimization methods, resulting in a series of continuous-time convex subproblems. Finally, we use the Matlab GPOPS solver to solve the convex subproblem and verify the correctness and effectiveness of the path planning method.