AGV Collision-Free Trajectory Optimization Considering Probability Constraints

This paper addresses the optimal collision-free trajectory planning problem for autonomous ground vehicles (AGVs) in the presence of probability constraints. Within a computational framework based on optimal control, a strategy employing a conservative approximation function is proposed to solve the nonlinear trajectory optimization problem for AGVs under probability constraints. This strategy utilizes an approximation function based on hyperbolic tangent functions to replace probability constraints in trajectory optimization with deterministic constraints expressed explicitly, thereby transforming the original AGV trajectory optimization problem with probability constraints into a parameterized nonlinear programming problem. Furthermore, it is proven that the optimal solution under this approximation strategy converges to the optimal solution of the original problem. Finally, numerical results validate the reliability and optimality of this strategy in solving collision-free trajectory optimization problems for AGVs under probability constraints.

1 Introduction

In recent years, Autonomous Ground Vehicles (AGVs) have garnered significant attention and research interest. Broadly, autonomous driving technology encompasses several key aspects [1]: self-positioning technology, environmental perception technology, trajectory planning, and motion control. Among these, trajectory planning stands as the decision-making core for AGVs, primarily tasked with designing feasible and safe maneuvering paths based on environmental information and task requirements to enable AGVs to accomplish missions [2].