A Lightweight Optimal Trajectory Planning for Smart Summon in Highly Complex and Irregular Parking Lot Scenarios

This paper focuses on the trajectory planning problem in highly complex and irregular parking lot scenarios. We formulate this problem as an optimal control problem (OCP) according to the requirements of the task and propose a planner to solve it numerically. To ensure optimal planning results, the proposed planner consists of three procedures, which are responsible for generating the guiding route, constructing the collision-free tunnel, and optimizing the trajectory, respectively. Firstly, we use the generalized Voronoi graph (GVG) to build an equidistant roadmap and propose an improved adaptive A* algorithm (IAA*) to improve the time efficiency of the guiding route generation process. Secondly, we employ a coarse trajectory to guide a homotopic route and replace the intractably scaled collision-avoidance constraints with within-tunnel constraints, which are small-scale and independent of the environment’s complexity. Our tunnel construction method ensures the integrity of the free spaces, so that repeated construction is not necessary. Finally, in the process of trajectory optimization, we propose a lightweight OCP iterative solution framework to search for the optimal solution with high computational efficiency, in which the customized OCP with only box constraints is quickly solved in each iteration. Besides, the theoretical analysis, numerical simulations, and real-world experiments had been carried out to demonstrate the effectiveness and efficiency of the method.

I. Introduction

Autonomous driving problems have been widely researched over the last ten years due to their increasing significance in industry and traffic fields [1], [2], [3], [4]. As the core technology, trajectory planning is the key to sophisticated control system design and steady motion [5], [6]. A number of related works can be found in existing mobile robot movement designs [7] and autonomous vehicle dynamic missions [8]. However, in unstructured environments with dense obstacles, effective trajectory planning remains a challenge [9] due to the tiny and tortuous solution spaces caused by the interaction of complicated vehicle kinematics and constrained driveable regions [10]. To address this challenge, this paper focuses on a general solution, which is suitable for smart summon problems in scenarios with irregularly placed static obstacles.