基于多属性决策的多车自主泊车路径全局规划方法

Multi-Attribute Decision-Based Global Planning for Multi-Vehicle Autonomous Parking

摘要:

虽然自主代客泊车技术已经可以代替驾驶员完成泊车操作，但在多车辆且缺乏调度的情况下，容易造成交通混乱。为此，提出了一种基于多属性决策的多车自主泊车路径全局规划方法，旨在解决泊位分配不合理问题并降低泊车整体成本。首先，建立停车场泊位拓扑地图，基于直线和回旋线构建引导路径；然后，设计引导路径择优的准则和子准则，构建层次分析（AHP）择优体系，确定最优泊位及其引导路径；最后，确定泊车起终点位置，以回旋线为基准，根据泊位的碰撞约束条件计算泊车路径。此外，模拟了多车同时进入停车场的场景，仿真结果表明，该方法可以有效降低泊车成本，提高泊车效率。

Abstract:

Although the autonomous valet parking technology is able to complete parking operations instead of drivers， the lack of scheduling of multiple vehicles can easily cause traffic chaos. This paper proposes a multi-attribute decision-based global planning method for multi-vehicle autonomous parking paths， which aims to solve the problem of unreasonable parking space allocation and reduce the overall cost of parking. First， a topological map of parking spaces is established to construct guided paths based on straight lines and clothoids. Then， the criteria and sub-criteria of the guided path selection are designed， and an analysis hierarchical process （AHP） path selection system is constructed to determine the optimal parking space and its guided path. Finally， the parking starting and ending positions are determined， and the parking path is calculated based on the collision constraints of the parking space using the clothoids as the reference. The scenario of multiple vehicles entering the parking lot at the same time is simulated， and the simulation results show that the method can effectively reduce the parking cost and improve the parking efficiency.