引入动态加权地图的混合A*算法

Hybrid A* Algorithm Incorporating a Dynamic Weighted Map

摘要: 本文围绕多机器人路径规划中的前端搜索问题展开研究。传统混合A*算法本质上主要面向单机器人场景，其在初始路径搜索过程中更关注障碍物约束和机器人自身的运动学约束，而对其他机器人带来的动态干扰考虑不足。针对这一问题，本文在搜索阶段引入了动态加权地图，将其他机器人可能造成的局部拥堵和潜在冲突转化为附加搜索代价，从而引导机器人在路径生成过程中尽量避开风险较高的区域。基于该方法获得的初始路径能够更好地适应多机器人场景下的实际通行需求。

Abstract: This paper focuses on the front-end search problem in multi-robot path planning. Traditional Hybrid A* is essentially designed for single-robot scenarios. During initial path search, it mainly emphasizes obstacle constraints and the robot’s own kinematic constraints, while giving insufficient consideration to the dynamic interference caused by other robots. To address this issue, a dynamic weighted map is introduced into the search stage, where local congestion and potential conflicts caused by other robots are transformed into additional search costs. In this way, the robot is guided to avoid high-risk areas as much as possible during path generation. The initial path obtained by this method is therefore better suited to the actual traffic requirements of multi-robot scenarios.