基于地图建立的无人地面车路径规划

 提出一种同时完成地图建立与路径规划的算法。该算法为两层控制结构，其上层实现子目标点的生成，下层完成局部路径规划及运动控制。根据系统实时性的要求，以N个系统周期为触发条件执行子目标点生成程序。其中无人地面车通过传感器不断获取环境信息并进行处理，完成网格占据方式的地图建立与实时更新；将地图建立的结果作为数据输入，利用A*路径规划算法生成子目标点。根据子目标点生成结果，在每个系统周期内，通过基于模糊控制的底层快速算法完成无人地面车到子目标点的运动控制。以Pioneer 3-AT型无人地面车为试验平台在未知的复杂环境中对该算法进行验证，取得了良好的地图建立和路径规划效果，证明了该算法具有良好的实时性和准确性。

Path Planning for Unmanned Ground Vehicle Based on Map Building

This paper puts forward a two-layered path planning algorithm to be used for map building and path planning at the same time. The function of the upper layer is to create sub-goals, and the lower layer is to plan local routes. Based on the system real-time demand, the task of creating sub-goals is carried out at intervals of N system cycles. But at every system cycle, the algorithms of planning the local routes and motion controls are all carried on for the safety of system. The map building algorithm makes use of the information acquired by the Unmanned Ground Vehicle (UGV) and the environment information is processed by sonar, which is combined into the local map, to generate and renew grid-occupation based map. Based on the map, the A-star algorithm creates sub-goals and plans the route which makes the assessment function reach a minimum. At every system cycle, the "following-wall" strategy based on fuzzy control is implemented to avoid obstacles due to sub-goals. In this paper, the strategy is divided into the left following-wall and the right following-wall to decide which boundary of the obstacle to follow using the information from sonar synchronously. The algorithm is verified on Pioneer 3-AT UGV in an unknown and complicated environment. It is shown that the system can achieve practical map building and path planning with a good accuracy. The system supported in this paper is easy to extend and re-structure and to be transplanted to other systems.