陈警, 罗斌, 张婧, 李佗, 王晨捷. 一种移动机器人激光模型全局路径规划方法[J]. 武汉大学学报 ( 信息科学版). DOI: 10.13203/j.whugis20220067
引用本文: 陈警, 罗斌, 张婧, 李佗, 王晨捷. 一种移动机器人激光模型全局路径规划方法[J]. 武汉大学学报 ( 信息科学版). DOI: 10.13203/j.whugis20220067
CHEN Jing, LUO Bin, ZHANG Jing, LI Tuo, WANG Chenjie. A Global Path Planning Method for Mobile Robot Laser Model[J]. Geomatics and Information Science of Wuhan University. DOI: 10.13203/j.whugis20220067
Citation: CHEN Jing, LUO Bin, ZHANG Jing, LI Tuo, WANG Chenjie. A Global Path Planning Method for Mobile Robot Laser Model[J]. Geomatics and Information Science of Wuhan University. DOI: 10.13203/j.whugis20220067

一种移动机器人激光模型全局路径规划方法

A Global Path Planning Method for Mobile Robot Laser Model

  • 摘要: 为了使移动机器人在全局路径规划中更好地适应地图环境,本文提出了一种改进A*和射线模型的激光模型全局路径规划方法,本方法借鉴了平面激光雷达实时扫描原理,由路径节点向目标点散发多条虚拟激光射线来感知障碍物边界,从而快速越过障碍物并到达目标点,同时结合Floyd优化算法,在较短时间内得出一条安全可靠稳定的全局路径。在本实验凹型环境中,激光模型搜索时间快于A*和射线模型99%,搜索过程减少96%;在本实验不可行区域中,激光模型上快于前两种算法98%,搜索过程减少97%;同时双蚁群交叉凹形环境算法对比,表现也相对较优。实验表明本方法对解决凹型陷阱及不可行区域耗时问题比较有效。

     

    Abstract: Objective: In order to make the global path planning of mobile robot better adapt to the complex map environment, a laser model global path planning method based on improved A* and ray model is proposed in this paper. Methods: Send a ray from the node to the end point, if an obstacle is encountered, send N diffuse rays to the left and right sides in turn. Judge whether the two adjacent rays cross the obstacle, so as to judge the boundary of the obstacle. By comparing the cost of the boundary nodes on both sides, select the optimal boundary node for the next node search, thus, the effect of quickly crossing the current obstacle and moving towards the target point can be realized. If the obstacle is encountered again on the way, the algorithm will be followed again. Combined with Floyd optimization algorithm, a safe, reliable and stable global path is obtained. Results: The comparative experiments of laser model algorithm, A* algorithm and original ray model were carried out in this paper. The experimental conditions were simulated by matlab, and the map was simulated. The experimental results were compared from the index dimensions of calculation time, total distance, number of nodes, number of corner nodes, total distance of search process, number of nodes in search process and so on. This paper carries out comparative experiments in different environments. It is concluded that in concave environment, the search time of laser model in this paper can be 99% faster than that of A* and ray model, and the search process can be reduced by 96%. In the infeasible region experiment, the laser model in this paper is 98% faster than the first two algorithms, and the search process is reduced by 97%. At the same time, the performance of the two ant colony cross concave environment algorithm is also relatively better. Conclusion: The ray model and A* algorithm are compared in concave trap environment, infeasible area environment and complex environment. It is concluded that the laser model in this paper has better global path planning results, fewer nodes, faster calculation process and fewer search path nodes in such environment. As a new method, the laser model algorithm in this paper can provide a new idea and direction for global path planning and navigation in the environment of unknown map in the real robot scene.

     

/

返回文章
返回