改进PRM在特定复杂环境下的路径规划  

Improved PRM Path Planning in Specific Complex Environments

作　　者：杨振舰 张国良 YANG Zhenjian;ZHANG Guoliang(School of Computer and Information Engineering,TCU,Tianjin 300384,China)

机构地区：[1]天津城建大学计算机与信息工程学院,天津300384

出　　处：《天津城建大学学报》2024年第5期346-353,共8页Journal of Tianjin Chengjian University

基　　金：天津市自然科学基金(19JCTPJC50000)。

摘　　要：针对传统概率路线图法(probabilistic roadmap methods,PRM)在特定复杂场景(场景中存在十字形的窄通道连通区域,且最终规划出来的路径必定会经过十字形的窄通道连通区域)下路径规划成功率低的问题,提出了一种改进的PRM方法.在学习阶段,首先使用Fast(features from accelerated segment test)角点检测算法来检测给定地图环境的角点;然后根据检测出来的角点计算地图的主要连接区域(十字形的窄通道连通区域中心的矩形),通过对这些主要连通区域额外随机采样一个点,从而增加了主要连通区域的采样点数量,提高了窄通道和自由区域的连通性.仿真结果表明,优化PRM路径规划的成功率比传统PRM的成功率高出70%,路径规划时间缩短了1.2 s,因此改进的方法大大提高了PRM算法的效率.The authors propose an improved PRM method to solve some problems about the traditional probabilistic roadmap methods(PRM)which has a low success rate in specific complex environments(a cross-shaped narrow channel connection area exists in this complex scene,and the final planned path must pass through the cross-shaped narrow channel connection area).In the learning phase,the authors first use the Fast(Features from Accelerated Segment Test)corner point detection algorithm to detect the corner points of a given map environment,and then calculate the main connection area(the rectangle in the center of the cross-shaped narrow channel connection area)of the map based on the detected corner points.By randomly sampling an additional point in these main connected areas,the number of sampling points in the main connected area can be increased and the connectivity of narrow channel can be improved.The free area simulation results show that the success rate of optimized PRM path planning is 70%higher than the success rate of traditional PRM,and the path planning time is shortened by 1.2 s,so the improved method greatly improves the efficiency of PRM algorithm.

关 键 词：路径规划 PRM算法 Fast算法 特定复杂环境 主要连通区域 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]