融合人工势场法的动态快速行进树路径规划算法  

Dynamic fast marching tree algorithm with integrated artificial potential fields

作　　者：吴旭鹏 贾小林[1,2] 顾娅军 Wu Xupeng;Jia Xiaolin;Gu Yajun(RFID&IOT Laboratory,School of Computer Science&Technology,Southwest University of Science&Technology,Mianyang Sichuan 621010,China;Mobile Internet of Things&Radio Frequency Identification Technology Key Laboratory of Mianyang,Mianyang Sichuan 621010,China)

机构地区：[1]西南科技大学计算机科学与技术学院RFID&IOT实验室,四川绵阳621010 [2]绵阳市移动物联射频识别技术重点实验室,四川绵阳621010

出　　处：《计算机应用研究》2024年第9期2745-2750,共6页Application Research of Computers

基　　金：国家自然科学基金面上项目(61471306);四川省自然科学基金资助项目(2022NSFSC0548);四川省重点研发计划资助项目(2020YFS0360);四川省教育厅教改资助项目(JG2021-1414)。

摘　　要：针对快速行进树算法(FMT*)由于随机采样导致的冗余探索问题以及不适用于动态环境的问题,提出一种融合人工势场法的动态快速行进树路径规划算法(APF-Dynamic FMT*),该算法设计了一种基于人工势场法的采样点引导函数,该函数根据环境信息动态的调整采样点生成范围,减少冗余探索。同时,该算法设计了一种路径树动态调整机制,当现有路径受到环境改变的影响时,能在未受影响的剩余路径树的基础上重新规划出新的优秀路径,适用于解决动态环境下的路径规划问题。仿真实验结果表明,APF-Dynamic FMT*算法在消耗相同计算资源的同时,显著提高了路径规划的成功率与路径质量,且当现有路径受动态环境影响后,能够高效地重新规划出可通行的优秀路径。Addressing the issues of redundant exploration and inapplicability in dynamic environments inherent in fast mar-ching tree algorithm(FMT*),this paper proposed the APF-Dynamic FMT*algorithm,which integrated artificial potential field method.This algorithm designed a sampling point guidance function based on artificial potential field method,which could dynamically adjust the sampling point generation range according to the environment information to reduce redundant exploration.Additionally,this algorithm designed a dynamic path tree adjustment mechanism,when the existing path was affected by the environment change,it could re-plan a new excellent path on the basis of the remaining path tree that is not affected,which was suitable for solving the path planning problem in dynamic environment.The results verify that the APF-Dynamic FMT*algorithm can significantly improve the success rate and path quality of path planning while consuming the same computational resources,and when the existing path is affected by the dynamic environment,it can efficiently re-plan the passable excellent path.

关 键 词：FMT*算法 人工势场法 动态路径规划 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]