基于改进D*Lite⁃APF算法的巡检机器人路径规划  

Inspection robot path planning based on improved D*Lite⁃APF algorithm

作　　者：胡粒琪 曾维[1] 陈才华 张鹏[2] 王艺儒 李铜 HU Liqi;ZENG Wei;CHEN Caihua;ZHANG Peng;WANG Yiru;LI Tong(Department of Artificial Intelligence,College of Computer Science and Cyber Security,Chengdu University of Technology,Chengdu 610059,China;Department of Communication Engineering,School of Mechanical and Electrical Engineering,Chengdu University of Technology,Chengdu 610059,China)

机构地区：[1]成都理工大学计算机与网络安全学院人工智能系,四川成都610059 [2]成都理工大学机电工程学院通信工程系,四川成都610059

出　　处：《现代电子技术》2024年第5期155-159,共5页Modern Electronics Technique

基　　金：国家科技重大专项专题(GFZX02040205.3);四川省重点研发项目(2023YFN0053)。

摘　　要：针对巡检机器人在动态场景下路径规划存在非全局最优、路径不平滑及局部避障效果不佳的问题,提出一种将改进D*Lite算法和人工势场法融合的算法。首先优化D*Lite算法启发代价函数,提升规划效率,并引入Dubins曲线平滑生成的全局路径;其次改进人工势场法势场函数并添加随机半径扰动点,解决局部碰撞问题,提高避障性能;最后将两种优化算法有效融合,实现全局规划和局部避障。仿真实验结果表明,相较于单一D*Lite算法,融合算法在路径长度、时间花销、路径拐点及扩展节点数方面均表现更优,能在确保全局路径最优的情况下有效避障。In view of the non⁃global optimality,unsmooth path and poor local obstacle avoidance effect in the path planning of inspection robots in dynamic scenes,an algorithm that integrates the improved D*Lite algorithm and the artificial potential field method is proposed.The D*Lite algorithm is optimized to heed the cost function,improve the planning efficiency,and introduce the global path generated by Dubins curve smoothing.The artificial potential field normal potential field function is improved and random radius disturbance points are added to avoid the local collision and improve the obstacle avoidance performance.The two optimization algorithms are integrated effectively to realize global planning and local obstacle avoidance.Simulation results show that,in comparison with the single D*Lite algorithm,the fusion algorithm performs better in path length,time cost,path inflection point and number of extension nodes,and can avoid obstacles effectively under the condition of ensuring the optimal global path.

关 键 词：巡检机器人 路径规划 D*Lite Dubins曲线 人工势场法 避障 

分 类 号：TN99-34[电子电信—信号与信息处理]