摘要
快速探索随机树(RRT)算法虽能高效解决复杂空间的路径规划问题,但存在路径质量差、无法保证最优解及对动态障碍物处理不足等问题。此外,算法效率受随机性影响,探索过程冗余,导致搜索成本偏高。文中提出了一种基于广义正态分布优化(GNDO)算法对RRT算法进行优化。将RRT算法已经探索到的路径节点作为输入,用GNDO算法通过计算每个点的适应度值进行取舍和调节,找到最优路径;同时,为了加快RRT算法的收敛速度,引入自适应步长策略调整父节点与目标节点之间的距离。仿真实验结果证实:与RRT和RRT^(*)相比,经过GNDO优化后的GNDO-RRT算法路径质量显著提升,路线更为平滑,时间消耗也远低于RRT^(*)优化算法。
Path planning,a key research field in robotics,computer science,and artificial intelligence,aims to identify the optimal path for mobile entities,such as robots,drones,and autonomous vehciels.With technological advances,path planning is gaining growing importance in various applications,particularly in complex and dynamic environments.Path planning encompasses not only algorithm design and implementation,but also closely relates to environmental perception,decision-making processes,and control strategies.Path planning methods primarily divide into two categories:global path planning and local path planning.The former focuses on identifying the optimal path in known environments,typically relying on comprehensive environmental information and detailed map data.Common global path planning algorithms include the A algorithm,Dijkstra’s algorithm,and the Rapidly-exploring Random Tree(RRT).These algorithms effectively compute the shortest path from the starting point to the target point,making them suitable for path planning in static environment.In contrast,the latter emphasizes real-time path adjustments in dynamic environments to avoid obstacles and other moving elements.Local path planning algorithms,such as the Dynamic Window Approach and Aritificial Potential Fields,quickly respond to environmental changes,ensuring the safety of mobile entities,especially in complex dynamic environments and unknown scenarios.The Rapidly-exploring Random Tree(RRT)algorithm,discussed in this paper,serves as a widely applied sampling-based intelligent algorithm in global path planning.The RRT algorithm effeciently addresses path planning problems in complex spaces by randomly sampling the space and gradually expanding the tree structure.However,the RRT algorithm has severe drawbacks in applications,including poor path quality,in anability to an inability to guarantee optimal solutions,and insufficient handling of dynamic obstacles.Additionally,the algorithm’s efficiency suffers from inherent randomness,leading to potentially redundant exploration and high search costs,particularly in high-dimensional spaces.To address these challenges,this research proposes an optimization method for the RRT algorithm based on the Generalized Normal Distribution Optimization(GNDO)algorithm.This method utilizes the path nodes already explored by the RRT algorithm as input,calculating the fitness values of each point through the GNDO algorithm for selection and adjustment to identify the optimal path.The GNDO algorithm effectively balances exploration and exploitation during the path optimization process,thereby enhancing path quality.Furthermore,to accelerate the convergence speed of the RRT algorithm,we introduce an adaptive step size strategy to adjust the distance between parent nodes and target nodes,facilitating faster path searches.In the experimental section,we systematically evaluate the GNDO-RRT algorithm within a simulation environment.Our experimental results indicate the GNDO-RRT algorithm,significantly improves path quality compared to traditional RRT algorithms,generating smoother paths while consuming less time.Our research may provide new insights for the path planning and lay a solid foundation for future applications in dynamic environments.
作者
陈胜锦
杨光永
崔光海
徐天奇
CHEN Shengjin;YANG Guangyong;CUI Guanghai;XU Tianqi(School of Electrical and Information Engineering,Yunnan Minzu University,Kunming 650500,China)
出处
《重庆理工大学学报(自然科学)》
北大核心
2025年第3期127-132,共6页
Journal of Chongqing University of Technology:Natural Science
基金
国家自然科学基金项目(61761049,61261022)
云南省教育厅科学研究基金项目(2023Y0502)
云南民族大学2022年硕士研究生科研创新基金项目(2022SKY006)。
作者简介
陈胜锦,男,硕士研究生,主要从事机器人路径规划研究,E-mail:chenshengjin2021@163.com。
