Collision avoidance path planning for mobile robots using an improved Bi-RRT* algorithm
Jinxu Sun, Lei Zhang, jingjing yang, Xiaolong Liu, Yutao Feng
Tianjin University of Technology and Education Yantai Automobile Engineering Professional College Changchun Technical University of Automobile
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
To address the high path redundancy, poor smoothness, and low planning efficiency exhibited by the bi-directional rapidly-exploring random tree (Bi-RRT) algorithm in complex dynamic environments, an improved Bi-RRT* algorithm for collision avoidance path planning in mobile robots is proposed. The method employs a Social LSTM network to predict the trajectories of dynamic obstacles. The predicted positions over a 3 s horizon are utilized to guide real-time obstacle avoidance decisions. A bidirectional target-biased sampling strategy and an adaptive step-size mechanism are adopted to enhance search efficiency and adaptability in dynamic environments. Additionally, a via-node optimization is introduced to eliminate redundant nodes in the initial path. Finally, the generated path is smoothed using a cubic cardinal spline. Experimental results demonstrate that in mixed obstacle scenarios, the average planning time is reduced by 52.1%, the path length is shortened by 12.5%, the success rate reaches 96.7%, and the average curvature is decreased by 62.2% compared with the BD‐RRT algorithm. These findings validate the effectiveness of the proposed algorithm in terms of real‐time performance, smoothness, and kinematic feasibility.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
工程Robotic Path Planning Algorithms
Reinforcement Learning in Robotics · Robotics and Sensor-Based Localization
参考文献 23
此处列出前 3 条