文章摘要
魏立新* **,侯仕杰* **,孙浩* **,吴绍坤***.复杂环境下基于目标指引的RRT*路径规划算法[J].高技术通讯(中文),2021,31(6):589~597
复杂环境下基于目标指引的RRT*路径规划算法
RRT* path planning algorithm based on target guidance in complex environment
  
DOI:10.3772/j.issn.1002-0470.2021.06.003
中文关键词: 复杂环境; 目标指引RRT*; 模糊推理策略; 动态环境; 树莓派智能小车
英文关键词: complex environment, target guidance RRT*, fuzzy inference strategy, dynamic environment, Raspberry Pi smart car
基金项目:
作者单位
魏立新* **  
侯仕杰* **  
孙浩* **  
吴绍坤***  
摘要点击次数: 983
全文下载次数: 627
中文摘要:
      机器人路径规划算法需应对运动过程中遇到的各种复杂环境。针对快速扩展随机树 (RRT)算法规划时间长、产生新节点随机性大、盲目性强的缺点,提出基于目标指引的RRT*路径规划算法。该算法在障碍物和目标点处分别产生虚拟势场,引入引力函数和斥力函数使得生成的随机点具有目标性,随机点朝向目标点方向产生,降低盲目性和随机性;回归策略和动态自适应步长策略减少规划时间和产生冗余点的数量。当环境复杂时,提出带有预测机制的模糊推理策略,以解决机器人在U型陷阱下易产生的局部死锁现象。在动态环境下,提出重规划策略使机器人拥有动态避障能力。最后,在树莓派智能小车上进行了实验测试,结果验证了该算法的有效性。
英文摘要:
      Robot path planning algorithm needs to deal with various complex environments encountered in the process of movement. According to the disadvantages of the algorithm of rapidly-exploring random trees(RRT), such as long planning time, large randomness and strong blindness of new nodes, an algorithm of RRT* path planning based on target guidance is proposed. In this algorithm, virtual potential fields are generated at obstacles and target points respectively. By introducing the gravitational function and the repulsive function, the generated random points are targetable. The random points are generated towards the target point, which reduces the blindness and randomness. Regression and dynamic adaptive step-size strategies reduce planning time and the number of redundant points. When the environment is complex, a fuzzy inference strategy with predictive mechanism is proposed to solve the local deadlock in U-shaped traps. In dynamic environment, the reprogramming strategy is proposed to enable the robot to have dynamic obstacle avoidance ability. Finally, the experiment is carried out on the Raspberry Pi smart car. The results show that the algorithm is effective.
查看全文   查看/发表评论  下载PDF阅读器
关闭

分享按钮