| 吴芳,赵云波.基于轨迹预测与改进人工势场法的机械臂动态避障规划方法[J].高技术通讯(中文),2023,33(9):937~946 |
| 基于轨迹预测与改进人工势场法的机械臂动态避障规划方法 |
| Dynamic obstacle avoidance for manipulator based on trajectory prediction and improved artificial potential field method |
| |
| DOI：10. 3772/ j. issn. 1002-0470. 2023. 09. 005 |
| 中文关键词: 轨迹预测； 改进人工势场法； 动态避障； 路径规划 |
| 英文关键词: trajectory prediction， improved artificial potential field method， dynamic obstacle avoidance， path planning |
| 基金项目: |
| 作者 | 单位 | | 吴芳 | （浙江工业大学信息工程学院杭州 310023） | | 赵云波 | |
|
| 摘要点击次数: 984 |
| 全文下载次数: 778 |
| 中文摘要: |
| 在人机共存环境中，人可能会成为机器人执行任务过程中的动态障碍物，因此，在机器人的运动过程中需要动态地避障规划，从而避免机器人危害到用户安全。人工势场法是常用的动态路径规划算法，具有实现简单、计算实时性高等优点。传统的人工势场法根据虚拟的引力场和斥力场得到合力，从而引导机器人的运动，但是当引力和斥力等大反向时，存在局部极小值问题。针对该问题，本文提出了基于平面位置采样的改进人工势场法。每次计算得到合力向量后，以该向量的指向为中心，在垂直于地面、包含合力向量的平面上以特定的角度间隔分别逆时针和顺时针方向采样90°范围内的运动方向，然后分别计算引力值和斥力值，最后根据引力值和斥力值的加权和最小确定机器人的最佳运动方向。为了应对用户运动导致机器人运动路径突变的情况，本文依据用户手臂运动和头部转动的关联关系，通过检测用户的头部姿态，并利用手臂当前的运动信息预测手臂接下来的运动位置。最后设计了基于轨迹预测与改进人工势场法的机械臂动态避障规划方法，实验结果表明，该方法可以有效地进行动态避障，并且规划的路径更加平滑、长度更短。 |
| 英文摘要: |
| In the human machine coexistence environment, humans may become dynamic obstacles in the task execution of the robot. Therefore, dynamic obstacle avoidance planning is required in the movement of the robot, so as to prevent the robot from endangering the safety of users. The traditional artificial potential field method computes the resultant force according to the virtual attractive force and repulsive force, then the robot moves in the direction of resultant force. However, when the attractive force and repulsive force have the same length, but in opposite directions, there is a local minimum problem. To solve this problem, an improved artificial potential field method based on position sampling is proposed in this paper. After calculating the resultant force, sample are taken anti clockwise and clockwise at angular intervals on a plane perpendicular to the ground and containing the resultant force, and then the attractive force and repulsive force of each sampling point are calculated. The optimal movement of the robot is determined according to the minimum sum of attractive force and repulsive force. Secondly, in order to deal with the peak point in the robot motion path caused by user motion, according to the correlation between user arm motion and head pose, the next pose of arm is predicted according to the user’s head pose and the current motion information of the arm. Finally, a dynamic obstacle avoidance planning algorithm of manipulator based on trajectory prediction and improved artificial potential field method is proposed. The experimental results show that this algorithm can effectively avoid dynamic obstacles, and the planned path is smooth and short. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
