宋燕利,张舒磊,宋康,等.基于ABSS-ARRT算法的焊接机械臂避障路径规划研究[J].精密成形工程,2024,16(11):168-177.
SONG Yanli,ZHANG Shulei,SONG Kang,et al.Obstacle Avoidance Path Planning of Welding Robot Arm Based on ABSS-ARRT Algorithm[J].Journal of Netshape Forming Engineering,2024,16(11):168-177. |
| 基于ABSS-ARRT算法的焊接机械臂避障路径规划研究 |
| Obstacle Avoidance Path Planning of Welding Robot Arm Based on ABSS-ARRT Algorithm |
| 投稿时间：2024-03-11 |
| DOI：10.3969/j.issn.1674-6457.2024.11.020 |
| 中文关键词: 路径规划 RRT*算法 自适应目标偏置 动态采样域 APF新节点生成 |
| 英文关键词: path planning RRT* algorithm adaptive target bias dynamic sampling domains APF new node generation |
| 基金项目:湖北省重点研发专项(2023BAB202);湖北省“楚天英才计划” 科技创新团队;广西科技重大专项(2023AA04007);教育部创新团队发展计划(IRT_17R83);新能源汽车科学与关键技术学科创新引智基地项目(B17034);中央高校基本科研业务费专项资金(WUT:2023III00JL) |
|
| 摘要点击次数: 137 |
| 全文下载次数: 16 |
| 中文摘要: |
| 目的 针对RRT*算法(Rapid-exploration Random Tree*)在机械臂避障路径规划中存在的搜索效率低、采样点质量差，算法整体效率低、收敛较慢等缺点，提出一种自适应偏置-步长-采样域策略及融合人工势场法的ABSS-ARRT*算法(Adaptive Bias-step-Sampling Domain and Improved Artificial Potential Field RRT*)。方法 在RRT*算法中融入自适应目标偏置采样策略和动态采样域策略，提出一种融合人工势场法(Artificial Potential Field，APF)思想的改进新节点生成策略，引入引力和斥力权重系数，同时采用自适应步长策略，使算法的性能得到提升。结果 通过二维和三维地图中的验证，相较于RRT算法和RRT*算法，ABSS-ARRT*算法在平均迭代次数、收敛时间、路径节点数量、路径总长度及平均成功率均表现出优越性，其中二维地图中路径长度分别缩短了21.8%和3.23%，平均迭代时间分别下降了35.6%和52.0%，三维地图中路径长度分别缩短了28.9%和19.5%，平均迭代时间分别下降了75.9%和72.5%，同时在MATLAB中对改进的RRT*算法在机械臂上进行可行性验证。结论 所提出的ABSS-ARRT*算法能够在复杂静态环境中为机械臂快速智能地规划出一条无碰撞高质量路径，验证了该算法的优越性和可行性。 |
| 英文摘要: |
| Aiming at the shortcomings of RRT algorithm (Rapid-exploration Random Tree) in obstacle avoidance path planning of robot arm, such as low search efficiency, poor quality of sampling points, low overall efficiency of the algorithm and slow convergence, the work aims to propose an Adaptive Bias-step-Sampling Domain and Improved Artificial Potential Field RRT. An adaptive target bias sampling strategy and a dynamic sampling domain strategy were introduced into RRT algorithm, and an improved new node generation strategy based on Artificial Potential Field (APF) was proposed. The gravity and repulsive force weight coefficients were introduced, and the adaptive step size strategy was added. The performance of the algorithm was improved. Compared with RRT algorithm and RRT algorithm, the 2D and 3D maps showed advantages in the average number of iterations, convergence time, number of path nodes, total path length and average success rate. In 2D maps, the average path length was shortened by 21.8% and 3.23%, and the average iteration time was reduced by 35.6% and 52.0%. In the three-dimensional map, the path length was shortened by 28.9% and 19.5% on average, and the average iteration time was reduced by 75.9% and 72.5%. Meanwhile, the feasibility of the improved RRT algorithm on the robot arm was verified in MATLAB. The proposed ABSS-ARRT algorithm can quickly plan a collision-free high-quality path for a robot arm in a complex environment, which verifies the superiority and feasibility of the algorithm. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
渝公网安备 50010702501719号