基于YOLOv4-Tiny和RRT-Connect算法的机械臂自主抓取仿真
Simulation of Autonomous Grab of Manipulator Based on YOLOv4-Tiny and RRT-Connect Algorithm
DOI: 10.12677/MOS.2023.123254, PDF,    国家自然科学基金支持
作者: 余吉雅, 张艳超, 张文博：浙江理工大学信息科学与工程学院，浙江 杭州
关键词: 目标检测；路径规划；ROS；机械臂仿真；Object Detection； Path Planning； ROS； Manipulator Simulation
摘要: 机械臂自主抓取一直是机器人领域的研究重点，路径规划和目标检测问题是机械臂自主抓取的核心内容，由于深度神经网络的发展，目标检测精度得到了大幅度的提升，本文根据YOLOv4-tiny算法进行草莓的目标检测，获得了93.5%的检测精度，并为机械臂选择合适的路径规划算法，通过在二维障碍地图和三维障碍地图进行模拟实验，得到RRT-connect算法的路径规划效率和成功率均大于RRT算法，并在三维障碍地图中获得了100%的成功率。在仿真环境中，通过Gazebo软件搭建机械臂仿真环境，并通过Moveit软件对机械臂进行控制和路径规划，将YOLOv4-tiny算法和RRT-connect算法应用在机械臂抓取仿真系统中，最终平均获得了87%的成功抓取率。
Abstract: Autonomous grasping of manipulator has always been the focus of research in the field of robotics. Path planning and object detection are the core contents of autonomous grasping of manipulator. Due to the development of deep neural network, the accuracy of object detection has been greatly improved. In this paper, the object detection of strawberry is carried out according to YOLOv4-tiny algorithm, and the detection accuracy is 93.5%. The appropriate path planning algorithm is select-ed for the manipulator. Through simulation experiments on two-dimensional obstacle map and three-dimensional obstacle map, it is concluded that the path planning efficiency and success rate of RRT-connect algorithm are greater than those of RRT algorithm. The simulation environment of the manipulator is built by Gazebo software, and the control and path planning of the manipulator are carried out by Moveit software. The YOLOv4-tiny algorithm and RRT-connect algorithm are applied to the manipulator grasping simulation system. Finally, the average successful grasping rate of 87% is obtained.
文章引用：余吉雅, 张艳超, 张文博. 基于YOLOv4-Tiny和RRT-Connect算法的机械臂自主抓取仿真[J]. 建模与仿真, 2023, 12(3): 2773-2781. https://doi.org/10.12677/MOS.2023.123254

参考文献

[1] Li, Y., Guo, Z., Shuang, F., et al. (2022) Key Technologies of Machine Vision for Weeding Robots: A Review and Benchmark. Computers and Electronics in Agriculture, 196, 106880. [Google Scholar] [CrossRef] 
[2] 刘亚, 艾海舟, 徐光佑. 一种基于背景模型的运动目标检测与跟踪算法[J]. 信息与控制, 2002, 31(4): 315-319, 328.
[3] 张捍东, 郑睿, 岑豫皖. 移动机器人路径规划技术的现状与展望[J]. 系统仿真学报, 2005, 17(2): 439-443.
[4] Wjpd, A., Yt, A., Rong, L.B., et al. (2020) Detection and Segmentation of Overlapped Fruits Based on Optimized Mask R-CNN Application in Apple Harvesting Robot. Computers and Electronics in Agriculture, 172, 105380.
[5] Onishi, Y., Yoshida, T., Kurita, H., et al. (2019) An Automated Fruit Harvesting Robot by Using Deep Learning. Robomech Journal, 6, Article No. 13. [Google Scholar] [CrossRef] 
[6] Lehnert, C., Sa, I., Mccool, C, et al. (2016) Sweet Pepper Pose Detection and Grasping for Automated Crop Harvesting. Proceedings of the 2016 IEEE International Conference on Robotics and Auto-mation (ICRA), Stockholm, 16-21 May 2016, 2428-2434. [Google Scholar] [CrossRef] 
[7] Bochkovskiy, A., Wang, C.-Y. and Liao, H.-Y.M. (2020) Yolov4: Optimal Speed and Accuracy of Object Detection. arXiv: 200410934.
[8] Karaman, S., Walter, M.R., Perez, A., et al. (2011) Anytime Motion Planning Using the RRT. Shanghai, 9-13 May 2011, 1478-1483. [Google Scholar] [CrossRef] 
[9] Jr, J. and Lavalle, S.M. (2000) RRT-Connect: An Efficient Approach to Single-Query Path Planning. Proceedings of the 2000 IEEE International Conference on Robotics and Automation, ICRA 2000, San Francisco, 24-28 April 2000, 995-1001.
[10] Quigley, M., Gerkey, B.P. and Conley, K., et al. (2009) ROS: An Open-Source Robot Operating System. ICRA Workshop on Open Source Software, 3, 5.
[11] Koenig, N. and Howard, A. (2004) Design and Use Paradigms for Gazebo, an Open-Source Multi-Robot Simulator. 2004 IEEE/RSJ Interna-tional Conference on Intelligent Robots and Systems (IROS), Sendai, 28 September-02 October 2004, 2149-2154.
[12] 孙亮, 马江, 阮晓钢. 六自由度机械臂轨迹规划与仿真研究[J]. 控制工程, 2010, 17(3): 388-392.

为你推荐