矿用智能巡检机器人无标定视觉伺服控制研究

李静; 黄友锐; 韩涛; 兰世豪; 陈宏茂; 甘福宝

doi:10.13272/j.issn.1671-251x.2021030077

矿用智能巡检机器人无标定视觉伺服控制研究

doi: 10.13272/j.issn.1671-251x.2021030077

安徽理工大学电气与信息工程学院, 安徽淮南 232001

基金项目:

国家自然科学基金项目（61772033）。

详细信息

作者简介:
李静(1995-),女,安徽宿州人,硕士研究生,主要研究方向为机器人智能控制,E-mail:2293415900@qq.com。

中图分类号: TD67
计量
- 文章访问数: 96
- HTML全文浏览量: 23
- PDF下载量: 19
- 被引次数: 0
出版历程
- 收稿日期: 2021-03-27
- 修回日期: 2021-08-25
- 刊出日期: 2021-11-20

Research on uncalibrated visual servo control of mine intelligent inspection robot

School of Electrical and Information Engineering, Anhui University of Science and Technology, Huainan 232001, China

摘要

摘要: 针对矿用智能巡检机器人无标定视觉伺服控制中采用基于传统的卡尔曼滤波(KF)的图像雅可比矩阵存在估计值不准确、鲁棒性差的问题，提出了一种具有长短期记忆(LSTM)的卡尔曼滤波算法(KFLSTM算法)。KFLSTM算法使用LSTM弥补由KF算法产生的估计误差，将滤波增益误差、状态估计向量误差、观测误差用于LSTM的在线训练，利用训练后的LSTM模型对雅可比矩阵进行最优估计，通过提高雅可比矩阵估计值的准确性和稳定性来改善视觉伺服控制的实时性和鲁棒性。建立了基于KFLSTM算法的无标定视觉伺服模型，将最优雅可比矩阵作为控制器的输入，使控制器输出较准确的关节角速度，从而控制机械臂的实时运行。将KFLSTM算法应用到矿用智能巡检机器人六自由度视觉伺服仿真实验中，结果表明：应用KFLSTM算法得到的图像误差收敛速度相较于传统KF算法提高了100%～142%，图像特征误差更小，定位精度为0.5像素，且机器人末端执行器运动平稳，具有较强的抗噪声干扰能力，可有效提高矿用智能巡检机器人的作业精度与效率，并增强其工作的稳定性与安全性。
- 矿用智能巡检机器人 /
- 无标定视觉伺服 /
- 图像雅可比矩阵 /
- 卡尔曼滤波 /
- 长短期记忆神经网络
Abstract: In order to solve the problem of inaccurate estimation and poor robustness of image Jacobian matrix based on traditional Kalman filtering(KF)in the uncalibrated visual servo control of mine intelligent inspection robot, Kalman filtering algorithm with long and short-term memory(LSTM)(KFLSTM algorithm)is proposed.The KFLSTM algorithm uses the LSTM to compensate for the estimation error generated by the KF algorithm, uses the filter gain error, state estimation vector error and observation error for online training of the LSTM, and uses the trained LSTM model for optimal estimation of the Jacobian matrix to improve the real-time and robustness of visual servo control by improving the accuracy and stability of the Jacobian matrix estimation.The uncalibrated visual servo model based on the KFLSTM algorithm is established, and the most optimal Jacobian matrix is used as the input of the controller, which makes the controller output more accurate joint angular velocity so as to control the real-time operation of the manipulator.The KFLSTM algorithm is applied to the six-degree-of-freedom visual servo simulation experiment of the mine intelligent inspection robot.The results show that the image error convergence speed obtained by the KFLSTM algorithm is 100%-142% higher than that of the traditional KF algorithm, the image characteristic error is smaller, the positioning precision is 0.5 pixels, and the robot end effector moves smoothly.Moreover, the method has strong anti-noise interference capability, which can improve the precision and efficiency of the mine intelligent inspection robot effectively and enhance its stability and safety.
- mine intelligent inspection robot /
- uncalibrated visual servoing /
- image Jacobian matrix /
- Kalman filtering /
- long and short-term memory networks

HTML全文

参考文献(14)

[1]	潘祥生,陈晓晶.矿用智能巡检机器人关键技术研究[J].工矿自动化,2020,46(10):43-48. PAN Xiangsheng,CHEN Xiaojing.Research on key technologies of mine-used intelligent inspection robot[J].Industry and Mine Automation,2020,46(10):43-48.
[2]	康庆生,孟正大,戴先中.图像雅可比矩阵伪逆估计在视觉伺服中的应用[J].机器人,2006,28(4):406-409. KANG Qingsheng,MENG Zhengda,DAI Xianzhong.Application of pseudo-inverse estimation of image Jacobian matrixto visual servoing[J].Robot,2006，28(4):406-409.
[3]	赵杰,李牧,李戈,等.一种无标定视觉伺服控制技术的研究[J].控制与决策,2006,21(9):1015-1019. ZHAO Jie,LI Mu,LI Ge,et al.Study on uncalibrated visual servoing technique[J].Control and Decision,2006,21(9):1015-1019.
[4]	赵清杰,陈云蛟,张立群.基于粒子滤波的雅可比矩阵在线估计技术[J].北京理工大学学报,2008,28(5):401-404. ZHAO Qingjie,CHEN Yunjiao,ZHANG Liqun.On-line estimation of Jacobian matrix based on particle filter[J].Transactions of Beijing Institute of Technology,2008,28(5):401-404.
[5]	徐鹏,杨宏韬,李岩,等.基于扩展H∞粒子滤波算法的视觉伺服雅可比矩阵在线辨识[J].广西大学学报(自然科学版),2017,42(3):1051-1057. XU Peng,YANG Hongtao,LI Yan,et al.On-line identification of visual servo Jacobian matrix based on extended H∞ particle filter[J].Journal of Guangxi University(Natural Science Edition),2017,42(3):1051-1057.
[6]	WANG Fasheng,SUN Fuming,ZHANG Junxing,et al.Unscented particle filter for online total image Jacobian matrix estimation in robot visual servoing[J].IEEE Access,2019(99).DOI: 10.1109/ACCESS.2019.2927413.
[7]	梁喜凤,彭明,路杰,等.基于自适应无迹卡尔曼滤波的采摘机械手视觉伺服控制方法[J].农业工程学报,2019,35(19):230-237. LIANG Xifeng,PENG Ming,LU Jie,et al.Servo control method of picking manipulator based on adaptive traceless Kalman filter[J].Transactions of the Chinese Society of Agricultural Engineering,2019,35(19):230-237.
[8]	王新梅,魏武,刘玮,等.鲁棒卡尔曼滤波下的图像雅可比矩阵带时延补偿的估计[J].控制理论与应用,2015,32(8):1052-1057. WANG Xinmei,WEI Wu,LIU Wei,et al.Estimation of image Jacobian matrix with time-delay compensation basedon robust Kalman filtering[J].Control Theory & Applications,2015,32(8):1052-1057.
[9]	王洪斌,李萍.基于改进的卡尔曼滤波图像雅可比矩阵估计[J].武汉理工大学学报,2009,31(14):145-148. WANG Hongbin,LI Ping.On-line estimation for image Jacobian matrix based on improved Kalman filter[J].Journal of Wuhan University of Technology,2009,31(14):145-148.
[10]	章晓峰,李光,肖帆,等.基于BP神经网络的包装分拣机器人视觉标定算法[J].包装学报,2019,11(4):74-81. ZHANG Xiaofeng,LI Guang,XIAO Fan,et al.A visual calibration algorithm for packaging sorting robot based on BP neural network[J].Journal of Packaging Science and Technology,2019,11(4):74-81.
[11]	TSAI Z R,CHANG Y Z.Enhancing design of visual-servo delayed system[J].Journal of Electronic Science and Technology of China,2018,16(3):232-240.
[12]	CHI Gaoxuan.Application study of fuzzy ARTMAP neural network in robot servo system[J].Journal of Physics Conference Series,2018,1026(1):012024.DOI: 10.1088/1742-6596/1026/1/012024.
[13]	GU Jinan,WANG Hongmei,PAN Yuelong,et al.Neural network based visual servo control for CNC load/unload manipulator[J].Optik International Journal for Light and Electron Optics,2015,126(23):4489-4492.
[14]	NADI F,DERHAMI V,REZAEIAN M.Visual servoing control of robot manipulator with Jacobian matrix estimation[C]//Proceedings of the 2014 International Conference on Robotics and Mechatronics,2014：405-409.DOI: 10.1109/ICRoM.2014.6990935.