摘要
基于卷积神经网络设计了一款柔性关节机械臂的控制率方案。针对机械臂执行器动力学、关节柔性的特点,采用滑模控制法设计了一套控制率算法,克服了干扰与设备中的不确定性问题。为了满足滑模控制对输入控制量的要求,应用卷积神经网络原理设计了基于视觉的机械臂控制优化方法。由仿真与实际系统实验测试结果可知,该系统的联合预测精度与跟踪控制精度较高、误检概率低,具有良好的工程应用前景。
Based on convolutional neural network,a control rate scheme of flexible joint manipulator is designed.According to the characteristics of the manipulator's actuator dynamics and joint flexibility,a set of control rate algorithm is designed by using the sliding mode control method,which overcomes the problems of disturbance and uncertainty in the equipment.In order to meet the requirement of sliding mode control for input control quantity,a vision-based control optimization method of manipulator is designed by applying the principle of convolutional neural network.It can be seen from the simulation and actual system experimental test results that the system has high joint prediction accuracy and tracking control accuracy,low false detection probability,and has a good engineering application prospect.
作者
周荣亚
刘刚
Zhou Rongya;Liu Gang(School of Railway Equipment Manufacture,Shaanxi Railway Institute,Shaanxi Weinan,714000,China;School of Materials Science and Engineering,Xi'an University of Technology,Shaanxi Xi'an,710048,China)
出处
《机械设计与制造工程》
2022年第10期47-51,共5页
Machine Design and Manufacturing Engineering
基金
陕西省自然科学基础研究计划面上项目(2020JM-455)。
关键词
卷积神经网络
控制率
机械臂
滑模控制
convolutional neural network
control rate
manipulator
sliding mode control
