基于递归神经网络的机器人手臂轨迹跟踪控制  

Trajectory Tracking Control of Robot Arm Based on Recurrent Neural Network

作　　者：张迪 ZHANG Di(School of Computer and Software Engineering,Sias University,Zhengzhou 451150,China;Henan Intelligent Manufacturing and Digital Twin Engineering Research Center,Zhengzhou 451150,China)

机构地区：[1]郑州西亚斯学院计算机与软件工程学院,河南郑州451150 [2]河南省智能制造数字孪生工程研究中心,河南郑州451150

出　　处：《机械工程与自动化》2024年第4期19-21,共3页Mechanical Engineering & Automation

基　　金：河南省科技攻关项目(242102210088)。

摘　　要：针对机器人手臂在连续变化的过程中不能及时跟随运动矩阵变化而导致的控制精度不高的问题,提出了一种基于递归神经网络的机器人手臂轨迹跟踪控制方法。首先建立了机器人手臂的动力学数学模型,并设计了机器人手臂的轨迹跟踪控制器。然后基于递归神经网络建立了轨迹跟踪线性变化参数模型,对机器人手臂运动轨迹进行实时调整,并通过求解静态反馈控制数值,实现了控制器参数的在线调整来适应外界环境的变化,从而有效改善了机器人手臂运动控制的精度。实验结果表明:所提方法具备更好的鲁棒性和较高的控制精准度,平均轨迹跟踪误差仅为0.008 m。研究结果可为机器人手臂的高精度控制提供理论支持。Aiming at the problem that the robot arm can not follow the change of the motion matrix in time in the process of continuous change,which leads to low control accuracy,a trajectory tracking control method of robot robot arm based on recurrent neural network is proposed.Firstly,a mathematical model of robot arm dynamics was established,and a trajectory tracking controller for the robot arm was designed.Then,a linear parameter model for trajectory tracking was established based on recurrent neural networks,which can adjust the motion trajectory of the robot arm in real-time.By solving static feedback control values,online adjustment of controller parameters was achieved to adapt to changes in the external environment,effectively improving the accuracy of robot arm motion control.The experimental results show that the proposed method has better robustness and higher control accuracy,and the average trajectory tracking error is only 0.008 m.The research results can provide theoretical support for high-precision control of the robot arm.

关 键 词：递归神经网络 机器人 参数模型 多关节手臂 鲁棒控制 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TP183[自动化与计算机技术—控制科学与工程]