机械臂神经网络控制优化与仿真  

Neural network control optimization and simulation of robot arm

作　　者：许洋洋 王莹 薛东彬[2] XU Yangyang;WANG Ying;XUE Dongbin(School of Mechanical and Electrical Engineering,Zhengzhou University of Industrial Technology,Zhengzhou 451150,China;School of Mechanical and Electrical Engineering,Henan University of Technology,Zhengzhou 450007,China)

机构地区：[1]郑州工业应用技术学院机电工程学院,郑州451150 [2]河南工业大学机电工程学院,郑州450007

出　　处：《中国工程机械学报》2018年第5期416-420,共5页Chinese Journal of Construction Machinery

基　　金：河南省自然科学基金资助项目(2013B460024)

摘　　要：为了提高机械臂末端连杆运动轨迹控制的稳定性,在径向基函数(RBF)神经网络控制器的基础上,采用混合算法优化RBF神经网络控制器.用两个径向基神经网络单元作为自适应控制器,其中一个作为输入端的控制器,另一个作为机械臂的辨识器.将混合算法优化应用到这两个神经网络单元中,以改善网络结构参数对神经网络控制和辨识性能的影响,在Matlab环境下进行了仿真实验,并与RBF神经网络控制器跟踪效果进行对比.仿真结果显示:在受到不确定因素干扰时,机械臂末端连杆采用改进RBF神经网络控制器产生的误差较小,系统反应速度较快,转矩波动较小.机械臂末端连杆采用改进RBF神经网络控制器,具有抗干扰的能力,快速保持系统输出的稳定性.In order to improve the motion trajectory control stability of end link mechanical arm,on the basis of radial basis function(RBF)neural network controller,hybrid algorithm is used to optimize RBF neural network controller.Two radial basis neural network elements are used as adaptive controllers,one as the input controller,the other as the robot arm identifier.At the same time,the hybrid algorithm was applied to the two neural network units to improve the impact of network structure parameters on the control and identification performance of the neural network.The simulation experiment was conducted in the Matlab environment,and the tracking effect was compared with that of RBF neural network controller.The simulation results show that when the mechanical arm terminal link is disturbed by uncertain factors,the error caused by the improved RBF neural network controller is small,the system reaction speed is fast,and the torque fluctuation is small.The end link of the mechanical arm adopts the improved RBF neural network controller,which has the ability of anti-interference and fast maintaining the stability of the system output.

关 键 词：多连杆 机械臂 径向基函数(RBF)神经网络 混合算法 仿真 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]