摘要
为了提高工业机器人双臂控制精度,考虑瞬态性能设计了一种自适应神经网络反步控制方法。首先建立了工业机器人双臂的数学模型,然后设计了瞬态性能函数,根据瞬态性能函数分步设计轨迹控制律和转速控制律,并利用神经网络逼近未知项,最终利用所设计的自适应律估计出神经网络权值向量,实现了工业机器人双臂高精准控制。仿真验证结果表明,所设计的控制方法能够确保工业机器人双臂所夹持目标物稳定、准确跟踪轨迹指令信号,轨迹跟踪最大误差仅为0.2 cm,内力跟踪的最大误差仅为0.1 N。在实测验证中,定位的平均误差仅为0.33 cm,内力的平均误差仅为0.21 N,单次运行平均耗时仅为1.50 s,控制精度和运行效率均得到了大幅提升。
To improve the control accuracy of industrial robot dual-arm,an adaptive neural network back-stepping control method was designed considering the transient performance.Firstly,the mathematical model of the robot dual-arm was established,and then the transient performance function was designed.According to the transient performance function,the trajectory control law and speed control law were designed step by step,and the neural network was used to approximate the unknown term.Finally,the weight vector of the neural network was estimated by the designed adaptive law to realize the high precision control of the industrial robot dual-arm.The simulation results show that the designed control method can ensure the target clamped by dual-arm of the industrial robot track the trajectory command signal stably and accurately,and the trajectory tracking maximum error is only 0.2 cm,and the maximum error of internal force tracking is only 0.1 N.In actual verification,the average error of positioning is only 0.33 cm,and the average error of internal force is only 0.21 N,and the average running time is only 1.50 s.The control accuracy and operation efficiency have been greatly improved.
作者
熊蕊
XIONG Rui(School of Information Engineering,Zhengzhou Tourism College,Zhengzhou 450000,China)
出处
《现代制造工程》
CSCD
北大核心
2022年第8期53-59,共7页
Modern Manufacturing Engineering
基金
河南省科技攻关项目(182102210150)。
关键词
工业机器人
双臂
瞬态性能
自适应神经网络
反步控制
industria robot
dual-arm
transient performance
adaptive neural network
back-stepping control
