多关节机械臂的分数阶滑模变结构神经网络自适应控制  

Fractional Sliding Mode Variable Structure Neural Network Adaptive Control of Multi-joint Manipulator

作　　者：李正楠 张锦 殷玉枫[1] 武奎扬[1] 孙煊广 LI Zhengnan;ZHANG Jin;YIN Yufeng;WU Kuiyang;SUN Xuanguang(College of Mechanical Engineering,Taiyuan University of Science and Technology,Taiyuan 030024,China;Shanxi Transportation Vocational College,Taiyuan 030031,China)

机构地区：[1]太原科技大学机械工程学院,太原030024 [2]山西交通职业技术学院,太原030031

出　　处：《机械科学与技术》2021年第2期270-275,共6页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金项目(U1610118);山西省交通运输厅科技计划项目(2019-1-9);山西省研究生创新项目(2019SY473)。

摘　　要：在分数阶滑模控制方法的基础,设计了新型的滑模函数趋近律,提升了多关节机械臂轨迹跟踪控制系统的鲁棒性和响应速度。机械臂在建模过程中存在多种建模误差,工作过程中存在外界干扰等不确定因素,因此采用神经网络对各种不确定项进行逼近,对径向基函数神经网络自适应控制中存在的误差,提出了神经网络误差的自适应补偿项,减小了径向基函数神经网络对不确定项的逼近误差。结果表明,该控制算法有效地提高了轨迹的跟踪速度和跟踪精度,并降低了系统中存在的抖颤。Based on the fractional sliding mode control method,the new approximation law of sliding mode function is designed to improve the robustness and response speed of the trajectory tracking control system of a multi-joint manipulator.There are various modeling errors in the modeling process of the multi-jointmanipulator and uncertain factors such as external interference in the working process.The neural network is used to approximate various uncertainties,and the adaptive compensation term of the error in the radial basis function neural network adaptive control is proposed.The approximation error of the radial basis function neural network is reduced.The results show that the control algorithm effectively improves the tracking speed and tracking accuracy and reduces the chattering in thecontrolsystem.

关 键 词：多关节机械臂 滑模控制 神经网络 自适应 

分 类 号：TH113.2[机械工程—机械设计及理论]