引用本文:梁相龙,姚建勇.带有输出约束的液压机械臂自适应神经网络力跟踪控制[J].控制理论与应用,2025,42(1):138~148.[点击复制]
LIANG Xiang-long,YAO Jian-yong.Adaptive neural network force tracking control of hydraulic manipulators with output constraints[J].Control Theory and Technology,2025,42(1):138~148.[点击复制]
带有输出约束的液压机械臂自适应神经网络力跟踪控制
Adaptive neural network force tracking control of hydraulic manipulators with output constraints
摘要点击 3330  全文点击 37  投稿时间:2022-08-15  修订日期:2024-11-28
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2023.20725
  2025,42(1):138-148
中文关键词  液压机械臂  导纳控制  动态面控制  神经网络  力跟踪控制  未知环境
英文关键词  hydraulic manipulator  admittance control  dynamic surface control  neural network  force tracking control  unknown environment
基金项目  国家重点研发计划项目(2021YFB2011300), 国家自然科学基金项目(52075262)资助.
作者单位E-mail
梁相龙 南京理工大学 xlliang.njust@gmail.com 
姚建勇* 南京理工大学 jerryyao.buaa@gmail.com 
中文摘要
      为解决带有输出约束的液压机械臂动力学模型未知问题并提升液压机械臂在未知环境下的力跟踪性能, 本文提出了一种基于积分障碍李雅普诺夫函数的自适应神经网络导纳控制方法. 首先, 分析了液压机械臂的机械和液压系统动力学模型, 根据阻抗控制原理, 提出了基于环境参数估计的参考轨迹自适应生成方法; 然后, 考虑系统 输出受限和机械系统动力学模型未知, 利用径向基函数神经网络设计自适应神经网络控制器; 同时, 引入动态面控制方法以避免对虚拟信号进行直接求导, 并通过李雅普诺夫方法分析了闭环控制系统的稳定性; 最后, 利用MATLAB/Simulink, Simscape Multibody 和Simscape Fluids仿真平台对液压机械臂进行仿真研究, 结果表明所设计的控制律对未知机械系统动力学具有良好的鲁棒性, 可以实现良好的位置和力跟踪控制, 且确保系统输出不超过预设的范围.
英文摘要
      In order to solve the problem of unknown dynamics of the hydraulic manipulator with output constraints and improve the force tracking performance of the hydraulic manipulator in unknown environments, an adaptive neural network admittance control method integrating integral barrier Lyapunov function is proposed in this paper. Firstly, the mechanical and hydraulic system dynamics of the hydraulic manipulator are addressed, and according to the principle of impedance control, an adaptive generation method of reference trajectory based on environment parameters estimation is developed. Then, an adaptive radial basis function neural network tracking control is developed for the hydraulic manipulator with unknown mechanical system dynamics and output constraints. Meanwhile, the dynamic surface control approach is introduced to circumvent the direct derivation of virtual signals, and the stability of the closed-loop control system is analyzed via Lyapunov technique. Finally, using the MATLAB/Simulink, Simscape Multibody and Simscape Fluids simulation platforms to simulate the hydraulic manipulator, the results indicate that the developed control law has good robustness with respect to unknown mechanical system dynamics, achieves satisfactory position and force tracking performance, and ensures the system output does not exceed the prescribed range.