机械臂非奇异快速终端滑模迭代学习轨迹跟踪控制研究  

Research on Trajectory Tracking Control of Non-Singular Fast Terminal Sliding Mode Iterative Learning for Robot Manipulators

作　　者：陈涛 李晓娟[1] 刘建璇 王立忠 CHEN Tao;LI Xiaojuan;LIU Jianxuan;WANG Lizhong(School of Mechanical Engineering,Xinjiang University,830047,Urumqi)

机构地区：[1]新疆大学机械工程学院,乌鲁木齐830047

出　　处：《西安交通大学学报》2025年第1期125-135,147,共12页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(52265003);新疆维吾尔自治区天山雪松青年拔尖人才资助项目(20227SYCCX0061)。

摘　　要：针对机械臂建模参数精准性与扰动不确定性的精确轨迹跟踪控制问题,提出了一种非奇异快速终端滑模控制与迭代学习控制相融合的控制方法。首先,为保证跟踪误差的收敛速度,避免收敛中的奇异性问题,设计采用饱和函数趋近律的非奇异快速终端滑模控制器。其次,为进一步提高轨迹跟踪精度,设计误差迭代学习控制器,并对所设计的控制器进行了收敛性分析。最后,在Simulink软件中搭建所提方法的控制系统,进行迭代控制与对比控制仿真实验,并同步开展机械臂跟踪控制真机实验。结果表明:在迭代实验中,关节最大平均稳态误差提升了72%;在对比实验中,与比例微分(PD)型迭代学习控制和PD型线性滑模控制相比,最大平均稳态误差分别提升了97%、51%,最大响应调整时间分别减少70%、50%;在真机实验中,机械臂跟踪误差稳定在[-0.05,0.05] rad区间内。实验结果充分验证了所提控制方法的正确性与有效性,为解决机械臂轨迹跟踪中的不确定性问题提供了一种有效的控制方案。In response to the challenge of precise trajectory tracking control for robot manipulators,influenced by the accuracy of modeling parameters and disturbance uncertainties,a novel control approach that combines non-singular fast terminal sliding mode control with iterative learning control is presented.First,to ensure the convergence speed of tracking errors and prevent singularity issues during convergence,a non-singular fast terminal sliding mode controller employing the law of approach to saturation is designed.Second,to further improve trajectory tracking accuracy,an error iterative learning controller is developed,and the convergence of these controllers is analyzed.Finally,the control system based on the proposed method is implemented in Simulink for iterative and comparative control simulation experiments.Additionally,real-machine experiments for robot manipulator tracking control are carried out.The experimental results show that:in the iterative experiment,the maximum average steady-state error in joints increases by 72%;in the comparative experiment,compared to PD-type iterative learning control and PD-type linear sliding mode control,the maximum average steady-state error rises by 97%and 51%,respectively,while the maximum response adjustment time decreases by 70%and 50%,respectively;in the real-machine experiment,the robot manipulator tracking error stabilizes within the range of[-0.05,0.05]rad.These findings thoroughly validate the effectiveness and accuracy of the proposed control method,offering an effective control solution for addressing uncertainties in robot manipulator trajectory tracking.

关 键 词：机械臂 迭代学习控制 非奇异快速终端滑模控制 轨迹跟踪 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]