刘昱,孙岩,张钰婷,等.气动六自由度平台机器人无模型迭代学习控制[J].河南理工大学学报（自然科学版）,2026,45(6):136-146.

Liu Y, Sun Y, Zhang Y T,et al.Model-free iterative learning control of a pneumatic six-degree-of-freedom platform robot[J].Journal of Henan Polytechnic University(Natural Science) ,2026,45(6):136-146.

气动六自由度平台机器人无模型迭代学习控制

刘昱, 孙岩, 张钰婷, 杨洁

北京石油化工学院 自动化系，北京 102617

摘要: 目的 并联机器人具有刚度大、承载能力强、精度高、动力性能好等优点，具有广阔的发展前景。为进一步提高气动六自由度平台机器人的位姿控制精度，针对无模型自适应迭代学习控制开展研究。  方法 首先，针对金属间隙密封气缸构建三阶数学模型，为后续控制算法提供一个精确的参考模型；其次，在深入分析金属间隙密封气缸驱动的气动伺服系统基础上，提出一种基于无模型自适应迭代学习控制的六自由度平台机器人位姿控制策略并设计控制器；最后，基于MATLAB验证该控制器的有效性。  结果 仿真结果表明，无模型迭代学习控制方法具有更强的鲁棒性：在阶跃响应仿真中，无模型迭代学习控制方法比PID方法的平均上升时间减小59%，平均调节时间缩短40%，均方误差的平均降幅约为31%，最大误差的平均降幅约为63%；在正弦跟踪实验中，无模型迭代学习控制方法比PID方法的均方误差平均降幅约为56%~57%，最大误差平均降幅约为47%~52%。  结论 对于气动六自由度平台机器人，无模型迭代学习控制方法比传统PID控制具有更加优异的控制性能，有较好的应用前景。

关键词:六自由度平台机器人;无模型控制;迭代学习控制;PID控制;金属间隙密封气缸

doi:10.16186/j.cnki.1673-9787.2023120004

基金项目:国家自然科学基金资助项目（6227020586）

收稿日期:2023/12/01

修回日期:2024/04/08

出版日期:2026/10/09

Model-free iterative learning control of a pneumatic six-degree-of-freedom platform robot

Liu Yu, Sun Yan, Zhang Yuting, Yang Jie

Department of Automation， Beijing Institute of Petrochemical Technology， Beijing 102617，China

Abstract: Objectives Parallel robots possess advantages such as high stiffness， strong load-carrying capacity， high precision， and excellent dynamic performance， offering broad development prospects. To further improve the pose control accuracy of a pneumatic six-degree-of-freedom （6 DOF） platform robot， this study focuses on model-free adaptive iterative learning control.  Methods First， a third-order mathematical model is established for the metal-clearance-sealed cylinder to provide a roughly accurate reference model for subsequent control algorithms. Second， based on an in depth analysis of the pneumatic servo system driven by the metal-clearance-sealed cylinder， a pose control strategy for the 6-DOF platform robot based on model-free adaptive iterative learning control is proposed and its controller is designed. Finally， the effectiveness of the controller is verified using MATLAB. Results Simulation results demonstrate that the model free iterative learning control method exhibits stronger robustness. In step response simulations， compared with the PID method， the model-free iterative learning control method reduces the average rise time by approximately 59% and shortens the average settling time by about 40%，the average reduction in the mean square error is approximately 31%，the average reduction in the maximum error is approximately 63%. In sinusoidal tracking experiments， compared with the PID method， the model-free iterative learning control method achieves average reductions of approximately 56%~57% in the mean square error and 47%~52% in the maximum error.  Conclusions For the pneumatic 6-DOF platform robot， the model-free iterative learning control method exhibits superior control performance compared to traditional PID control and shows promising application prospects.

Key words: six-degree-of-freedom platform robot; model-free control; iterative learning control; PID control; metal clearance-sealed cylinder