西安理工大学学报2016，Vol.32Issue(4)：437-442,6.DOI:10.19322/j.cnki.issn.1006-4710.2016.04.011
基于模型分块逼近的三关节机器人鲁棒滑模控制
Research on three-j oint robot’s robust sliding mode control based on model’s partitional approximating
摘要
Abstract
Generally,the dynamic model of robot with three-j oint is undetermined due to three-j oint robot’s uncertain structure parameters,working environment’s external interfere and struc-tural vibration.Accordingly,it is difficult to control the robot’s joints’position stabilizing and traj ectory tracking and controller’s design due to the dynamic model’s uncertainty.Therefore, three designed RBF(Radical Basis Function)neural networks are used to respectively model the three undetermined terms of the undetermined robot dynamic model,with partition approxima-ting the three-joint robot.Three undetermined terms’estimation information is respectively ob-tained,with the robot’s estimation model obtained.The neural networks’weights are obtained through the adaptive algorithm.The robust sliding mode control law is designed based on the ro-bot’s estimation model.The control law’s robust term is used to overcome the neural networks’ modeling error.The control system’s stability is proved by defining Lyapunov function.The simulation experiments test verifies that three joints can trace ideal trajectory and reach an ideal position in 1 s,and stabilization error and tracking error can fast and stably approximate to zero.关键词
三关节机器人/模型分块逼近/关节控制/RBF神经网络Key words
robot with three-joint/model’s partitional approximating/joints’control/RBF neu-ral network分类
信息技术与安全科学引用本文复制引用
马莉丽,钟斌..基于模型分块逼近的三关节机器人鲁棒滑模控制[J].西安理工大学学报,2016,32(4):437-442,6.基金项目
国家自然科学基金资助项目（51005246）；中国人民武装警察部队工程大学基础研究基金资助项目 （）
