重庆理工大学学报2025，Vol.39Issue(1)：83-92,10.DOI:10.3969/j.issn.1674-8425(z).2025.01.011
机械臂新型固定时间非奇异终端滑模控制
A new fixed time nonsingular terminal sliding mode control for robot arms
摘要
Abstract
During the actual operation of a robotic arm with unknown initial state,there exist nonlinear factors such as unknown external disturbances and modeling errors,making the robotic arm unable to work accurately and quickly.To address the issue,we propose a fixed time algorithm based optimized NTSMC scheme for robotic arms.First,new segmented sliding variables are built to solve the singularity problem caused by terminal sliding mode control and achieve faster convergence speed within a fixed time.Then,a reinforcement learning algorithm based on actor critic architecture is adopted to approach and compensate for model uncertainty and external disturbances in real-time.Next,with the consideration of input saturation,an auxiliary system is designed to alleviate it.Finally,comparative experiments are conducted.Our method delivers superior performances with faster fixed time convergence speed and higher control accuracy under the same control parameters.关键词
机械臂/固定时间算法/NTSMC/强化学习Key words
robotic arm/fixed time algorithm/NTSMC/reinforcement learning分类
信息技术与安全科学引用本文复制引用
杨佳,杨理,许强,余希瑞,吴佩林..机械臂新型固定时间非奇异终端滑模控制[J].重庆理工大学学报,2025,39(1):83-92,10.基金项目
国家自然科学基金项目(52177129) （52177129）
重庆市教委科学技术研究重点项目(KJZD-K201901102) （KJZD-K201901102）
重庆市技术创新与应用发展专项面上项目(cstc2020jscx-msxmX0210) （cstc2020jscx-msxmX0210）
重庆理工大学研究生创新项目(gzlcx20233095) （gzlcx20233095）
