控制理论(社会学)
非线性系统
计算机科学
李雅普诺夫函数
自适应控制
鲁棒控制
控制(管理)
稳健性(进化)
Lyapunov重新设计
控制系统
理论(学习稳定性)
控制工程
Lyapunov优化
非线性控制
非线性动力系统
滑模控制
Lyapunov稳定性
李雅普诺夫方程
线性系统
作者
Zhe Zhang,何在通,Can Ding,Zhiqiang Miao,Min Liu,Yaonan Wang
标识
DOI:10.1109/tie.2026.3699855
摘要
For complex dynamic systems subject to unknown external disturbances and motion constraints (with industrial manipulator motion control as a representative application), this article aims to ensure motion safety while maintaining high-quality trajectory tracking. To this end, we establish an online robust control framework: by learning models to estimate time-varying disturbances in real time, and designing a dynamically updated mechanism to adaptively adjust safety margins, thereby avoiding failures or excessive conservatism caused by fixed margins. Concurrently, we propose a hierarchical energy safety constraint that, on the one hand, limits overall motion intensity, and on the other, focuses solely on constraining boundary normal motions most likely to cause boundary violations. This enhances motion flexibility near boundaries without compromising safety. The aforementioned stability and safety conditions are unified into a real-time solvable optimisation decision to generate safety-critical control inputs with minimal deviation from the desired control. Simulation results demonstrate that during obstacle avoidance and time-varying disturbance tasks, this method consistently operates outside the safety boundary. Compared with alternative approaches, it exhibits reduced tracking error, faster convergence, and smoother disturbance responses, thereby validating the proposed framework’s integrated advantages in safety, robustness, and performance.
科研通智能强力驱动
Strongly Powered by AbleSci AI