Adaptive Finite-Time Parameter Estimation and Control for Constrained Robotic Systems

摘要

In adaptive robot control, accurate and fast parameter estimate is crucial since it can improve the control performance for the robotic systems. Most existing works only ensure accurate parameter estimates with exponential convergence rate or obtain finite-time (FT) convergence with nonzero estimation error. This paper proposes a novel adaptive FT parameter estimation and control scheme for uncertain robotic systems with the consideration of unavoidable joint position constraints. A new parameter estimation law is constructed by exploiting both online history data and current data such that accurate parameter estimation is ensured with a weak interval excitation (IE) condition instead of the typically stringent persistent excitation (PE) condition. Then, the parameter updating law is integrated into a nonsingular integral terminal sliding mode based adaptive controller to ensure zero parameter estimation error and tracking error in finite time without singular problem. Besides, a modified unified time-varying asymmetric barrier function (UTABF) is employed to handle the joint position constraints directly. The UTABF can not only deal with both constrained and unconstrained cases uniformly but also make the proposed control more user-friendly in design and less demanding in implementation. The effectiveness of the developed control scheme is verified by theoretical analysis and experiments.