Adaptive Parameter Estimation for Nonlinear Systems with Relaxed PE Condition

This paper presents a novel adaptive parameter identification method based on an augmented regression for linearly parameterized nonlinear systems, which relaxes the persistence excitation (PE) condition. By setting filter operations, an augmented system is constructed. Then, the augmented system is fused to decouple the estimated parameters. Furthermore, an adaptive estimator based on the parameter estimation error is designed to guarantee that the parameters exponentially converge to the true values. In this parameter estimation framework, not only the required PE condition can be relaxed but also the parameter estimation performance can be improved. Finally, comparative simulation results are provided to illustrate the effectiveness of the proposed method.