New Results on Deterministic Learning of Sampled-Data Nonlinear Systems

Abstract

In this paper, our main concern is to establish new exponential stability-based identification results for a class of Euler nonlinear sampled-data systems using deterministic learning. At first, a new deterministic learning law is designed based on the Lyapunov function method. Rigorous analysis is provided to show that the resulting closed-loop linear time-varying (LTV) systems (containing tracking errors and parameter estimation errors) is exponentially stable. All the states of the closed-loop system converge to a small neighborhood around the origin exponentially. Thus, locally-accurate identification performance can be achieved under the new deterministic learning algorithm. Finally, simulation results on Duffing oscillator system are given to show the effectiveness of the proposed method.