Stochastic Adaptive Nonlinear Control With Filterless Least Squares
Wuquan Li, Miroslav Krstić
Abstract
For stochastic strict-feedback nonlinear systems with unknown parameters in the drift terms or the diffusion terms, we develop new least-squares identification schemes without regressor filtering. A key new ingredient in the proposed estimator design is a weighted term with design parameters, which is introduced to deal with the nonlinear terms and stochastic noise. With such an estimator, new adaptive controllers are designed to guarantee that the equilibrium at the origin of the closed-loop system is globally stable in probability, and the states are regulated to zero almost surely. Besides, by suitably selecting the estimator parameters, we prove that the proposed least-squares estimators are convergent, as well as strongly consistent in some special cases. Finally, two simulation examples are given to illustrate the least-squares identification and the adaptive control design.