Robust adaptive finite-time parameter estimation for linearly parameterized nonlinear systems

This paper studies a novel adaptive parameter estimation framework for linearly parameterized nonlinear systems. Appropriate parameter error information is derived by defining auxiliary filtered variables and used to drive the parameter adaptation, which guarantees exponential error convergence. The proposed method is further improved via a sliding mode approach to achieve finite-time (FT) error convergence. The case with bounded disturbances or noises is also studied. The parameter estimation is obtained without using the state derivatives and is independent of observer/predictor design. The online computation of the regressor matrix inverse can be avoided. Simulation examples are included to illustrate the effectiveness.