Sliding-mode observer for nonlinear systems with unknown inputs and noisy measurements
Abstract
An estimator for nonlinear systems with unmatched unknown inputs is presented. The observer is of sliding mode type and guarantees the estimation of the system's estate and unknown unmatched inputs up to an arbitrarily small tolerance. It is also assumed that measurements are contaminated by noise. Adaptation is used to update some of the estimator's gains online, which compensate for the system's uncertainties and noise. The estimator design and the convergence analysis follows the trend of literature on discontinuous sliding-mode observers however, our design results in a continuous dynamic system which steers the trajectories to an arbitrarily small neighbourhood of the sliding mode.