PERFORMANCE MEASURES OF THE VARIABLE STRUCTURE FILTER

A new method for state estimation, referred to as the Variable Structure Filter (VSF), has recently been proposed. The VSF is a model based predictor-corrector method. It uses an internal model to provide an initial estimate of the states and subsequently refines this initial estimate by a corrective term that is a function of the system output and the upper bound of uncertainties. As such, the VSF can explicitly cater for uncertainties in its internal model. In this paper, a conceptual discussion of the VSF strategy and its performance in terms of stability, accuracy, and convergence is provided. The impact of modeling uncertainties on the performance of the VSF is discussed and quantified. The analysis is augmented by comparative simulation studies to further illustrate the concept.