A Comparative Study of the Variable Structure Filter

A new strategy for state estimation has recently been proposed referred to as the Variable Structure Filter. This filter is model-based and can be applied to discrete-time systems that are subject to random disturbances and measurement noise. The advantage of the Variable Structure Filter is that it can be formulated to accommodate and explicitly account for modeling uncertainties. A proof of stability for the filter is provided. In this paper, the Variable Structure Filter is applied to a third order system with an uncertain mathematical model. A performance comparison is presented by comparing the state estimation errors of the Variable Structure Filter to those obtained from an equivalent Kalman Filter and a digital Infinite Impulse Response filter.