Identification for systems with bounded noise

In this paper, we present a set-membership identification algorithm for systems with unknown but bounded disturbance. The algorithm contains a weighting factor which is selected according to whether the new observed data contains sufficient information. The proposed approach ensures that the estimation error is bounded and nonincreasing. Furthermore, it is shown that the parameter estimates provided by the algorithm will converge to a region containing the true parameters, and its upper bound is also given.