Exact description of feasible parameter set for bounded-noise models

The problem of estimating feasible parameter sets for linear discrete-time systems with unknown-but-bounded noise is investigated in this paper. Considering that the feasible parameter set of a linear discrete-time system with bounded noise is a convex polytope and that a convex polytope can be transformed into a polyhedral cone with its dimension increased by one, we transform the problem of estimating feasible parameter sets into calculating polyhedral cones. An algebra-based recursive method is proposed to identify the vertices of the polytope which is related to the polyhedral cone and a parametric description of the feasible parameter set is then obtained. Simulation examples show the effectiveness of the proposed method.