Robust recursive instrumental variable method with modified weights for bilinear system identification

Based on P.J. Huber's minimax principle (1981), both symmetric and nonsymmetric robust recursive instrumental variable methods with modified weights are developed for bilinear system identification. A novel approach is used to prove a general convergence theorem for the proposed robust methods. The robustness and convergence of the proposed approaches are demonstrated by comparison with the corresponding nonrobust approaches in two simulation examples. It is shown that the suggested methods are far superior to the nonrobust ones.<
>