Stable Weighted Multiple Model Adaptive Control With Improved Convergence Rate

The paper is concerned with the convergence rate of the weighting algorithm and the stability of weighted multiple model adaptive control (WMMAC). First a new weighting algorithms with convergence assurance is proposed. Second, the stability of the WMMAC system is proved based on virtual equivalent system (VES) concept and methodology, which is independent of specific 'local' control strategy. The obtained results indicate that the stability of a WMMAC system depends mainly on the 'stabilizing' property of each local controller and the convergence of weighting algorithm, i.e., the weight corresponding to the true (or the closest) model of the plant converges to 1, and others converge to 0.