Multiple Model Reference Adaptive Tracking Control of Multivariable Systems with Blending

Abstract

This paper develops a multiple fixed model blending-based adaptive online parameter identification scheme and a multiple model reference adaptive control (MMRAC) for multiple input, multiple output (MIMO) linear time-invariant (LTI) systems with polytopic parameter uncertainty. The parameter identification scheme produces bounded estimates that asymptotically converge to the unknown system’s matrices. The proposed MMRAC is developed by combining the proposed adaptive parameter identification scheme with a state-space model reference control approach. This MMRAC scheme is proven to asymptotically track signals generated by a LTI reference model. A set of simulation test results are presented to illustrate the stability and effectiveness of the proposed MMRAC scheme.