Blending Based Multiple Model Pole Placement Adaptive Control of Multivariable Systems

In this article, we develop a blending based multiple model pole placement adaptive control (MMPPAC) scheme for square multivariable systems defined in state space. We consider linear, time invariant systems with known, compact, convex polytopic uncertainty. A state feedback controller is developed to drive all the states to the origin. The control architecture consists of a multiple model online identification scheme, and a control scheme that guarantees boundedness of all closed-loop signals, and asymptotic regulation of the plant’s state vector to the origin. Numerical simulations of motion control of lateral vehicle dynamics illustrate the stability and efficacy of the proposed MMPPAC scheme, with robustness to dynamical uncertainty and measurement noise.