Adaptive internal model control for multiple-input multiple-output systems

The Internal Model Control (IMC) structure has become an extremely popular one in process control applications. Previous research in the literature has focused on nonadaptive IMC. Adaptive IMC has also been empirically introduced using neural networks and fuzzy logic. Stability issues in the area of adaptive IMC control have been investigated for single-input single-output (SISO) systems as well as the special case of decentralized multi-input multi-output (MIMO) systems. This dissertation extends the earlier stability work to the domain of general MIMO continuous-time systems. An adaptive observer is developed and analyzed for the identification of an m x r multi-input multi-output (MIMO) system with unknown parameters. Using the adaptive observer as the internal model, and the Certainty Equivalence principle, a continuous-time adaptive internal model controller is then proposed for the control of a general m x r MIMO system with unknown parameters. The resulting closed loop system is analyzed for stability and the efficacy of the adaptive control schemes is demonstrated via simulations.