A state-space approach to the multivariable continuous-time self-tuning control

In this paper, we propose a new self-tuning control for continuous-time linear time invariant MIMO systems. We first develop a self-tuning multi-input state feedback control, and then develop a self-tuning control for a general MIMO system with an input-output description. In the latter case, we show that once the system's dynamics of the MIMO system is properly parameterized into a nonminimal state space description, the self-tuning control problem can be transformed into that in the former case, and therefore be solved similarly. The unique feature of our state space approach compared with the conventional polynomial equation approach is that the only a priori information on the system required is an upper bound of the system's observability indices.