Decentralised feedback controllers for uncertain interconnected dynamic systems

This paper presents a method for the design of decentralised controllers for interconnected dynamic systems with structured uncertainties. The design problem is formulated as a constrained linear quadratic optimal regulator problem, where the cost function is designed to incorporate the system uncertainties. Conditions are then derived for choosing the state and control weighting matrices that guarantee closed-loop stability in the presence of the uncertainties. Once these matrices are appropriately chosen, a generalised Riccati equation is solved and a robust decentralised controller is obtained. Two numerical examples are given to illustrate the controller design method.