ABSTRACT
In this chapter, we continue further into the decentralized-control techniques for interconnected systems, where we focus herein on methods for designing classes of reliable decentralized controllers to deal with possible actuator and/or sensor failures. Thus we study hereafter the problem of designing reliable decentralized feedback control for linear interconnected systems where we focus initially on subsystems with internal time-delays and additional timedelay couplings, under actuator and/or sensor failures. Then we specialize the result to delay-free subsystems. We equally treat continuous-and discrete-time system representations. The failures are described by a model that takes into consideration possible outages or partial failures in every single actuator/single sensor of each decentralized controller. The decentralized control design is performed through two steps. First, a decentralized stabilizing reliable feedback control set is derived at the subsystem level through the construction of appropriate Lyapunov-Krasovskii functional (LKF) and, second, a feasible linear matrix inequalities procedure is then established for the effective construction of the control set under different feedback schemes. Two schemes are considered: the first is based on state-measurement and the second utilizes static output-feedback. The decentralized feedback gains in both schemes are determined by convex optimization over linear matrix inequalities (LMIs).
