Sampled-data Networked Control Systems: A Lyapunov-Krasovskii Approach

The main goal of this thesis is to develop computationally efficient methods for stability analysis and controller synthesis of sampled-data networked control systems. In sampled-data networked control systems, the sensory information and feedback signals are exchanged among different components of the system (sensors, actuators, and controllers) through a communication network. Stabilization of sampled-data networked control systems is a challenging problem since the introduction of multirate sample and holds, time-delays, and packet losses into the system degrades its performance and can lead to instability. A diverse range of systems with linear, piecewise affine (PWA), and nonlinear vector fields are studied in this thesis. PWA systems are a class of state-based switched systems with affine vector field in each mode. Stabilization of PWA networked control systems are even more challenging since they simultaneously involve switches due to the hybrid vector fields (state-based switching) and switches due to the sample and hold devices in the network (event-based switching). The objectives of this thesis are: (a) to design controllers that guarantee exponential stability of the system for a desired sampling period; (b) to design observers that guarantee exponential convergence of the estimation error to the origin for a desired sampling period; and (c) given a controller, to find the maximum allowable network-induced delay that guarantees exponential stability of the sampled-data networked control system. Lyapunov-Krasovskii based approaches are used to propose sufficient stability and stabilization conditions for sampled-data networked control systems. Convex relaxation techniques are employed to cast the proposed stability analysis and controller synthesis criteria in terms of linear matrix inequalities that can be solved efficiently