Control and Communication-Schedule Co-design For Networked Control Systems

Abstract: In a networked control system (NCS), the control loop is closed through a communication medium. This means that sensor measurements and/or control signals can be exchanged through a communication link. NCSs have many benefits, such as wiring reduction (elimination in the case of wireless communication), installation cost reduction, and simplification of upgrades and restructuring. However, network congestion, impairments of the wireless links (such as bandwidth limitations, packet losses, delays, and noises) may degrade system performance and even cause instability. These issues have motivated a great deal of research over the past 20 years and have given rise to a number of approaches to prevent congestion and compensate for delays and/or packet losses. An interesting class of NCSs that has not received enough attention is an NCS whose systems are uncertain and subject to state and inputs hard constraints.These hard constraints may stem from the system itself, its environment, or be proposed by the designer in order to guarantee safety or a certain performance. The contribution of this thesis is introducing a design framework that guarantees robust constraint satisfaction for a class of multi-agent NCSs with a shared communication medium that is subject to bandwidth limitation and prone to packet losses. The proposed framework is built upon reachability analysis to determine the communication demand for each system such that local constraints are satisfied and scheduling techniques to guarantee satisfaction of the communication demands. The thesis explores offline and online scheduling designs under various communication topologies, optimal control designs under state and output feedback, and scheduling and control co-design for NCSs with hard constraints.