Passivity analysis and passivation in the design of cyber-physical systems.

摘要

This dissertation focuses on the analysis and control of cyber-physical systems (CPS) using dissipativity and passivity theory. Cyber-physical systems, as a new generation of systems with integrated computational and physical capabilities, present significant challenges in control design and analysis, due to non-traditional modeling, uncertain environment and highly coupled discrete-event and continuous-time dynamics. On the other hand, it is well known that passive and dissipative systems have modeling, compositionality advantages and stability-guaranteed performance, which are desirable requirements in CPS design. However, it is not straightforward to apply dissipativity and passivity theory to CPS directly in general.;The main contribution of this dissertation is to provide systematic and computational methods of passivity analysis and passivation for continuous, networked and hybrid dynamical systems, which provide modeling foundations for CPS. These methods are originally developed for classical nonlinear systems. They include passivity analysis and passivation for interconnected systems using passivity indices and a transformation-based passivation scheme for individual systems. Later, it is shown that the proposed methods can address the issues in the design of CPS, by considering hybrid systems and networked control systems (NCS), respectively. For hybrid systems, the transformation-based passivation scheme provides valuable results on preserving passivity of switched systems under quantization. For networked control systems, the problems of passivity analysis and passivation using passivity indices for interconnected event-triggered feedback systems are investigated. The co-design of passivity levels and event-triggering conditions demonstrates how the trade off between required passivity levels and communication resource utilization can be achieved in NCS.;Overall, this dissertation provides new approaches to passivity analysis and passivation of CPS with the focus being on hybrid systems and networked control systems. Numerical simulations and relevant examples are also provided to demonstrate the practical applications of these methods.