Consensus And H_∞Filtering Of Multi-agent Systems Based On Various Data-triggering Machnisms

The increasing technological advancement of signal processing and communication network makes it feasible that multi-agent systems tend to be operated by digital con-trollers widely in some modern applications. In this case, information transmission is usu-ally either time-triggered or event-triggered. The advantages of the time-triggered scheme lies in easy implementations of its analysis and design. The event-triggered scheme can benefit from some savings of network communication resources. Thus, in this thesis these two information transmission schemes will be respectively employed to investigate con-sensus and H∞filtering of multi-agent systems. Based on analysis and summary of topic backgrounds and the existing literatures, the contributions of this thesis can be concluded as follows:1. For a class of multi-agent systems with Lipschitz nonlinear dynamics and Marko-vian switching topologies, a sampled-data consensus protocol depending on topology switching is proposed by considering effects of communication delays. Employing Lyapunov-Krasovskii functional and weak infinite operator method, a sufficient condition, which can ensure exponential mean-square consensus of multi-agent systems, is derived. Correspondingly, the topology-dependent sampled-data consensus controller is given.2. By introducing two buffers to deal with asynchronous problem arising from network-induced delays, a new network-based multi-agent system framework is estab-lished. Under this framework, a new network-based consensus protocol is proposed. By constructing a novel Lyapunov-Krasovskii functional with graph information, a delay-dependent criterion for consensus of multi-agent systems is given. Based on this criterion, a sufficient condition on the existence of network-based consensus controller is obtained.3. A new distributed event-triggered transmission strategy for sampled-data consen-sus of multi-agent systems is presented. Under this strategy, an event-triggered consensus protocol is proposed, with which consensus of the multi-agent system can be achieved asymptotically. A co-design algorithm for obtaining the consensus controllers and the threshold parameters of event-triggering conditions in a unified framework is proposed. Numerical simulations illustrate that the quantity of sampled-data transmission can be sig-nificantly reduced by using the proposed strategy.4. Distributed H∞consensus filtering in sensor networks is investigated by adopting different transmission triggering mechanisms. Based on the time-triggered and event-triggered schemes, the constructions of H∞consensus filters are respectively given, and then a method is obtained to design the H∞consensus filter parameters and threshold parameters of triggering conditions.