PhD Oral Exam - Amir Amini, Electrical and Computer Engineering
This event is free
School of Graduate Studies
When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.
Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.
Recently, distributed multi-agent systems (MAS) have been widely studied for a variety of engineering applications, including cooperative vehicular systems, sensor networks, and electrical power grids. To solve the allocated tasks in MASs, each agent autonomously determines the appropriate actions using information available locally and received from its neighbours. Many cooperative behaviours in MAS are based on a consensus algorithm. Consensus, by definition, is to distributively agree on a parameter of interest between the agents. Depending on the application, consensus has different configurations such as leader-following, formation, synchronization in robotic arms, and state estimation in sensor networks. Consensus in MASs requires local measurements and information exchanges between the neighbouring agents. Due to the energy restriction, hardware limitation, and bandwidth constraint, strategies that reduce the amount of measurements and information exchanges between the agents are of paramount interest. Event-triggering transmission schemes are among the most recent strategies that efficiently reduce the number of transmissions. This dissertation proposes a number of event-triggered consensus (ETC) implementations which are applicable to MASs. Different performance objectives and physical constraints, such as a desired convergence rate, robustness to uncertainty in control realization, information quantization, sampled-data processing, and resilience to denial of service (DoS) attacks are included in realization of the proposed algorithms. A novel convex optimization is proposed which simultaneously designs the control and event-triggering parameters in a unified framework. The optimization governs the trade-off between the consensus convergence rate and intensity of transmissions. This co-design optimization is extended to an advanced class of event-triggered schemes, known as the dynamic event-triggering (DET), which is able to substantially reduce the amount of transmissions. In the presence of DoS attacks, the co-design optimization simultaneously computes the control and DET parameters so that the number of transmissions is reduced and a desired level of resilience to DoS is guaranteed. In addition to consensus, a formation-containment implementation is proposed, where the amount of transmissions are reduced using the DET schemes. The performance of the proposed implementations are evaluated through simulation over several MASs. The experimental results demonstrate the effectiveness of the proposed implementations and verify their design flexibility.