EE Special Seminar Cooperative Control of Multi-Agent Dynamical Systems
EE Special Seminar
Cooperative Control of Multi-Agent Dynamical Systems
Dr. Nguyen Dinh Hoa, Dept. of Automatic Control, Hanoi Univ. of Science and Technology
Friday June 27, 2014, 10.00-11.30 am.
Room 210, EE Bldg., EE Dept., Faculty of Engineering, Chulalongkorn University
This talk presents several results on the control of multi-agent dynamical systems which includes two parts. The first part focuses on the design of decentralized stabilizing controllers for linear multi-agent systems. On the other hand, the second part devotes to the analysis of synchronization in nonlinear multi-agent systems.
In the first part, we consider multi-agent systems in which each agent is represented by a general LTI system and the agents are collaborated in the following sense. Each agent sends out its information and receives the information from other connected agents to cooperatively achieve a global objective besides the local objective of each agent. One typical global objective is the stability of the whole multi-agent system which not only depends on the dynamics of each agent but also depends on the structure of information exchange among agents. Thus, for a given multi-agent system in which the agents are initially independent, our goal is to design a stabilizing controller such that the agents interconnect in a decentralized fashion with a specifically chosen or given information structure. In order to do so, we propose a method based on the LQR approach with a properly selected performance index such that both global and local objectives can be achieved by the designed decentralized LQR controller. Some numerical examples are introduced to illustrate the effectiveness of the proposed method.
In the second part, we present a study on the synchronization of nonlinear multi-agent systems in which each agent is an oscillator and the agents are coupled through a nonlinear function. Since each agent oscillates with a distinct frequency, the global objective for this type of multi-agent system is to make all agents oscillating at the same frequency or in other words synchronized. Then the question here is under which condition of the information exchange among agents, the synchronization can occur, with the assumptions that the natural frequencies of agents and the nonlinear coupling function are known. One answer for this question will be provided in this talk revealing that the information exchange among agents should be large enough to compensate the difference among the natural frequencies of agents.
Dinh Hoa Nguyen received his B.S degree in Automatic Control from Hanoi University of Science and Technology in 2007, M.Eng degree in Electrical Engineering in 2009 from Chulalongkorn University, and PhD degree in Information Science and Technology, The University of Tokyo in 2014. In 2009, he was awarded the International Scholarship in the ICCAS-SICE International Joint Conference, and the IEEE Control Systems Society Travel Support for the 7th Asian Control Conference. His research interests include multi-agent control systems, cooperative control, robust control, iterative learning control and optimal/optimization control.