Veranstaltung: Cooperative Control of Multi-Agent Systems
- Vorlesung und Praxisübungen
- Prof. Dr.-Ing. Jan Lunze
- Dr. Li (extern)
- Angeboten im:
Termin nach Absprache mit dem Dozenten.
This is an introductory course about cooperative control of multi-agent systems and multi-robot systems. A networked multi-agent system consists of a set of dynamical agents that interact over a communication network for collective behaviors; each agent has its own state variable and dynamics. The communication network is modeled as a graph with directed edges or links corresponding to the allowed information flow between agents. Information in the communication network only travels directly between immediate neighbors in the graph. A fundamental problem is the design of distributed protocols or control laws that guarantee certain collective behaviors of the agents (for example, consensus or synchronization in the sense that the states of all the agents reach a same value). For cooperative control of agents on graphs, all control protocols must be distributed in the sense that the control law of each agent is only allowed to depend on information from its immediate neighbors in the graph . The study of networks of coupled dynamical systems arises in many fields of research with many fascinating topics, which have many interesting applications, particularly in multi-robot systems in control field. This course will introduce the basic concepts, the topics of some control problems, and future research trends.
- Introduction to cooperative control
- Network models
- Consensus problem
- Synchronization mainly for coupled oscillators and chaoticc systems and the variants
- Flocking behavior of holonomic agents
- Flocking behavior of nonholonomic agents
- Formations of vehicles in cyclic pursuit
- Formation control of holonomic agents
- Formation control of nonholonomic mobile robots or vehicles
- Advanced topics and conclusion of the course
For further information on these lecture please refer to the website of the Institute of Automation and Computer Control (www.atp.rub.de).