Non-Uniformly Sampled Coordinated Control of Multi-Agent Systems and Its Applications

Project: Research

View graph of relations

Description

A multi-agent dynamic system is a dynamic system consisting of a number ofsubsystems named agents. Coordinated control of multi-agent systems has attractedgreat attention in the past few years because these systems have advantages such ashigh efficiency and great flexibility. Typical application examples include removal of alarge amount of hazardous material by a team of mobile robots and surveillance of anenormous area by a group of artificial satellites. In practice it is common for simpleembedded microprocessors and low-cost communication and actuation modules to beused in implementation of multi-agent control systems. They are responsible forcommunicating with neighbors, processing the information, computing control algorithm,and driving the agent. Normally such low-cost on board hardware would lead to limitedcommunication capacity and limited onboard energy resource. Unfortunately, most of theexisting results focus on control algorithm development in a continuous-time settingwithout considering these practical constraints. More recently, some efforts are devotedto addressing those constraints by adopting sampled-data algorithms. However, theyonly consider multi-agent systems with single integrator or double integrator agentdynamics, which are not the case for many engineering systems in practice.The objective of this project is to investigate several fundamental problems incoordinated control and implementation of multi-agent systems in more practicalsettings. We aim to develop a distributed non-uniformly sampled control methodologyfor coordination of multi-agent systems. This includes non-uniform sampling algorithmdesign, sampled controller synthesis, and stability analysis for multi-agent systems withmore general agent dynamics, time-varying communication topologies, parametervariation and environmental disturbances. The theoretical results will also be applied toseveral practical engineering problems such as formation of mobile robots andcoordination of micro air vehicles.The existing continuous and periodic sampling methods are inherently demanding forcommunication bandwidth and capacity. To aim for a wider range of applications withmore flexibility, a non-uniformly sampled control strategy will be developed. Thisapproach only requires information exchange and control execution when certaincondition is satisfied. With such a mechanism, the number of communications, controllercomputations, and actuator updates could be significantly reduced. The success of thisresearch project will enable us to address a number of critical and fundamentalproblems in coordinated control and implementation of multi-agent systems. Theanticipated theoretic results will facilitate the applications of multi-agent systems inmore practical situations. This project thus holds both significant academic interest andpractical importance.

Detail(s)

Project number9042049
Grant typeGRF
StatusFinished
Effective start/end date1/01/1521/12/18

    Research areas

  • control systems,distributed control,sampled control ,mobile robots ,