Distributed Tracking Control of Networked Heterogeneous Dynamic Systems by Pure Output Feedback

Project: Research

View graph of relations

Description

Systems of networked robotic vehicles have been widely applied for various tasks in recent years. Examples include large area surveillance, emergency building survey, and ocean trash collection. Such systems belong to the so-called multi-agent systems (MASs), and each robotic vehicle in the system is usually referred to as an agent. To achieve coordinated maneuvering, the cooperative control of MASs has been extensively studied in the robotics and control community since the early 2000s. Numerous researchers, including the PI, have recently been working on the cooperative control of heterogeneous MASs with time-varying network topologies. This research has been motivated by real-world missions where agents often have different dynamics or are subject to uncertainties, with communication network topologies among agents being time-varying due to environment change or vehicle failure. It has been proved that a reference internal model is necessary and sufficient for the output consensus/tracking of heterogeneous linear MASs, which is the most fundamental problem concerning this topic. Furthermore, most approaches to this problem often require that certain state information of the local controllers can be exchanged through a communication network. However, this is not feasible in many practical scenarios for the following reasons: 1) communication resources are too limited to exchange such large amounts of information; 2) communication between agents is not available or reliable due to environmental constraints; or 3) agents are not equipped with communication units. To address this issue, it is desirable to develop distributed control strategies based on pure output feedback in the sense that: 1) only the output information of individual agents is used by the controller, and 2) only the output information of neighboring agents is exchanged through the communication network. In this project, we aim to investigate tracking control problems of heterogeneous MASs with time-varying network topologies based on pure output feedback. In particular, several types of heterogeneous uncertain MASs will be studied, including general linear, second-order nonlinear, and a typical class of higher order nonlinear MASs. New stabilization techniques will be proposed. The developed control approaches will be validated via their application to tracking control of a group of mobile robots. The anticipated results of this research project will not only provide an alternative way to solve fundamental problems in robotic vehicles networks but also facilitate applications of coordinated maneuvering in real mission scenarios. This project is thus of both significant academic interest and practical importance. 

Detail(s)

Project number9042784
Grant typeGRF
StatusActive
Effective start/end date1/01/20 → …