Distributed Coordination Of Heterogeneous Multi-Agent Systems

Distributed coordination of multi-agent systems has received a major attention ofmulti-disciplinary researchers coming from applied mathematics, statistical physics, bi-ology, communication, system and control, computer science in the past decade. This ispartly due to its broad applications of multi-agent systems in many areas, such as theformation control of robotic systems, the cooperative control of unmanned aerial vehi-cles, the attitude alignment of satellite clusters, the target tracking of sensor networks,and so on. This dissertation considers two classes of distributed coordination problems–consensus and containment. The main results of this dissertation are as follows:1. The consensus of heterogeneous multi-agent systems with agents modeled byfrst-order and second-order integrator is studied. First, we study the consensus ofheterogeneous multi-agent systems under the connected undirected graph and leader-following network, and obtain some sufcient conditions for solving the consensus prob-lem by using the graph theory, Lyapunov method and Lasalle’s invariance principle.Then, using the same methods and under the same networks, we get some sufcientconditions for solving the consensus problem which the second-order integrator agentscannot obtain the velocity information. Finally, we study the the consensus of het-erogeneous multi-agent systems under directed fxed and switching topologies, and getsome necessary and sufcient conditions under directed fxed topology and some suf-cient conditions under directed switching topologies by using the variable substitution,algebraic graph theory and previous results for frst-order multi-agent systems.2. The fnite-time consensus of heterogeneous multi-agent systems is also studiedin this dissertation. The continuous protocols with and without velocity information aregiven for heterogeneous multi-agent systems. By adopting the graph theory, homoge-neous domination method and adding a power integrator method, we get some sufcientconditions for solving the fnite-time consensus problem under strongly connected graphand leader-following network.3. The containment control of heterogeneous multi-agent systems is consideredunder directed graph. When the leaders are frst-order integrator agents, we designthe linear protocol for heterogeneous multi-agent systems. By using graph theory andvariable substitution, we know that the second-order integrator agents converge to theconvex hull spanned by the frst-order integrator agents if and only if the directed graphcontains a directed spanning forest. When the leaders are second-order integrator agents,we design the nonlinear protocol for heterogeneous multi-agent systems and obtain the necessary and sufcient condition for solving the containment control in fnite time.4. Some classes of fnite-time consensus problems of homogeneous multi-agent sys-tems are also investigated. First, we consider the fnite-time consensus of stochasticmulti-agent systems in probability. By using algebraic graph theory, stochastic Lya-punov method and probability theory, some sufcient conditions for solving the fnite-time consensus in probability are obtained under diferent communication topologies.Second, we study the fnite-time consensus of second-order multi-agent systems withoutvelocity measurements and some sufcient conditions for solving the fnite-time consen-sus are gotten by adopting the graph theory, Lyapunov theory and Lasalle’s invarianceprinciple. Through system transformation, the fnite-time consensus of second-ordermulti-agent systems with input delays is also studied and some sufcient conditions areobtained in the end.