Premium
Synchronization of heterogeneous multiagent systems: A distributed observer and compensation framework
Author(s) -
Wang Bohui,
Zhao Yu,
Zhang Langwen
Publication year - 2020
Publication title -
advanced control for applications: engineering and industrial systems
Language(s) - English
Resource type - Journals
ISSN - 2578-0727
DOI - 10.1002/adc2.43
Subject(s) - synchronization (alternating current) , computer science , control theory (sociology) , observer (physics) , distributed computing , bounded function , controller (irrigation) , multi agent system , compensation (psychology) , synchronization networks , dual (grammatical number) , topology (electrical circuits) , control (management) , mathematics , artificial intelligence , computer network , psychology , mathematical analysis , channel (broadcasting) , art , physics , literature , quantum mechanics , combinatorics , psychoanalysis , agronomy , biology
This article proposes a distributed observer and compensation framework to address the synchronization problem of heterogeneous multiagent systems under directed communication and dynamic jointly connected topology scenarios. Different from the previous results, the synchronization problem is solved in our framework which neither relies on the solution of regulator equations which can be regarded as a high gain and usually calculated by a forced equivalence principle, nor assumes the global connectivity condition for all the agents to obtain a bounded synchronization error level. To do this, first, we introduce a dual design principle, where a stabilization condition with an appropriate coupling strength for the synchronization algorithm is proposed. Within this framework, a class of synchronization dynamics called synchronization distributed observers and tracking controller named control protocol are designed. While the synchronization behaviors can be achieved between synchronization distributed observers and control protocol, all the agents will synchronize to their leader in a situation where the system dynamics of the leader can be acquired in advance by all individuals. Finally, simulations of heterogeneous single‐link manipulators are presented to show the effectiveness of our approach.