Open AccessTrust‐based self‐organising network control
Author(s) -
Haus Tomislav,
Palunko Ivana,
Tolić Domagoj,
Bogdan Stjepan,
Lewis Frank L.,
Mikulski Dariusz G.
Publication year - 2014
Publication title -
iet control theory and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.059
H-Index - 108
eISSN - 1751-8652
pISSN - 1751-8644
DOI - 10.1049/iet-cta.2014.0333
Subject(s) - computer science , mechanism (biology) , reputation , negotiation , multi agent system , convergence (economics) , control (management) , trajectory , decentralised system , network topology , autonomous agent , task (project management) , order (exchange) , compromise , distributed computing , artificial intelligence , computer network , engineering , business , social science , philosophy , physics , systems engineering , epistemology , finance , astronomy , sociology , political science , law , economics , economic growth
This study proposes a mechanism for decentralised trust‐based self‐organising network control. The driving force of this work is the idea that some agents are more versed in certain tasks, such as communication, data processing, localisation or trajectory tracking, at a given moment. The author's control objective is to reach an agreement throughout the network regarding each agent's suitability for a certain task. In particular, this study is interested in cultivating network trust towards each agent. Trust values reflect the reputation of each agent within the group. The authors initialise individual trust values towards each agent based on local observations and observations received from the neighbours. Next, a trust‐dependent consensus‐based mechanism is employed. Since this mechanism is decentralised, scenarios with unequal trusts values towards one agent need to be handled with caution in order not to compromise the collective behaviour. To that end, they employ adaptive control concepts when negotiating trust values. They formally prove convergence of the proposed adaptive mechanism for the network topologies that can be represented with directed weighted graphs. The proposed mechanism is illustrated through a case study of decentralised formation control of autonomous agents. They provide an in‐depth analysis of several possible scenarios regarding the proposed control structure.