Distributed weighting strategies for improved convergence speed of first-order consensus

Publikationstyp
Conference Paper
Publikationsdatum
2017-07-03
Sprache
English
Author
Mirali, Furugh 
Werner, Herbert 
Institut
Regelungstechnik E-14 
TORE-URI
http://hdl.handle.net/11420/3550
Start Page
942
End Page
947
Article Number
7963074
Citation
Proceedings of the American Control Conference: 7963074, 942-947 (2017)
Contribution to Conference
American Control Conference, ACC 2017 
Publisher DOI
10.23919/ACC.2017.7963074
Scopus ID
2-s2.0-85027011701
Publisher
IEEE
This paper is concerned with consensus algorithms for multi-agent systems. Motivated by a recently proposed distributed algorithm that guarantees average consensus for undirected and time-invariant communication graphs using the so-called equal neighbour model, which assumes agents to weight the incoming information equally, we propose a generalisation of this weighting strategy and extend the average consensus algorithm to this generalised case. We propose four special weighting strategies with guaranteed convergence to average consensus and compare them with the equal neighbour model. Simulation studies show that proposed weighting strategies outperform the equal-weight strategy. For the special case of the star topology this is analytically shown.
DDC Class
600: Technik