Power Consumption Characteristics of Autonomous Decentralized Clustering Based on Local Interaction


IEICE TRANSACTIONS on Information and Systems   Vol.E97-D   No.12   pp.2984-2994
Publication Date: 2014/12/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2014PAP0001
Type of Manuscript: Special Section PAPER (Special Section on Parallel and Distributed Computing and Networking)
Category: Wireless Network
autonomous decentralized control,  local-action theory,  mobile ad hoc network,  clustering,  

Full Text: PDF>>
Buy this Article

Mobile ad hoc networks (MANETs) consist of mobile terminals that directly connect with one another to communicate without a network infrastructure, such as base stations and/or access points of wireless local area networks (LANs) connected to wired backbone networks. Large-scale disasters such as tsunamis and earthquakes can cause serious damage to life, property as well as any network infrastructure. However, MANETs can function even after severe disasters have destroyed regular network infrastructure. We have proposed an autonomous decentralized structure formation technology based on local interaction, and have applied it to implement autonomous decentralized clustering on MANETs. This method is known to configure clusters that reflect the network condition, such as residual battery power and the degree of each node. However, the effect of clusters that reflect the network condition has not been evaluated. In this study, we configure clusters using our method, the back-diffusion method, and a bio-inspired method, which is a kind of autonomous decentralized clustering that cannot reflect the network condition. We also clarify the importance of clustering that reflects the network condition, with regard to power consumption and data transfer efficiency.