Clustering algorithms for wireless ad hoc networks

Abstract

Efficient clustering algorithms play a very important role in the fast connection establishment of ad hoc networks. In this paper, we describe a communication model that is derived directly from that of Bluetooth, an emerging technology for pervasive computing; this technology is expected to play a major role in future personal area network applications. We further propose two new distributed algorithms for clustering in wireless ad hoc networks. The existing algorithms often become infeasible because they use models where the discovering devices broadcast their Ids and exchange substantial information in the initial stages of the algorithm. We propose a 2-stage distributed O(N) randomized algorithm for an N node complete network, that always finds the minimum number of star-shaped clusters, which have maximum size. We then present a completely deterministic O(N) distributed algorithm for the same model, which achieves the same purpose. We describe in detail how these algorithms can be applied to Bluetooth for efficient scatternet formation. Finally, we evaluate both algorithms using simulation experiments based on the Bluetooth communication model, and compare their performance.