Cluster-based multi-channel system for improving performance of large-scale wireless multi-sink sensor networks

This paper presents a simple and distributed clustering algorithm suitable for large-scale wireless sensor networks (WSNs) consisting of several thousands of sensor nodes and few sink nodes. A two-tiered hierarchical architecture is used to increase scalability and ensure performances and durability of such a system: Level 1 called sensor network is partitioned into several equilibrate clusters with one leader or sink by cluster; Level 2 also called sink network is composed by N sinks nodes placed in planned manner into monitored region and is connected through IEEE 802.11 radio interfaces. A multi-channel system is used to create a cellular structure by assigning one frequency channel per cluster. We use simulation technique to evaluate and compare the impact of two distributed schemes (a single channel one and a multi-channel one) on network capacity like traffic load, energy consumption, medium access delay, end-to-end delay and data delivery ratio.