BN-LEACH: An improvement on LEACH protocol using Bayesian networks for energy consumption reduction in wireless sensor networks

A Wireless Sensor Network (WSN) consists of thousands number of sensor nodes which have constraints on computation, energy and memory resources. Limited energy resource affects the lifetime of the WSNs. Hence many protocols have been proposed to reduce energy consumption. LEACH is one of the most popular clustering protocols with the aim of maintaining the energy efficiency of sensor nodes. Almost all improvements of LEACH protocol give no guarantee that cluster-heads are uniformly distributed in the network. In this paper, we propose a protocol named BN-LEACH to select cluster-heads using a Bayesian Network (BN) model based on three factors — distance to the Base Station (BS), remaining energy and density. This model calculates the probability of becoming cluster-head for each sensor node. Appropriate cluster-heads are chosen according to a dynamic zoning method and a greedy approach in order to distribute cluster-heads uniformly. By comparison with LEACH, LEACH-C and WEEC protocols, which are improvements on LEACH, simulation results show that our proposed protocol balances the energy consumption of the sensor nodes, prolongs the network lifetime and extends the First Node Death (FND) further more from above-mentioned protocols.