Abstract

Energy being major constraint and data gathering the core operation, considerable attention has been given to energy efficient data gathering in wireless sensor network research and clustering has emerged as an efficient architecture to achieve it. Low energy adaptive clustering hierarchy (LEACH), proposed in past, guided significant portion of this research and is used as a benchmark in this area. However, use of probabilistic and randomized approach for cluster head selection results in their variable number, during different rounds of data gathering. Use of spread spectrum technique, to allow interference free parallel communication in different clusters, results in increased data size. The variable number of cluster heads requires the spreading factor to be set corresponding to probable maximum number of cluster heads, to avoid interference. This additional increase in data size increases amount of energy required to transmit it to the base station, directly affecting network lifetime and energy efficiency, both. To reduce the spreading factor and to improve data gathering efficiency a distributed clustering scheme is proposed in this paper, which restricts the number of cluster heads whenever more than the desired numbers of nodes select them to acquire the role. Results of the simulations, carried out in NS2, show that with the proposed scheme the amount of data received at base station per unit of energy consumed is significantly more than the benchmark scheme used. Proposed scheme is observed to provide improvement of 58% in network lifetime and 48.76% in energy efficiency over LEACH.