An Energy Efficient Clustering Algrithm Based on DEEC Protocol and K-mean Method

Abstract

In WSN the sensor nodes are usually powered by batteries and thus have very limited lifetime if no power management is performed. Because of this major limitation, energy-efficient techniques are main research challenge in this context to solve it. Dividing the network in clusters is an effective technique to achieve this goal. This algorithm is based on creating virtual sub-groups of sensor nodes in order to minimize routing calculations and to reduce the size of cluster head data aggregation. Nowadays, a lot of heterogeneous clustering protocols for WSN are created. Nevertheless, these protocols need to find the optimal clusters formation in the network that conserve CHs and theirs member nodes energy consumption. A new approach is proposed combining between an efficient clustering algorithm K-means and our proposed DEEC.  This approach has been employed to enhance DEEC protocol performances. Numerical simulation proves that the improved protocol entitled KM-DEEC achieves a satisfactory results compared to others DEEC protocol versions.