A power efficient cluster head selection algorithm based on Affinity Propagation in heterogeneous sensor networks

Wireless sensor networks (WSNs) are composed of a large number of sensor nodes with limited computation capability, power and memory. Balance of energy consumption between nodes can reduce the number of dead nodes and prolong the network lifetime. In this paper, based on the novel clustering algorithm Affinity Propagation (AP), a power efficient cluster head selection algorithm (PECBA) is proposed to select cluster heads considering the energy balance for multi-level heterogeneous WSNs. In our proposed algorithm, the optimal cluster heads are first chosen according to the distance between any two nodes and the ratio of each node´s residual energy to network´s average energy, and then the non-cluster head nodes are assigned to different cluster heads according to the energy balancing criteria proposed in this paper. Simulation results have shown that the proposed algorithm has better performance in both the balance of energy consumption and the network lifetime compared with DEEC. Besides, by comparing our proposed PECBA and the common AP, the energy balancing criterion used in PECBA makes our proposed algorithm more power efficient.