An Improved SEED Clustering Model for Wireless Sensor Networks

It is important to develop clustering methods to collect data efficiently in wireless sensor networks (WSNs). Among the clustering methods in the literature, the most popular on behalf of balanced energy depletion and increasing the life of the network is heterogeneous clustering consisting of nodes with different characteristics. In this study, Sleep-awake Energy Efficient Distributed (SEED) clustering method that is a heterogeneous clustering, has been improved. In this sense, the mechanism of the SEED has been developed on behalf of the data sending-receiving, and energy consumption. According to the proposed method, the nodes in the WSN perceive the data in specified time periods and do not transmit and receive data by staying asleep at certain times. The most important difference of the proposed algorithm from the SEED method is that the remaining energy of the nodes and the network average energy are added to the threshold value in the cluster head (CH) selection. Moreover, cluster formation and CH selection enables more effective method than SEED algorithm by providing cluster members to communicate with CHs, and then the data transmission process is also included in the method process. Thus, energy consumption is reduced and network life is elongated by choosing the optimum CHs. The proposed method has been compared with both the SEED algorithm and other heterogeneous clustering methods existing in the literature in the simulation environment. The results of the simulations show the advantages of the recommended method.