此方法主要使用剩餘能源與時間優先權的觀念，利用各節點的剩餘能源比來計算出各節點的群組等待時間，而群組等待時間最短的節點，將獲得鄰近地區最高的優先權，而成為此一地區的群組核心，進而收集群組成員（cluster member）。如此的群組核心選取方式，可以在良好的同步機制運作下，就選取到能源狀況比較好的群組核心，進而節省訊息交換的封包傳輸能源消耗。 A Wireless Sensor Network consists of hundreds or thousands of sensor nodes. Sensor nodes are battery-powered and usually deployed in an unprotected environment to collect the needed information. Because of the enormous number of sensor nodes, it is difficult to recharge sensor nodes after distribution. So, how to increase energy efficiency and prolong the network lifetime is the design focus.
Clustering is the easier and more efficient way to save energy in wireless sensor networks. Having cluster heads process the data collected by cluster members have the following energy efficient factors: 1.Considering the data collected by large amount of sensor nodes, cluster heads help the aggregation of data and reduce the number of data transmissions. 2.Clustering reduces the number of nodes that need to relay data, therefore reduces the channel-competing situation. 3.Periodical reselection of clusters/cluster heads can avoid the heavy loads on particular nodes, letting all nodes take turns balances loading and extends the entire network lifetime.
The importance of cluster heads is stated above. Besides data aggregation and transmission, the method of selecting cluster heads also determines the size of a cluster, the position of a cluster head and the number of cluster heads in a network, thus affecting the energy consumption of a wireless sensor network. Implementing an efficient cluster head selection scheme is one of the key points in clustering.
This paper presents a new clustering approach. It uses a new cluster-head selection method to reach the goals of low clustering information overhead and optimal cluster-head selection.
It uses node residual energy to compute the clustering waiting time of each node. The node, which has the shortest waiting time, will have the highest priority to become a cluster head. This method can choose nodes with more energy as cluster heads and further decreases energy consumption for clustering overhead.
