A cluster-based organization with two phases to separate priority calculation and network activity operations is developed. In the election-state phase, clusters are organized by those nodes with sufficient residual energy level(REL) that are distributed as evenly as possible in the network. For each network round, each node decides its role(such as, cluster header(CH) or a normal node) and its priority level for activity operation. The priority level also affects the active/sleep scheduling in the node. In the steady-state phase, the positive priority levels of active nodes determine their activity order to avoid transmission collision, and to reduce redundant data transmission and sensing coverage. Since the data reporting to the sink is usually performed by a CH delivery chain, subsidiary CHs are selected to assist with data delivery and to share the loading of CHs. The formulae for computing priority levels are analyzed and verified by the examples with valid parameters. The experimental results show that the two phases efficiently consume node energy and achieve energy savings.
