Abstract :
Wireless sensor networks consisting of a mass of inexpensive, low-power and multifunctional sensor nodes are deployed in military or citizen field, collecting information or monitoring environments. Gathering data in energy efficient way is critically important to prolong the lifetimes of the networks. In {3], data collection problem is studied in homogeneous network. An adaptive clustering protocol LEACH is presented and makes a breakthrough in this area. By localized data fusion and randomized cluster heads rotation, energy is averagely depleted among nodes and LEACH achieves 7-8 times improvement than three traditional protocols measuring the dieing times of nodes; In this paper, we propose EDAC (Energy-Driven Adaptive Clustering) protocol, which is an improvement over LEACH in heterogeneous networks. Unlike the homogeneous network, if we apply LEACH in heterogeneous networks, the averagely energy dissipation mechanism will result in early death of "powerless" nodes and can not use the advantage of the "powerful" nodes. In EDAC, we use energyariven cluster heads rotation, which enables cluster heads change asynchronously and coordinate energy consuming closely based on the heterogeneous property of nodes?? energy. Time synchronization is relaxed and topology maintenance is localized by event triggering rotation. Also, the energy cost in topology maintenance can be kept small even when only little data to be transmitted in networks. Simulation results show that the lifetimes of nodes in EDAC can be 1-2.5 times as those in LEACH in heterogeneous network, and that it also works as well as LEACH in homogeneous networks.
