Dynamic algorithm for non-uniform deployment based on route graph in mobile sensor networks

Abstract

The paper presents an algorithm for optimal deployment in mobile wireless sensor networks in terms of coverage area and the number of active nodes. This technique is based on movements of route tree branches and can be applied to aggregate networks with one sink. For predict energy exhaust moment, the structure of route tree is used. Based on this estimation movement vectors for active nodes are calculated and nodes are moved. Because of computational simplicity, the algorithm may be implement decentralized in nodes microprocessors. For comparison of proposed method with existing deployment algorithms (ST, EMAND) a number of networks simulations has been done. The initial number of network nodes was chosen as independent variable parameter in the simulation. Based on simulation results performance estimation has been evaluated. The proposed algorithm demonstrates significant advantage in terms of coverage area (1.2-1.6 times over EMAND and ST) and average number of active nodes (1.6-2.8 times).