Geometrical sensor selection in large-scale high-density sensor networks

In this paper, we consider large-scale high-density sensor networks consisting of small battery-powered sensor nodes. As these sensors are heavily limited in terms of energy consumption and thus the lifetime of the entire network is restricted, it is reasonable to introduce a sensor power as well as a total network power constraint. Both power constraints can simultaneously hold by means of smart power allocation methods. For a large number of sensor nodes the complexity of the utilized selection algorithm can become intolerably high. In order to simplify the power allocation procedure as well as a consecutive selection of most reliable sensor nodes, we propose an analytic-geometric pre-selection of 3-dimensional subspaces, in which more reliable sensor nodes are located. Our investigation is based on the distance between uniformly distributed sensor nodes, the target object and the fusion center as well as a free space signal propagation model. We present analytical solutions for the selection procedure and derive simplified equations in order to directly determine the region of active sensor nodes in closed-form.