Stochastic sensor scheduling in Wireless Sensor Networks with general graph topology

In this paper we present a stochastic strategy for energy-efficient joint sensors selection/information routing in Wireless Sensor Networks with general communication graph topology. Such a scheme generalizes the early stochastic Sensor scheduling approach proposed in [1], [2], where the communication graph was limited to be a tree. The idea is to associate at each edge of the (directed) communication graph a probability to transmit and thus to perform the sensor selection and the information routing through a stochastic mechanism. Being the problem of optimally determining the edge probabilities computationally intractable, a heuristic based on the use of spanning trees is introduced together with some theoretical results that allow evaluation of the optimality gap. A final numerical example shows the effectiveness of the proposed method.