An Implementation of Decentralized Consensus Building in Sensor Networks

In cognitive sensor networks, achieving consensus among the sensor nodes without requiring centralized control is an important attribute that can enable quick and reliable network decisions. Decentralized consensus building can be achieved through iterative information exchange among sensor nodes. While much of the literature has concentrated on developing theory, minimizing the number of iterations is necessary in practice to reduce energy consumption. In this work, we present an approach aimed at solving this practical problem. Specifically, the contributions of this work are fourfold. First, existing theoretical, continuous-time results are reformulated so that they can be implemented in physical, discrete-time hardware. Second, we show that the number of iterations needed to achieve consensus can be minimized a priori for a given network topology. Third, we illustrate our results through numerical analysis and network simulation. Finally, we present results from a hardware implementation.