oa Distributed algorithms in wireless sensor networks: An approach for applying binary consensus in large testbeds

Our work represents a new starting point for a wireless sensor network implementation of a cooperative algorithm called the binary consensus algorithm. Binary consensus is used to allow a collection of distributed entities to reach consensus regarding the answer to a binary question and the final decision is based on the majority opinion. Binary consensus can play a basic role in increasing the accuracy of detecting event occurrence. Existing work on binary consensus focuses on simulation of the algorithm in a purely theoretical sense. We have adapted the binary consensus algorithm for use in wireless sensor networks. This is achieved by specifying how motes find partners to update state with as well as by adding a heuristic for individual motes to determine convergence. In traditional binary consensus, individual nodes do not have a stop condition, meaning nodes continue to transmit even after convergence has occurred. In WSNs however, this is unacceptable since it will consume power. So in order to save power sensor motes should stop the communication when the whole network converges. For that reason we have designed a tunable heuristic value N that will allow motes to estimate when convergence has occurred. We have evaluated our algorithm successfully in hardware using 139 IRIS sensor motes and have further supported our results using the TOSSIM simulator. We were able to minimize the convergence time reaching optimal results. The results also show that as the network gets denser the convergence time will lower. In addition, convergence speed depends on the number of motes present in the network. During the experiments none of the motes failed and our algorithm converged correctly. The hardware as well as the simulation results demonstrate that the convergence speed depends on the topology type, the number of motes present in the network, and the distribution of the initial states.