A scalable fault tolerant diffusion scheme for data fusion in sensor networks

In sensor networks, sensor nodes are usually deployed randomly over an area to collect the information of interest. Data fusion is the phase of processing the collected information by sensor nodes before they are sent to the end user. This paper introduces a distributed consensus algorithm that allows the nodes of a sensor network to track the average of n sensor measurements. The study of the above mentioned algorithm showed that it is robust to asynchronism and dynamic topology changes. It is also put in evidence that the algorithm is fully distributed and does not require any global coordination. Moreover, The proposed method doesn’t involve explicit point-to-point message or routing, it diffuses information across the network. Accordingly, simulation results are provided illustrating the effectiveness of the studied algorithm.