A Scalable Protocol for Cooperative Time Synchronization Using Spatial Averaging

Time synchronization is an important aspect of sensor network operation. However, it is well known that synchronization error accumulates over multiple hops. This presents a challenge for large-scale, multi-hop sensor networks with a large number of nodes distributed over wide areas. In this work, we present a protocol that uses spatial averaging to reduce error accumulation in large-scale networks. We provide an analysis to quantify the synchronization improvement achieved using spatial averaging and find that in a basic cooperative network, the skew and offset variance decrease approximately as $1/\bar{N}$ where $\bar{N}$ is the number of cooperating nodes. For general networks, simulation results and a comparison to basic cooperative network results are used to illustrate the improvement in synchronization performance.