Clock synchronization for random mobile sensor networks

Mobile sensor nodes have found a large range of applications, e.g., social networks, habitant monitoring. Clock synchronization is a critical issue for most applications with mobile sensor networks. However, due to the limited communication range, the mobile sensor nodes can only exchange their information once they are sufficiently close for contact. Moreover, the random movement makes the performance analysis of any time synchronization protocols more challenging. In this paper, we introduce the relation graph to model the random contact of mobile sensor nodes, and evaluate the performance by the probability that the network can be synchronized within an arbitrary time. We adapt the previous maximum time synchronization protocol, which can drive the clocks of all sensor nodes to a common value by utilizing their own neighboring information. Furthermore, we provide an analytical lower bound for the probability that the time synchronization can be finished within any given time. For linearizable graph, we also provide an efficient way to calculate the exact probabilities. Extensive numerical examples show the effectiveness of our results.