LAUSR

Clock synchronization is a significant basis for many operations in wireless sensor networks. The distributed consensus-based clock synchronization has gained popularity for its robustness and scalability. However, most consensus-based clock synchronization protocols either ignore communication delay, or consider bounded delay without distribution model. In this article, based on the consensus theory, the clock synchronization problem with truncated exponential delay is investigated. We use the maximum likelihood method to estimate relative drift between nodes. A recursive method is presented to cope with the maximum likelihood estimation, which largely reduces the computational complexity and the storage overhead. Then, the estimated relative drift is adopted in the two main parts of consensus clock synchronization: drift compensation and offset compensation. We give the rigorous theoretical proof of the convergence of the network-wide synchronization. Simulations further verify the theoretical analysis and show that the synchronization performance of the proposed algorithm is better than existing similar schemes under truncated exponential delays.