Kalman-consensus filter for time synchronization in wireless sensor networks

Providing scalable and robust time synchronization among nodes is one of the challenges for designing and operating a large scale wireless sensor network. In this paper, a novel algorithm based on Kalman-consensus filter is proposed to achieve accurate time synchronization in multihop wireless sensor networks. The proposed approach does not need to select a reference node or construct a spanning tree in the network. Every node exchanges time messages with its neighbors and uses Kalman-consensus filter to estimate the time offset and skew. By this way, the virtual global time is computed and converges to a common value. Simulation results show that the proposed algorithm effectively reduces the global synchronization error. Moreover, the fully distributed algorithm is robust to dynamic network topology due, for example, to node failure and new node appearance.