A Convergence-Accelerated Distributed Time Synchronization Algorithm for Energy-Harvesting Wireless Sensor Networks

Time synchronization is an essential problem for energy-harvesting wireless sensor networks (EH-WSNs), which is closely related to efficient resource schedules, energy harvesting, data fusion, location, etc. With the advantage of being more robust than master controlling synchronization, distributed time synchronization algorithms are usually used to EH-WSNs for cooperating sleeping nodes. This paper proposes a novel accelerated time co-synchronization algorithm based on the storage-and-prediction method to improve the convergence rate. In this algorithm, each node in the network first predicts the estimated current time state value according to previous time state values stored in the local node, and then adjusts the time state value according to the estimated time state value deviations between all its adjacent nodes. Theoretical analysis in a more general case shows that the proposed algorithm can improve the convergence rate of distributed time synchronization when selecting the appropriate parameter, and the closed-form solution of the optimal parameter is also given. Finally, the simulation of comparing the classical algorithm with the proposed algorithm based on different scenarios is completed.