Clock Synchronization in Wireless Sensor Networks Based on Bayesian Estimation

Clock synchronization is essential for the operation of upper layer applications in Wireless Sensor Networks. When the network hops needed for clock synchronization message transmission is large, synchronization error will accumulate and synchronization accuracy may be reduced significantly. Moreover, in the existing synchronization algorithms, large number of communication resources and node energy will be expended in sending and receiving time messages. To solve the problem, this paper proposes a Bayesian estimation-based time synchronization (BETS) algorithm which uses synchronization error compensation to reduce the amount of time message interaction in clock synchronization. The key idea of BETS is to calibrate the prior information of synchronization error with a small amount of field sampling time information, which will eliminate the impact of environment on clock synchronization accuracy. In addition, the gradient descent method is used to estimate the relative clock drift rate, which provides the reference for setting algorithm execution cycle and ensures clock synchronization during network operation time. In order to evaluate the theoretical lower bound of the performance of BETS, the Bayesian Cramér-Rao bound (BCRB) is derived. Both simulation and hardware experiments show that BETS algorithm makes full use of the prior information of synchronization error, hence fewer time messages are required in synchronization and the resource constraints of WSNs are satisfied.