Research On Underwater Acoustic Sensor Network Time Synchronization

The underwater sensor network plays a key role in marine military applications,submarine exploration,marine environmental monitoring,and national security.Clock synchronization is critical to a underwater sensor network and is tihe basis for applications such as node location,data fusion,power management,and transmission scheduling.The underwater acoustic sensor network has problems such as long propagation delay,limited transmission bandwidth,and node mobility.The existing synchronization algorithm of underwater sensor network focuses on the synchronization of the pairwise nodes,and there is short of research on the global synchronization.To solve the problems above,this paper proposes a clock synchronization algorithm:K_Sync,for pairwise nodes in a underwater sensor network,and a global clock mutual synchronization algorithm:CO-Sync.The K-Sync algorithm uses the time correlation of the relative speed and relative distance of the sensor nodes to establish a linear model of the system.The Kalman filter is used to estimate the normalized cdock frequency and cdock offset to synchronize the clock between the two nodes.Compared with the existing methods under different scenarios,the simulation results show that the K-Sync algorithm has good accuracy in the normalized clock frequency,cumulative clock bias and other key parameters.In particular,the K-Sync algorithm can achieve high synchronization accuracy,under the condition of small synchronization interval,short response time,when node moves with the empirical motion model.Aiming at the problem of global cdock synchronization in sensor networks,a CO-Sync algorithm is proposed and the validity of the algorithm is proved.The algorithm does not require a global reference node.The node acquires clock information relative to its neighboring nodes to compensate for its own clock parameters.After multiple iterations,the node achieves global clock synchronization.In the case of chain,ring,grid,and random topology,and the case of different communication distances under random topology,the simulation results show that the algorithm can converge under different network topologies and the stronger the network is connected,the faster the convergence rate.