Research On Cluster-based Time Synchronization Algorithm For Multi-hop In Wireless Sensor Networks

Tremendous advances in communication, embedded computation and sensor technology have occurred in the development of wireless sensor networks. Wireless sensor networks (WSN)shows large potential in a wide range of applications such as : military; environmental monitoring; medical sanitation; smart spaces; It was the promising technique in the 21st century.Time synchronization which plays an important role in the actual application is a fundamental service of WSN. The research of time synchronization for single-hop is mature, but the multi-hop time synchronization is relatively weak which could not satisfy the actual application requirement of WSN. So we proposed a new time synchronization algorithm combing with clusters. It aimed at the application of a large scale, high-density WSN.First, the paper introduced the fundamental principles of time synchronization, evaluation criterion of the performance and some importance in design. Then, analyzed and compared several typical exiting algorithms of time synchronization. Based on these, we presented a Cluster-based Time Synchronization algorithm for Multi-hop in Wireless Sensor Networks (CBTS). It starts from the Leach algorithm which distributes the network into different clusters. And each cluster is composed of a cluster head and some cluster members. Then, we begin to synchronize the nodes based on clustering. The algorithm works in two phases: horizontal time synchronization phase and vertical time synchronization phase. In the horizontal synchronization phase, it adopts traditional pair-wise packet exchange mechanism to finish the time synchronization between the Base Station and cluster heads through establishing a hierarchical topology structure. In the vertical synchronization phase, it adopts pair-wise packet exchange and unidirectional reference broadcast mechanism to finish the time synchronization between cluster heads and cluster members. Moreover, it used linear least square to estimate the clock offset of nodes which achieved a good precision. Eventually, the local clocks of all nodes in the network can synchronize to each other.Finally, we simulated the CBTS algorithm on the NS2.The results showed that CBTS has a good synchronization precision and low message cost. So the CBTS algorithm is more suitable for large scale WSN.