A Study On The Time Synchronization In Wireless Sensor Network

With the rapid development of Micro-electro-Mechanism Systems (MEMS), Embedded System on Chip and wireless communication technologies, the wireless sensor network (WSN) which integrates above these key technologies become the hot research spotTime synchronization is a key supporting technology in wireless sensor network. In the practical applications of WSN, only when the data collected by the sensor node is integrated with time information does it have practical significance. In wireless sensor network, node localization, data fusion, TDMA scheduling, coordination of sleep and the other applications require sensor nodes based on precise time synchronization. Research on the time synchronization technology has great significance for promoting the extensive application of wireless sensor network.Firstly, the paper makes the specific research on time synchronization technology in wireless sensor network, describes the time model, and summarizes the key factors affecting the time synchronization and the design ideas of the time synchronization algorithm. The time synchronization algorithms in wireless sensor network which have been proposed at home and abroad are also classified and compared.Secondly, a time synchronization algorithm based on maximum likelihood estimation for wireless sensor networks called M-BTS is proposed. The algorithm improves the accuracy of the BTS by employing the maximum likelihood estimation method to estimate the maximum possible value of the time offset. It is proved in the theory and experiments that the maximum possible value is synodical estimation. Our experiment shows that this algorithm has higher precision and stability than BTS.Finally,since most of current research on time synchronization in WSN at home is limited to the single-hop time synchronization, or only improves time synchronization precision and energy consumption of the classic algorithm, which cannot meet the need of the large-scale sensor network applications, a compensated time synchronization algorithm with multiple reference nodes is proposed which is aiming at the large-scale multi-hop network environment. The urgent solving problem of the large-scale multi-hop network is the progressional growth of accumulation errors along with the increase of the number of the hop. So we propose the algorithm uses the special position of the boundary nodes to collect the time information from different reference nodes and compare them. The boundary nodes calculate and return the different value to the reference node which includes lager error. After some synchronization the time offset can be decreased among the different reference nodes. So the time synchronization of the whole network is improved. The result of simulation shows that this algorithm can greatly improve the accuracy of wireless sensor network time synchronization and has certain feasibility.