Research On Time Synchronization Of Wireless Sensor Networks

Time Synchronization is a necessary mechanism and key technology in wireless sensor networks, which is increasingly widespread concerned in recent years. Based on the researches on the subject of WSNs in background, necessity and significance in the first, this paper is developed. We put forward three algorithms on average time synchronization of WSNs, mainly analyzing the robustness, scalability, synchronization accuracy and reduction of the number of message exchanges, and maintenance energy efficient in use of the algorithm. In addition, we implement strict mathematical proof and convergence analysis to each algorithm, and the experimental simulation is also executed.In the first scheme ATSP algorithm, the phase offset, frequency drift and random time delay in the process of information transmission are all considered, and the contrast experiments with TPSN algorithm is also made. Under the analyses of mathematical theory and simulation experiment, time synchronization approach of ATSP algorithm is proved high-accuracy, yet at the cost of a large amount of message exchanges. In order to improve this shortage, we put forward the second average time synchronization algorithm.The second scheme CATS algorithm, inspired by RBS algorithm, makes a division of the intersected cluster structure model between each other about sensor network, and a local average time synchronization approach in a cluster is put forward. Similarly, it also considers the phase offset and frequency drift, and the contrast experiments with RBS algorithm is also made. A good energy conservation and high synchronization accuracy is demonstrated, but CATS algorithm may apply to an ideal condition, no considering the random time delay in the process of information transmission.The third scheme ACATS algorithm, using a hear-thought inspired by PBS algorithm, makes an improvement in the local average time synchronization method in a cluster. It reduces the number of message exchange needed among nodes in a cluster in the synchronous process, and it also considers the phase offset, frequency drift and random time delay in the process of information transmission, applying to the actual sensor networks.Project supported by the Natural Science Foundation of China (No.61072139).