A precision adaptive average time synchronization protocol in wireless sensor networks

Time synchronization provides an essential service for wireless sensor networks. Most of previous works focus on improving synchronization accuracy. However, other design concerns, such as global time continuity, power efficiency and adaptability, arise from real applications´ requirements and receive more and more attentions. During synchronization, most of approaches choose a specific sensor node´s (denoted as leader node) local time to be the global time, which would be disturbed easily due to network dynamics (leader node´s re-elections or power down). The disorder global time reference could confuse other sensor nodes and even crash applications based on temporal sequence relationship. This paper first present an average based time synchronization algorithm, which can provide continuous global time in a cluster with similar synchronization precision during leader re-elections. Furthermore, in order to meet power efficiency and adaptability requirements from applications, a self measure and query based adaptive mechanism is integrated into it to adjust synchronization period dynamically. Compared with previous adaptive algorithms, the query table method provides one-step convergence while satisfying synchronization error tolerance.