Feedback-Based Clock Synchronization in Wireless Sensor Networks: A Control Theoretic Approach

Abstract

Most wireless sensor networks (WSNs) employ a sleep clock to enable sensor on-off mode to save energy. Since the sleep clock usually works at a relatively low frequency, it is important to correct the long-term synchronization error caused by instability and nonlinearity. In this paper, the time synchronization issue in a WSN is formulated as a closed-loop control problem. Using the proportional-integral (PI) control principle, we propose a feedback-based synchronization (FBS) scheme to compensate the clock drift caused by both internal perturbation and external disturbance. Synchronization accuracy and FBS dynamics are analyzed in terms of response time and overshoot. We also derive a formula to determine the controller parameters for different synchronization accuracy and response-time requirements. Extensive experiments have been conducted to evaluate this synchronization scheme. The results indicate that FBS is much more robust than the delay measurement time-synchronization (DMTS) protocol in different synchronization durations and sensor node configurations. It is shown that FBS achieves almost the same accuracy as the flooding time-synchronization protocol (FTSP) when synchronization period P is less than 60 s, and it outperforms FTSP when P is longer than 60 s. Furthermore, the experimental results reveal that FBS's response to an external disturbance is faster than that of FTSP in one- or two-hop WSN scenarios. In addition, FBS consumes fewer resources than FTSP.