Clock Skew Estimation for Timestamp-Free Synchronization in Industrial Wireless Sensor Networks

摘要

Existing packet-based synchronization approaches take extensive energy for industrial wireless sensor networks due to substantial timing message exchanges. Combined with acknowledgement (ACK), this article proposes an energy-efficient synchronization protocol without exchanging timestamps. Also, a mapping rule of ACK response time is presented for the joint estimation of skew and fixed delay without requiring a following response in each cycle, so that estimation can be implemented in parallel to data transmission. Then, the timestamp-free synchronization technique is extended to the receiver-only synchronization scenario to further reduce communication overhead. Assuming Gaussian variable delays, the maximum likelihood estimators and corresponding lower bounds for slave node and silent node are developed, as well as a low-complexity estimator for silent node. In addition, considering the dynamics of clock, a timestamp-free clock skew tracking algorithm using Kalman filtering is developed. Simulation results validate the efficiency of the proposed estimators.