Time Synchronization Algorithms in Low-Cost Wireless Sensor Network Systems

Time synchronization is an essential building block of sensor network systems. In this paper, we present a comprehensive study of least squares algorithms for time synchronization in sensor networks, with a focus on continuous monitoring and data collection applications. We propose a set of algorithms to address a number of issues in practical implementation on typical low-cost sensor network platforms, including a scaled signal model to achieve numerical stability in an ill-conditioned problem, sequential estimators for the scaled signal model to reduce computational complexity, a fast initialization scheme to improve energy efficiency, and outlier detection algorithms to improve robustness in long-term autonomous operations. The proposed algorithms are implemented and a measurement-based simulation method is employed to study the performance.