TY - GEN
T1 - The recursive time synchronization protocol for wireless sensor networks
AU - Akhlaq, Muhammad
AU - Sheltami, Tarek R.
PY - 2012
Y1 - 2012
N2 - Many applications of Wireless Sensor Networks (WSNs) require accurate time synchronization for data consistency and coordination. While the existing protocols for time synchronization provide sufficient accuracy, they consume high energy and poorly synchronize the distant nodes. We propose a Recursive Time Synchronization Protocol (RTSP) which provides global clock synchronization in an accurate and energy-efficient way. It achieves better performance by using a novel blend of techniques such as the MAC-layer time-stamping based on Start of Frame Delimiter (SFD) byte, fewer broadcasts by the reference node, compensation of the propagation delay and adjustment of the timestamps at each hop, estimation of the relative skew and offset using least square linear regression on two data points (2LR), adaptive re-synchronization interval, aggregation of the synchronization requests and energy-awareness. We also analyze the sources of errors and efficiency. Simulation results show an average accuracy of 0.3μs in a large multi-hop network while using only 1/5th of the energy consumed by FTSP in the long run, i.e., the RTSP outperforms all other protocols of its class including RBS, TPSN and FTSP.
AB - Many applications of Wireless Sensor Networks (WSNs) require accurate time synchronization for data consistency and coordination. While the existing protocols for time synchronization provide sufficient accuracy, they consume high energy and poorly synchronize the distant nodes. We propose a Recursive Time Synchronization Protocol (RTSP) which provides global clock synchronization in an accurate and energy-efficient way. It achieves better performance by using a novel blend of techniques such as the MAC-layer time-stamping based on Start of Frame Delimiter (SFD) byte, fewer broadcasts by the reference node, compensation of the propagation delay and adjustment of the timestamps at each hop, estimation of the relative skew and offset using least square linear regression on two data points (2LR), adaptive re-synchronization interval, aggregation of the synchronization requests and energy-awareness. We also analyze the sources of errors and efficiency. Simulation results show an average accuracy of 0.3μs in a large multi-hop network while using only 1/5th of the energy consumed by FTSP in the long run, i.e., the RTSP outperforms all other protocols of its class including RBS, TPSN and FTSP.
KW - Adaptive Re-synchronization Interval
KW - Drift Estimation
KW - Election of the Reference Node
KW - Energy Awareness and Efficiency
KW - Global Clock Synchronization
KW - Request Aggregation
UR - https://www.scopus.com/pages/publications/84859036703
U2 - 10.1109/SAS.2012.6166318
DO - 10.1109/SAS.2012.6166318
M3 - Conference contribution
AN - SCOPUS:84859036703
SN - 9781457717239
T3 - 2012 IEEE Sensors Applications Symposium, SAS 2012 - Proceedings
SP - 62
EP - 67
BT - 2012 IEEE Sensors Applications Symposium, SAS 2012 - Proceedings
ER -