Time Synchronization in WSNs With Random Bounded Communication Delays

Recently, it was pointed out that many popular consensus-based time synchronization algorithms, such as Average TimeSynch (ATS), are all divergent in a network with arbitrarily small random communication delays under any interconnection topology. This note proposes a least-squares estimation-based time synchronization (LSTS for short) algorithm. This algorithm modifies the ATS algorithm in relative drift estimator, low-pass filter, and consensus protocol. By analyzing the mathematic model of the algorithm, it is proved that LSTS ensures asymptotic convergence of the drift estimation and boundedness of the global time synchronization error under random bounded communication delays if the network topology satisfies some connectedness conditions.