A Distributed Convergent Clock Synchronization Algorithm Based on Exponential Smoothing for Cloud-Based Railway Interlocking System

Clock synchronization is one of the essential problems for the cloud-based railway interlocking system, which is related to task processing and information synchronization. To improve the convergence rate and performance of a classical distributed clock synchronization algorithm, this paper proposes an exponential smoothing predictor-based distributed clock synchronization algorithm for the cloud-based railway interlocking system. Firstly, a classical distributed clock synchronization algorithm is conducted to get a convergent clock for distributed computing nodes. Then, to improve the convergence rate and performance, we introduce an exponential smoothing predictor to the classical method. Moreover, the convergence condition and convergence rate of our proposed method were analyzed theoretically. Finally, the simulation results show that the proposed algorithm has a better convergence performance than a classical method and can meet the requirements of a cloud-based railway interlocking system.