Novel Control Theoretic Consensus-Based Time Synchronization Algorithm for WSN in Industrial Applications: Convergence Analysis and Performance Characterization

Authors:
Publisher website
Google Scholar
Abstract
Fourth Industrial Revolution (IR4.0) depends heavily on Internet of Things (IoT) and wireless sensor networks (WSN). Through WSN, data in an industrial environment can be captured. Time synchronization is an important issue in WSN to ensure the correct sequence of the data collected. Besides, gauging on a common perceived time reference, local clocks on each node within any interconnected WSN should be able to initiate messages exchanges. Referencing to an unsynchronized value of virtual global reference clock will render the nodes communication and applications in WSN useless. In this work, a newly improved control theoretic consensus-based time synchronization algorithm for WSN known as Time synchronization using Sliding Mode control and PID (TSMPID) is proposed. The novel algorithm uses an augmented sliding mode control (SMC) in the skew and relative skew estimation schemes, which is a useful switching action, to provide a salient feature in guaranteeing finite-time convergence in synchronized virtual time clock estimation. A proportional, integral and derivative (PID) term has been introduced in the offset estimation to improve the estimation performance. The main purpose is to achieve global clock synchronization with lower computational effort and reduced synchronization error. TSMPID has the characteristics of being totally distributed, asynchronous, scalable across different network topological structures and robust to ad-hoc nodes deployment. Ad-hoc communication link node deployment scenarios are simulated comprising 5, 7 and 9 WSN nodes. TSMPID requires low energy and exhibits a much lesser consensus error spike in comparison to other recent prior-arts time synchronization schemes.
Funding Information
  • USM Research University Grant (1001/PELECT/8014088)
  • USM Industry Matching Grant (1001/PELECT/8070023)

This publication has 47 references indexed in Scilit: