Distributed network time synchronization: Social learning versus consensus
Abstract
The objective of this thesis is to investigate social learning-based distributed
network time synchronization (SLDNTS) and compare it to a classic approach: consensus
DNTS (CDNTS). To achieve this objective, the thesis introduces a method for generating a
practical observation random variable (ORV) for SLDNTS and presents both the worst and
best ORV. Then, this thesis shows, through simulations, that SLDNTS is more robust than
CDNTS in convergence. CDNTS fails when timing measurement error is nonzero, i.e.,a
Gaussian random variable with mean equal to true time and variance equal to one is
applied as an observation random variable (ORV) at each node. On the other hand
SLDNTS shows quick convergence with small number of iterations even if nonzero timing
measurement error, i.e., a Gaussian random variable, is applied.
Description
Thesis (M.S.)-- Wichita State University, College of Engineering, Dept. of Electrical Engineering and Computer Science