Timescale Ensemble Performance Enhancement
through use of Artificial Neural Network
Shikha Maharana
ISRO Telemetry Tracking and Command Network
Indian Space Research Organization
Plot 12 &13, 3
rd
Main,2
nd
Phase, Peenya
Bangalore, 560058
shikha_maharana@istrac.org
Aakanksha Avnish Bhardwajan
ISRO Telemetry Tracking and Command
Network
Indian Space Research Organization
Plot 12 &13, 3
rd
Main,2
nd
Phase, Peenya
Bangalore, 560058
aakanksha_bhardwajan@istrac.org
T Subramanya Ganesh
ISRO Telemetry Tracking and Command
Network
Indian Space Research Organization
Plot 12 &13, 3
rd
Main,2
nd
Phase, Peenya
Bangalore, 560058
ganesht@istrac.org
Ramakrishna B N
ISRO Telemetry Tracking and Command Network
Indian Space Research Organization
Plot 12 &13, 3
rd
Main,2
nd
Phase, Peenya
Bangalore, 560058
ramki@istrac.org
Copyright © 2019 by ISTRAC/ISRO. Permission granted to INCOSE to publish and use.
Abstract. A timescale is the estimate of phase and frequency of the “perfect” clock derived from the
phase and frequency of the clocks which participate in the ensemble. An ensemble of clocks is the
first step towards the realization of a timescale. In the past, several methodologies have been de-
veloped to solve the timescale problem such as KAS-2, Multi-scale Ensemble Timescale (METS),
ALGOS (BIPM), AT1 (NIST) and algorithm for UTC (CRL). An adaptive ensemble algorithm based
on Artificial Neural Network (ANN) has been developed which shows improvements in terms of
betterment of frequency stability of the ensemble. The Neural Network in the ensemble algorithm
learns differently for different sample spaces of the input phase data that is provided. The weights of
the participating clocks are derived from their respective Allan deviations. The weights formulated
for the ensemble undergo a regressive iteration to obtain the training data set for parameterization of
weights. A single-layer perceptron feedback Neural Network that is proposed in this paper, dynam-
ically adapts the weights assigned to the participating clocks based on the paper clock performance.
This paper dwells on the comparison of performances between the ANN-based algorithms with that
of two other algorithms, namely Reduced Kalman filter based algorithm and METS. Results of
ANN-based algorithm for different clock combinations are also presented. The behavior of this al-
gorithm in the event of measurement data outage has also been presented in the paper.
Keywords: Active Hydrogen MASER (AHM), IRNSS Network Timing facility (IRNWT), Artificial Neural
Networks (ANN), Allan Deviation (ADEV), Inverse Allan Variance (IAV), Overlapped Allan Variance (OAV),
Multi-scale Ensemble Timescale (METS)