Citation: Bui, T.Q.T.; Elango, A.;
Landry, R.J. FPGA-Based
Autonomous GPS-Disciplined
Oscillatorsfor Wireless Sensor
Network Nodes. Sensors 2022, 22,
3135. https://doi.org/10.3390/
s22093135
Academic Editor: Yuh-Shyan Chen
Received: 16 February 2022
Accepted: 17 April 2022
Published: 20 April 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
FPGA-Based Autonomous GPS-Disciplined Oscillatorsfor
Wireless Sensor Network Nodes
Toan Quang The Bui * , Arul Elango
†
and René Jr. Landry
†
École de Technologie Supérieure, Université du Quebéc, Montréal, QC H3C 1K3, Canada;
arul.elango@ens.etsmtl.ca (A.E.); rene.landry@etsmtl.ca (R.J.L.)
* Correspondence: quang-the-toan.bui.1@ens.etsmtl.ca
† These authors contributed equally to this work.
Abstract:
Numerous devices in distributed wireless sensor arrays require a high-accuracy timing
reference. Although the GPS-disciplined oscillators have been developed for decades, the hardware
design still has performance limitations. In this context, we present the hardware implementation
for a GPS-disciplined oscillator with an automatic adaptive drift correction algorithm, which is
implemented in a low-cost, high-speed field-programmable gate array (FPGA) device. The system
design and the hardware implementation are presented to demonstrate the advantages of the pro-
posed oscillator. To verify this oscillator in real-time applications, we tested the device in multiple
environments and compared it to state-of-the-art designs. The experimental results showed that
our proposed device has a low cost and high performance. This device can achieve less than 80ns
and 356 ns in 1PPS signal drift in the indoor environment test and the outdoor environment test,
respectively, after 24 h of working without a GPS signal.
Keywords: TCXO; GPSDO; timing correction; Kalman filter; FPGA; wireless sensor network
1. Introduction
High-precision Global Positioning System (GPS)-disciplined oscillators (GPSDOs)
have been designed to work as a source of reference timing, whose output is controlled
to agree with the clock signals generated and broadcast by GPS in the Global Navigation
Satellite System (GNSS) [
1
,
2
]. GPSDOs are widely used in measurement devices and
communication systems due to their cost-effective, high-precision, and self-calibrating
operation. The work in [
3
] presented the use of GPSDOs in synchronizing network radar.
The GPS-disciplined clock was also used in an analog-to-digital converter (ADC) for phase
measurement application as presented in [
4
]. Moreover, GPS signals have been used in
wireless sensor networks (WSNs) for positioning the sensor nodes in recent research [
5
–
7
].
In this context, the timing calibration and synchronization between local oscillators in
devices and the references are important to obtain high-accuracy results. The commonly
used GPS modules can perform time synchronization with a resolution of 100 ns or smaller.
Numerous GPSDOs are available on the market and in research laboratories. Four devices
of one specific low-cost GPSDO type are characterized and compared using a tested GPSDO
combining a GNSS signal simulator and a reference signal stability measurement system
based on software-defined radio (SDR) and digital signal processing (DSP) [8].
However, there are several factors that would affect the accuracy of the frequency
correction [
9
]. The first is the random errors and outlines contained in one pulse per second
(PPS), which is the broadcast output of the GPSs [
10
]. The second factor is the frequency
nonlinear drift of the crystal oscillators because of the aging and the temperature [
11
].
Therefore, several works have been developed to filter designs to eliminate the noise and
adaptive algorithms are applied to compensate the temperature and aging effects [
12
,
13
].
However, most of the previous designs have a shortage of investigating the relationship
between frequency variation and temperature change. In this paper, we present an adaptive
Sensors 2022, 22, 3135. https://doi.org/10.3390/s22093135 https://www.mdpi.com/journal/sensors
