Content uploaded by Șorecău Mirela
Author content
All content in this area was uploaded by Șorecău Mirela on Jul 06, 2023
Content may be subject to copyright.
Citation: S
,orec˘au, M.; S
,orec˘au, E.;
Sârbu, A.; Bechet, P. Real-Time
Statistical Measurement of Wideband
Signals Based on Software Defined
Radio Technology. Electronics 2023,
12, 2920. https://doi.org/10.3390/
electronics12132920
Academic Editor: Filippo Costa
Received: 26 May 2023
Revised: 22 June 2023
Accepted: 30 June 2023
Published: 3 July 2023
Copyright: © 2023 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/).
electronics
Article
Real-Time Statistical Measurement of Wideband Signals Based
on Software Defined Radio Technology
Mirela S
,orecău1, 2, *, Emil S
,orecău1,2, Annamaria Sârbu 2and Paul Bechet 2
1Department of Electrotechnics and Measurements, Technical University of Cluj Napoca,
400114 Cluj-Napoca, Romania; emil.sorecau@campus.utcluj.ro
2Communications, IT and Cyber Defense Department, “Nicolae Bălcescu” Land Forces Academy,
550170 Sibiu, Romania; sarbu.annamaria@armyacademy.ro (A.S.); bechet.paul@armyacademy.ro (P.B.)
*Correspondence: sorecau.mirela@armyacademy.ro
Abstract:
The increase in channel bandwidth and peak-to-average power ratio (PAPR) of modern
communication standards poses a serious challenge to performing channel power (CP) and com-
plementary cumulative distribution function (CCDF) measurements in real-time using standard
measurement solutions based on spectrum analyzers (SA). Recently, Software Defined Radio (SDR)
technology has become a viable alternative to the conventional real-time spectrum monitoring ap-
proach based on SA due to its reduced cost. Therefore, in this paper, we propose a novel, innovative,
agile and cost-effective solution to enable both CP and CCDF measurements on a state-of-the-art SDR
platform. The proposed solution exploits the ability of the SDR equipment to access signal samples in
the time domain and defines both CP and CCDF-type measurements. The two measurement func-
tions are software implemented in GNU Radio by designing customized blocks and integrated into
a graphical user interface. The proposed system was first tested and parameterized in a controlled
environment using emitted signals specific to the IEEE 802.11ax family of wireless local area networks.
After parameterization, the SDR-based system was used for over-the-air measurements of signals
emitted in the 4G+, 5G and 802.11ax communication standards. By performing the measurement
campaign, we have demonstrated the capabilities of the measurement system in performing real-time
measurements on broadband channels (up to 160 MHz for IEEE 802.11ax). Altogether, we have
proved the usability of CP and CCDF measurement functions in providing valuable insights into
the power distribution characteristics of signals emitted by the latest communication standards. By
exploiting the versatility of SDR technology, we have enabled a cost-effective solution for advanced
real-time statistical measurements of modern broadband signals.
Keywords:
software defined radio; spectrum analyzer; channel power; complementary cumulative
distribution function; IEEE 802.11ax; LTE-A; 5G; real-time spectrum monitoring
1. Introduction
New communication technologies have led to significant shifts in the way the electro-
magnetic spectrum is used. Signals that generate electromagnetic fields, such as WLAN,
LTE or 5G NR, exhibit quasi-stochastic variations, and their massive deployment is leading
to an unprecedented increase in the density of electromagnetic spectrum consumption.
Since 1980, mobile communications have been the technology with the most significant
changes in system architecture, constantly focused on meeting the growing needs of users.
The introduction of the third-generation communication networks (3G) in 2000 brought
higher spectral efficiency and radio frequency bandwidths of up to 5 MHz [
1
,
2
]. The 3G
standard has been continuously improved to increase data throughput by implementing
new enhancements such as High-Speed Packet Access (HSPA+) [
3
]. As 3G networks were
unable to meet the ever-increasing demands for high data rates, the fourth-generation tech-
nology (4G) was introduced by the 3rd Generation Partnership Project (3GPP) to meet these
Electronics 2023,12, 2920. https://doi.org/10.3390/electronics12132920 https://www.mdpi.com/journal/electronics
Electronics 2023,12, 2920 2 of 24
requirements. Although LTE is being promoted as a 4G standard, it does not fully meet the
technical requirements set by the International Telecommunications Union (ITU) [
4
]. Even
so, LTE is capable of delivering data rates of up to 300 Mbps through the use of scalable
bandwidths from 1.4 MHz to 20 MHz, the implementation of a MIMO configuration of
up to 4
×
4 for download and 1
×
1 for upload, and the first use of OFDM (Orthogonal
Frequency Division Multiplexing).
ITU technical requirements for 4G were met through the implementation of the LTE-
Advanced (LTE-A) standard. LTE-A inherits existing LTE features but brings significant
changes: aggregation of up to five carriers (5
×
20 = 100 MHz) bandwidth, which implicitly
increases throughput; use of 8
×
8 MIMO configuration for download and 4
×
4 for
upload; implementation of new concepts such as Coordinated Multi-Point transmission
and reception (CoMP) and improved Self-organizing Network (SON) [1].
LTE-A also uses OFDM technology, but 3GPP is making some adjustments, such as
using OFDMA for the download and SC-FDMA for the uplink. The use of OFDMA solves
one of the most serious problems associated with increasing bit rates, namely intersymbol
interference. The immunity to the effects of multiple reflections on the receiver channel
capacity comes from the simple algorithm that the OFDMA system transmits information
using multiple orthogonal narrowband subcarriers, where the bit rate for each subcarrier is
inversely proportional to the total number of subcarriers. Depending on the bandwidth,
we can have N subcarriers and the transmission on each subcarrier is performed at a bit
rate of 1/N times the bit rate of the source signal [
5
]. Although OFDM is favored in 4G
wireless communications for its ability to eliminate multipath fading and increase spectral
efficiency, it has one major limitation: high peak-to-average power ratio (PAPR) [
6
–
8
]. The
main common interest of these three papers is to study and improve the performance of
wireless communication systems by addressing key issues such as reducing PAPR, im-
proving spectral efficiency, and optimizing resource allocation in OFDMA and SC-FDMA
technologies. The PAPR is further increased by the use of higher order modulation schemes:
e.g., for GSM PAPR was 0.0 dB, for WCDMA 3.5 dB, for LTE 11 dB and for 5GNR is 13 dB.
For LTE networks, in the uplink, the high PAPR signal can be handled by a NodeB, in
the downlink case the user equipment usually has limited battery power. Therefore, to
increase the transmission efficiency and reduce the cost of using mobile station power am-
plifiers, the SC-FDMA technique is used for the uplink, with a significantly reduced PAPR
compared to a standard OFDM signal [
9
]. The aforementioned research and many recent
studies [
10
,
11
] focus solely on PAPR reduction. However, 4G and beyond technologies
need customized electromagnetic measurement platforms capable of providing complete
signal characterization, including statistical analysis of the temporal evolution of the PAPR.
This identified technological limitation is to be addressed in this paper.
Although 4G is currently the most widely deployed mobile technology, a new genera-
tion of internationally standardized technologies has the potential to completely transform
mobile communications [
12
]. The fifth generation (5G) wireless network, called New Radio
(NR), has been developed to provide ultra-reliable low-latency communications (URLLC),
much higher data transmission speeds (about 100 times faster than 4G), massive MIMO,
enhanced mobile broadband communications, millimeter-wave communications, and mas-
sive machine-to-machine communications (mMTC) for implementing Internet of Things
services and applications [
13
,
14
]. Unlike 4G, which uses frequencies below 6 GHz, 5G NR
uses both frequency band 1 (FR1) between 450 MHz and 7125 MHz and FR2, also known
as the millimeter wave band, between 24 GHz and 50 GHz. Depending on the frequency
range, the bandwidths vary, so the maximum achievable bandwidth for FR1 is 100 MHz
and for FR2 it is 400 MHz [
15
]. The Cyclic Prefix Orthogonal Frequency Division Multi-
plexing (CP-OFDM) channel resource access technique for 5G NR [
16
] is similar to OFDM
in LTE, except that the subcarrier spacing is no longer 15 kHz, but a variable subcarrier
spacing is used, which implicitly also changes the cyclic prefix duration per symbol.
Statistics show that over the past few years, 50% of Internet traffic is carried by
LTE, with the other half carried by Wi-Fi [
17
]. The capacity of Wi-Fi networks has been
Electronics 2023,12, 2920 3 of 24
significantly improved with the release of IEEE 802.11ax, which aims to improve spectral
efficiency and end-user throughput in densely deployed Wi-Fi environments. OFDMA
is the radio access scheme implemented by the Wi-Fi 6 (802.11ax) standard to enable
efficient use and allocation of spectrum, reduce transmission latency, improve power
spectral density, and enhance the overall user experience. OFDMA provides performance
by dividing the channel or available 20/40/80/160 (80 + 80) MHz frequency bands into
hundreds of subchannels, allowing devices such as access points to transmit data to multiple
stations simultaneously and vice versa. Sixth generation Wi-Fi (IEEE 802.11ax) supports up
to 74 users at 160 MHz channel bandwidth, 37 users at 80 MHz, 18 users at 40 MHz and
9 users at 20 MHz for the mentioned bandwidths in the 5 GHz band [18,19].
In summary, both cellular and wireless communication standards have diversified
and evolved to meet the increasing user demands. Today, the electromagnetic spectrum is
occupied with communication channels of variable bandwidths: 5 MHz (3G), 20 MHz (4G,
3rd generation Wi-Fi), 100 MHz (5G), 160 MHz (IEEE 802.11ax); radio access techniques
have diversified, digital modulation methods have improved, and smart antennas have
changed the communication quality paradigm by using MU-MIMO and beamforming
techniques. The complexity of modern communication channels poses challenges for
efficient spectrum measurement and monitoring. To address these challenges, there is a
need to fill this technological gap and identify an agile measurement solution that can
provide real-time information about the spectrum together with the implementation of
advanced electromagnetic spectrum measurement capabilities. This issue is further detailed
in the next section of the paper, where we provide a detailed discussion of advanced
spectrum measurement capabilities.
2. Advanced Electromagnetic Spectrum Measurement Capabilities
There is a wide range of signals in the electromagnetic spectrum, with high variability
not only in the frequency domain, but also in the other dimensions: time, amplitude,
and space. New generation technologies exhibit random variations in the structure of
the signals, requiring the development of measurement techniques that address all four
dimensions of the electromagnetic spectrum.
Considering that new communication systems and technologies are evolving rapidly,
the ITU-R Recommendation SM 2039-2013 urges monitoring systems to extend their ca-
pabilities to accommodate signals specific to the latest generations of communication
technologies [
20
]. As a result, electromagnetic spectrum monitoring is becoming more
complex, especially when real-time monitoring capabilities must be addressed. The main
element of an electromagnetic spectrum monitoring system is the spectrum analyzer (SA),
which can be a swept SA or a real-time SA [21].
The difference between the two analyzers in terms of monitoring variable signals,
not only in amplitude and frequency but also in phase, is worth mentioning. The sweep
analyzer acts as a passband filter, given by the resolution bandwidth (RBW), that moves
like a window and gradually monitors the frequency domain of interest (SPAN) at a
sweep rate that is directly proportional to the ratio of SPAN to sweep time (SWT). Vector
signal analyzers can measure both the magnitude and phase of the input signal at a
single frequency within the intermediary frequency bandwidth of the instrument. As the
significance of speed in frequency analysis increases, the traditional capabilities of swept
SAs, which display measurements sequentially, are proving to be insufficient. To overcome
this limitation, real-time SAs have been developed to simultaneously display the power of
all frequency components within the achievable real-time bandwidth.
Real-time SAs offer a significant advantage over swept SAs by providing instanta-
neous and continuous measurements across the entire frequency range of interest. Rather
than sweeping through frequencies sequentially, real-time analyzers capture and process
the entire spectrum simultaneously, enabling users to observe all frequency components
in real-time. Real-time SAs utilize advanced signal processing techniques and high-speed
analog-to-digital converters to capture and analyze wideband signals in real-time. These
Electronics 2023,12, 2920 4 of 24
analyzers can handle complex and dynamic signal environments, making them well-
suited for applications involving broadband signals, wideband modulation schemes, and
frequency-hopping technologies but these capabilities significantly increase the cost of the
measurement system. This very high acquisition cost affects not only the scalability of
such systems but also the flexibility in the context of mobility. Many recent studies [
22
–
24
]
suggest a low-cost solution to replace such high-cost equipment in the form of Software
Defined Radio technology. Software-Defined Radio (SDR) technology is a wireless com-
munication system that uses software-based processing to perform tasks traditionally
handled by dedicated hardware components, enabling flexibility and reconfigurability in
radio communications.
As with SAs, the performance of the SDR platform can affect the ability to measure
signals with different bandwidths, resolutions, frequency ranges, etc. In [
25
], the Adalm
PLUTO SDR platform was used to measure LTE signals, with a discussion regarding
its ability to detect low amplitude signals. In addition, the Adalm Pluto SDR platform
limits the instantaneous acquisition frequency bandwidth to 20 MHz, which restricts the
monitoring of newer technologies, such as 5G or WLAN. Continuing in the same direction,
Helbet et al. [
26
] tested three SDR platforms with different analog-to-digital converter
(ADC) resolutions (HackRF one—8 bit, Adalm PLUTO—12 bit, NI USRP 2930—14 bit)
showing performances comparable to a classic SA. In the same paper, the authors report
that for levels close to the noise threshold, the HackRF one platform does not represent an
efficient measurement solution. Therefore, a very important parameter of SDR platforms
is the resolution of the analog-to-digital converter which impacts their ability to measure
low-amplitude signals.
To ensure the highest level of accuracy in measurements, spectrum monitoring systems
must adapt to the evolving signal characteristics of advanced technologies, such as 5G NR
and WLAN, mitigating for improved spectrum monitoring capabilities. A type of spec-
trum measurement used in wireless system networks is channel power (CP), performed to
measure the power transmitted within a specified frequency or channel bandwidth. CP mea-
surements serve multiple purposes, including validating the performance of transmitters,
ensuring compliance with government regulations, and minimizing system interference.
Modern SAs possess the capability to automatically compute CP by integrating power over
the channel bandwidth. In the traditional approach using SAs, it is difficult to implement
CP measurements on wideband signals from the latest communication standards due to the
overwhelming hardware requirements. The high cost of SAs further limits high-bandwidth
studies, and real-time implementations on the order of hundreds of MHz using SAs are
currently available only at very high cost.
CP measurements become an insufficient characterization of signals that have high
temporal variability, like those used by new-generation communication devices. Addition-
ally, these modern signals require a measurement tool capable of providing direct statistical
insights and graphical representation for the amplitude variation in time. To this extent, the
SA proves to be a valuable tool for obtaining an accurate and complex statistical assessment
of signal variation. SAs enable evaluation and analysis of PAPR with a high degree of
accuracy and can perform statistical analysis of measurements even over long periods of
time, but are limited by the bandwidth, which in the case of new technologies is very high.
In the context of OFDM signals, which are known for their high PAPR, it is essential to use
advanced statistical measurements such as Probability Density Function (PDF), Cumulative
Distribution Function (CDF), Complementary Cumulative Distribution Function (CCDF) or
Amplitude Probability Density (APD). The use of these advanced statistical measurement
functions proves to be of interest for designing power-efficient amplifiers [27].
The probability density function parameter describes the relative probability that the
measured power will take a particular value. The information contained in a PDF is similar
to that of a histogram but shows continuous rather than discrete values. Using a PDF
function, it is possible to determine the probability of a measured power falling between
two limits by integrating over that range. The PDF integration over the entire function
Electronics 2023,12, 2920 5 of 24
becomes what we call the CDF. The CDF curve tells us the percentage of time that the signal
power is below a given value, or more specifically, it is very useful when performing noise
analysis because it highlights the minimum power values. Using the CDF function, it is
also possible to obtain its complement, the CCDF curve. The values of the CCDF curve
are calculated as follows: CCDF = 1
−
CDF [
28
]. The CCDF curve is plotted as a graph,
with the power level expressed in dB above the average signal power on the Ox axis and
the percentage of time the amplitude reached the specific power value on the Oy axis. The
CCDF function, unlike the CDF, focuses on the maximum or peak power values. More
specifically, it provides information about the probability or percentage of time that a signal
is at a particular value or exceeds the average value of the signal [
29
,
30
]. The APD method,
also a statistical measurement, is suitable for studying the characteristics of a signal’s
amplitude variations. APD defines the probability of an amplitude occurring within a
defined bandwidth and time interval [
31
]. The statistical description of the power levels of
the signal, as in the case of APD measurements, is achieved by using time domain data,
with the note that for a complete description of the signal characteristics, it is necessary to
evaluate the CCDF curve.
However, the interest in peak values and signal power statistics makes CCDF statistical
measurements very important and not to be neglected in the quasi-stochastic characteriza-
tion of new signals. Many studies approach the technique of PAPR reduction of OFDM
signals using statistical CCDF measurements [
32
–
34
], but to analyze this parameter in de-
tail, a statistical analysis of its evolution in time is required. CCDF curves are also important
in the design of power amplifiers due to the nonlinear nature of some components, as well
as for testing and debugging devices to assess compression effects [35].
The motivation for this study stems from the authors’ interest in implementing various
applications on SDR platforms (both low-cost and higher-cost) to highlight their capabilities
in terms of performance, cost, and configuration flexibility. The SDR platforms match and
even exceed the performance of real-time spectral analyzers, considering that the price of
these instruments increases significantly for real-time monitoring of larger bandwidths.
Given the need to address the latest standards, such as 5G and IEEE 802.11ax, and to per-
form multiple measurements, there are few instruments on the market that are affordable.
For higher-order statistical analysis, the acquisition bandwidth must be equal to the channel
bandwidth, and for technologies such as 4G, 5G, 802.11 ac/ax, this is difficult to achieve.
This means that we need a system with high acquisition bandwidth, and traditional spec-
trum analyzers are not capable of analyzing such statistical measurements, and those that
do have these capabilities come at a particularly high price. To implement CCDF and APD
statistical analysis, the acquisition bandwidth must be equal to the channel bandwidth.
This is a significant limitation even for high-performance analyzers. A real-time bandwidth
of 160 MHz is a critical parameter. For this, the cost of the analyzer becomes very high and
the SDR solution is a viable alternative at an acceptable cost.
Therefore, we propose a cost-effective, real-time, large bandwidth, SDR-based mea-
surement system with integrated CP and CCDF measurement capabilities. Both CP and
CCDF measurements are implemented in software and the proposed system is tested
using a traditional SA and signal generator-based setup. After parameterization, the mea-
surement system is applied to 4G/5G and Wi-Fi signals, demonstrating the versatility,
scalability and adaptability of SDR platforms to modern communication standards through
software reconfiguration.
This research makes some notable contributions to the existing literature. Specifically,
it focuses on the use of SDR platforms for spectrum monitoring and analysis, providing
a cost-effective and efficient alternative to traditional spectrum analyzers. This approach
allows for flexibility and customization in the implementation of measurement functions.
The research presents the design and implementation of an SDR-based real-time spectrum
monitoring system capable of monitoring extended frequency bands while incorporating
channel power (CP) measurements and complementary cumulative distribution function
(CCDF). This integrated feature set enables comprehensive analysis of signal quality and
Electronics 2023,12, 2920 6 of 24
statistical parameters. In addition, the research includes performance evaluations compar-
ing the SDR system to a high-performance spectrum analyzer, highlighting the potential,
capabilities and limitations of SDR platforms in spectrum monitoring. As a result, this
research provides new insights and practical implementation strategies that differentiate it
from existing published material in this area.
The paper is structured as follows: the Materials and Methods section, which consists
of four subsections dealing with the hardware and software implementation of the mea-
surement system, the parameterization of the system, and its validation in realistic wireless
environments. This is followed by the Results and Discussion section, which is divided
into two subsections, and at the end the Conclusions of the paper are presented.
3. Materials and Methods
3.1. Hardware Configuration
The goal of this work is to develop an affordable measurement system that can
effectively replace a high-performance SA in performing advanced signal statistical mea-
surements. To evaluate the capabilities of the system, we followed a methodology that
involved first generating reference signals using a signal generator and then performing
tests with real signals to simulate different scenarios. The performance of the system was
then evaluated by comparing the results with measurements obtained using a SA with
real-time capabilities. This comparison method was used to validate the accuracy and
reliability of the measurement system results compared to the established spectral analyzer.
In order for a monitoring station to qualify as real-time, it must capture IQ compo-
nents in the baseband or intermediate frequency (IF) range and then perform processing
on these components. This allows complex analysis of the electromagnetic spectrum (am-
plitude analysis and phase analysis). SDR platforms are designed to enable flexible and
programmable radio systems using software-based signal processing techniques. In the
case of SDR technology, the acquisition bandwidth is dependent on the sampling frequency
(the channel bandwidth is given by the sampling frequency).
To ensure real-time measurement capability and high flexibility in the development
of the monitoring system, the main starting point was to use the Universal Software
Radio Peripheral (USRP)—an SDR platform from Ettus Research—together with the GNU
Radio Companion software, upgraded with some custom blocks developed in Python
scripts. Based on this concept, a computer was assembled with the Linux operating system
(Ubuntu 20.04) and GNU Radio installed. The USRP X310 was also chosen as the SDR
platform, with a UBX 10-6000 MHz Rx/Tx receiver board with a full-duplex wideband
transceiver and an instantaneous bandwidth of up to 160 MHz. The main features of this
SDR are shown in Table 1[36].
The connection between the computer and the SDR was made over multimode fiber
using a GbE SFP+ with two 10 Gbps ports to provide up to 10 Gbps transmission speeds
with low interference (Figure 1). The computer was equipped with an eighth-generation
Intel
®
Core
™
Pro i7 processor and 32 GB of RAM, with an additional 2 TB PCIe 4.0 NVMe
solid state drive (SSD) for I and Q data recording capabilities at 160 MHz bandwidth
(2 ×16 bits ×160 MS/s = 5.12 Gbps).
In parallel to this work, another system made of a signal generator and spectrum
analyzer was set up to serve as a reference system for the parameterization and testing of the
prototype system. The reference system consists of two Rohde and Schwarz instruments:
SMM100A signal generator and FSVR7 real-time spectrum analyzer with channel power
and CCDF capability and a laptop for SciPy control.
The parameterization and testing of the proposed system were first performed under
laboratory conditions, where the IEEE 802.11ax signals were provided by a vector signal
generator. Following the controlled environment, the measurement system was used in an
uncontrolled environment, in an over-the-air measurement campaign for signals emitted in
the 4G/5G and Wi-Fi 6 communication standards.
Electronics 2023,12, 2920 7 of 24
Table 1. Main features of the USRP X310 platform.
Hardware
Capabilities
•2 transceiver card slots
•Dual SFP+ Transceivers (1 GigE/10 GigE)
•PCI Express over cable (MXI) gen1 ×4
•External PPS input and output
•External reference (10 MHz, 11.52 MHz, 23.04 MHz, or 30.72 MHz) input and output
•Expandable via 2nd SFP+ interface
•Supported master clock rates: 200 MHz and 184.32 MHz
•Variable daughterboard clock rates
•External GPIO Connector with UHD API control
•External USB Connection for built-in JTAG debugger
•Internal GPSDO option
FPGA
Capabilities
•2 RX DDC chains in FPGA
•2 TX DUC chain in FPGA
•Timed commands in FPGA
•Timed sampling in FPGA
•Up to 120 MHz of RF bandwidth with 16-bit samples
UBX 160 USRP
Daughterboard
Capabilities
•10 MHz to 6 GHz
•Up to 160 MHz Bandwidth
•RF Shielding
•Phase Alignment Capable
•Full duplex operation with independent TX and RX frequencies
Electronics 2023, 12, x FOR PEER REVIEW 7 of 25
Table 1. Main features of the USRP X310 platform.
Hardware
Capabilities
• 2 transceiver card slots
• Dual SFP+ Transceivers (1 GigE/10 GigE)
• PCI Express over cable (MXI) gen1 ×4
• External PPS input and output
• External reference (10 MHz, 11.52 MHz, 23.04 MHz, or 30.72
MHz) input and output
• Expandable via 2nd SFP+ interface
• Supported master clock rates: 200 MHz and 184.32 MHz
• Variable daughterboard clock rates
• External GPIO Connector with UHD API control
• External USB Connection for built-in JTAG debugger
• Internal GPSDO option
FPGA
Capabilities
• 2 RX DDC chains in FPGA
• 2 TX DUC chain in FPGA
• Timed commands in FPGA
• Timed sampling in FPGA
• Up to 120 MHz of RF bandwidth with 16-bit samples
UBX 160 USRP
Daughterboard
Capabilities
• 10 MHz to 6 GHz
• Up to 160 MHz Bandwidth
• RF Shielding
• Phase Alignment Capable
• Full duplex operation with independent TX and RX frequencies
The connection between the computer and the SDR was made over multimode fiber
using a GbE SFP+ with two 10 Gbps ports to provide up to 10 Gbps transmission speeds
with low interference (Figure 1). The computer was equipped with an eighth-generation
Intel® Core™ Pro i7 processor and 32 GB of RAM, with an additional 2 TB PCIe 4.0 NVMe
solid state drive (SSD) for I and Q data recording capabilities at 160 MHz bandwidth (2 ×
16 bits × 160 MS/s = 5.12 Gbps).
Figure 1. Hardware components of the system.
In parallel to this work, another system made of a signal generator and spectrum
analyzer was set up to serve as a reference system for the parameterization and testing of
the prototype system. The reference system consists of two Rohde and Schwarz
instruments: SMM100A signal generator and FSVR7 real-time spectrum analyzer with
channel power and CCDF capability and a laptop for SciPy control.
The parameterization and testing of the proposed system were first performed under
laboratory conditions, where the IEEE 802.11ax signals were provided by a vector signal
generator. Following the controlled environment, the measurement system was used in
an uncontrolled environment, in an over-the-air measurement campaign for signals
emied in the 4G/5G and Wi-Fi 6 communication standards.
Figure 1. Hardware components of the system.
3.2. Software Implementation
The main challenge of this project was the implementation of the software, which
involved both digital signal processing and Python programming skills. Starting from
scratch with the GNU Radio libraries and the Python programming language, we managed
to implement the main functions of a SA, such as real-time frequency domain spectrum
analysis, CP and CCDF measurement functions, with the possibility of saving the results,
for bandwidths up to 160 MHz.
GNU Radio is an open-source software development toolkit that enables the efficient
design and implementation of software-defined radios (SDRs) and wireless communica-
tion systems through a wide range of signal-processing blocks [
37
]. It is written in the
Python and C++ programming languages. GNU Radio provides a large collection of signal
processing blocks for a variety of tasks such as modulation, demodulation, filtering, and
coding. These blocks can be used to build custom communication systems, such as those
for wireless communications, digital signal processing, and software-defined radio. The
blocks can be combined and connected in a variety of ways to form a flowchart that rep-
resents the structure of the processing system. In the development of the CCDF and CP
measurement application, the workflow (Figure 2) was divided into three stages: sample
reception, filtering and spectrum monitoring (Figure 3), CP measurement, display and data
save (Figure 4), CCDF calculation and plotting (Figure 5).
Electronics 2023,12, 2920 8 of 24
Electronics 2023, 12, x FOR PEER REVIEW 8 of 25
3.2. Software Implementation
The main challenge of this project was the implementation of the software, which
involved both digital signal processing and Python programming skills. Starting from
scratch with the GNU Radio libraries and the Python programming language, we
managed to implement the main functions of a SA, such as real-time frequency domain
spectrum analysis, CP and CCDF measurement functions, with the possibility of saving
the results, for bandwidths up to 160 MHz.
GNU Radio is an open-source software development toolkit that enables the efficient
design and implementation of software-defined radios (SDRs) and wireless
communication systems through a wide range of signal-processing blocks [37]. It is
wrien in the Python and C++ programming languages. GNU Radio provides a large
collection of signal processing blocks for a variety of tasks such as modulation,
demodulation, filtering, and coding. These blocks can be used to build custom
communication systems, such as those for wireless communications, digital signal
processing, and software-defined radio. The blocks can be combined and connected in a
variety of ways to form a flowchart that represents the structure of the processing system.
In the development of the CCDF and CP measurement application, the workflow (Figure
2) was divided into three stages: sample reception, filtering and spectrum monitoring
(Figure 3), CP measurement, display and data save (Figure 4), CCDF calculation and
ploing (Figure 5).
Figure 2. Workflow diagram.
The first approach was achieved using the USRP Signal Source, Band Pass Filter,
Frequency Sink and Complex to mag^2. The USRP Signal Source is designed to be flexible
and programmable, allowing users to modify the RF front end and digital processing to
meet their specific needs. For example, in our case, the receive channel/daughterboard,
center frequency, sampling frequency, and gain were set. The sampling frequency also
determines the measurement span. Since the sampling frequency must be an even
Figure 2. Workflow diagram.
Electronics 2023, 12, x FOR PEER REVIEW 9 of 25
submultiple of the master clock rates, a bandpass filter was used to measure only the
useful bandwidths. The spectrum can be viewed after reception as well as after filtering.
After this step, the flowgraph is split in two. All samples are sent on both paths and are
transformed from a complex-valued signal to its magnitude squared (Figure 3). The
complex to magnitude squared transformation is performed by calculating the square of
the magnitude of the complex signal (calculated as the square root of the sum of the
squares of the real and imaginary components).
Figure 3. Flowgraph of sample acquisition, filtering, and spectrum monitoring.
In the second step (Figure 4), to measure the channel power, we need the Moving
Average block, which helps to calculate the moving average of a specific time slice. Next,
the values are converted to dBm using a correction factor that is modified according to the
reference equipment, and the result is displayed using Number Sink and saved to a .csv
file using the Data Save block. The save block does not exist in the GNU Radio library and
was developed for this application using a Python script.
Figure 4. CP measurement, display and data save.
The last step (Figure 5) is to implement the CCDF function (Complementary
Cumulative Distribution Function plot). It is a graph showing the probability that a
random variable is greater than a certain value. In the CCDF plot, the horizontal axis
represents the possible values of the random variable, and the vertical axis represents the
probability of obtaining a value greater than the value given on the horizontal axis. In
GNU Radio, the CCDF calculation function is not yet implemented. To fulfill this
requirement, it was necessary to create scripts in Python to facilitate the calculation,
Figure 3. Flowgraph of sample acquisition, filtering, and spectrum monitoring.
Electronics 2023,12, 2920 9 of 24
Electronics 2023, 12, x FOR PEER REVIEW 9 of 25
submultiple of the master clock rates, a bandpass filter was used to measure only the
useful bandwidths. The spectrum can be viewed after reception as well as after filtering.
After this step, the flowgraph is split in two. All samples are sent on both paths and are
transformed from a complex-valued signal to its magnitude squared (Figure 3). The
complex to magnitude squared transformation is performed by calculating the square of
the magnitude of the complex signal (calculated as the square root of the sum of the
squares of the real and imaginary components).
Figure 3. Flowgraph of sample acquisition, filtering, and spectrum monitoring.
In the second step (Figure 4), to measure the channel power, we need the Moving
Average block, which helps to calculate the moving average of a specific time slice. Next,
the values are converted to dBm using a correction factor that is modified according to the
reference equipment, and the result is displayed using Number Sink and saved to a .csv
file using the Data Save block. The save block does not exist in the GNU Radio library and
was developed for this application using a Python script.
Figure 4. CP measurement, display and data save.
The last step (Figure 5) is to implement the CCDF function (Complementary
Cumulative Distribution Function plot). It is a graph showing the probability that a
random variable is greater than a certain value. In the CCDF plot, the horizontal axis
represents the possible values of the random variable, and the vertical axis represents the
probability of obtaining a value greater than the value given on the horizontal axis. In
GNU Radio, the CCDF calculation function is not yet implemented. To fulfill this
requirement, it was necessary to create scripts in Python to facilitate the calculation,
Figure 4. CP measurement, display and data save.
Electronics 2023, 12, x FOR PEER REVIEW 10 of 25
display, and saving of CCDF parameters. For this purpose, the following steps were taken
into account when developing the Python scripts:
1. The input data are transformed into a vector (a one-dimensional matrix) of a specified
size from which the CCDF is calculated;
2. The mean value is calculated for all measured samples in the vector, using the
reference point (0 dB) on the left side of the x-axis;
3. All samples are normalized to the mean, expressed in decibels (dB);
4. Each sample is assigned to a specific bin on the x-axis, ranging from 0 to 30 dB;
5. The total number of samples falling into each x-axis bin, relative to the total number
of samples measured, is calculated and ploed as a percentage;
6. For greater precision, the y-axis is ploed in logarithmic format.
Figure 5. Flowgraph for CCDF implementation in GNU Radio.
As can be seen in Figure 5, points 2 through 6 have been implemented in the block
wrien in Python and imported into GNU RADIO, called CCDF PLOT. This block
performs the CCDF calculation, displays the graph, and saves the data from the graph.
Figure 6 shows the results. In order not to overload the displayed graph, the display and
storage of the mean, peak and crest values from the CCDF calculation were implemented
separately.
Figure 6 shows the final result of the software implementation when analyzing a
wireless channel with a bandwidth of 40 MHz. On the main screen, the filtered wireless
channel is displayed in the frequency domain, and on the left side, the parameters such as
the Crest, Mean, Peak and Channel Power are also displayed. The second screen shows
the CCDF graph, with the amplitude above the mean signal value (in dB) ploed on the
Ox axis and the percentage of time the signal reached the corresponding amplitude level
(in %) on the Oy axis. The four most significant indicators are extracted as pairs of
percentages (10%, 1%, 0.1%, 0.01%)—amplitude above mean (dB) in the right area of the
graph, as can be observed in Figure 6.
Figure 5. Flowgraph for CCDF implementation in GNU Radio.
The first approach was achieved using the USRP Signal Source, Band Pass Filter, Fre-
quency Sink and Complex to magˆ2. The USRP Signal Source is designed to be flexible and
programmable, allowing users to modify the RF front end and digital processing to meet
their specific needs. For example, in our case, the receive channel/daughterboard, center
frequency, sampling frequency, and gain were set. The sampling frequency also determines
the measurement span. Since the sampling frequency must be an even submultiple of
the master clock rates, a bandpass filter was used to measure only the useful bandwidths.
The spectrum can be viewed after reception as well as after filtering. After this step, the
flowgraph is split in two. All samples are sent on both paths and are transformed from a
complex-valued signal to its magnitude squared (Figure 3). The complex to magnitude
squared transformation is performed by calculating the square of the magnitude of the
complex signal (calculated as the square root of the sum of the squares of the real and
imaginary components).
In the second step (Figure 4), to measure the channel power, we need the Moving
Average block, which helps to calculate the moving average of a specific time slice. Next,
the values are converted to dBm using a correction factor that is modified according to the
reference equipment, and the result is displayed using Number Sink and saved to a .csv
file using the Data Save block. The save block does not exist in the GNU Radio library and
was developed for this application using a Python script.
The last step (Figure 5) is to implement the CCDF function (Complementary Cumu-
lative Distribution Function plot). It is a graph showing the probability that a random
Electronics 2023,12, 2920 10 of 24
variable is greater than a certain value. In the CCDF plot, the horizontal axis represents
the possible values of the random variable, and the vertical axis represents the probability
of obtaining a value greater than the value given on the horizontal axis. In GNU Radio,
the CCDF calculation function is not yet implemented. To fulfill this requirement, it was
necessary to create scripts in Python to facilitate the calculation, display, and saving of
CCDF parameters. For this purpose, the following steps were taken into account when
developing the Python scripts:
1.
The input data are transformed into a vector (a one-dimensional matrix) of a specified
size from which the CCDF is calculated;
2.
The mean value is calculated for all measured samples in the vector, using the reference
point (0 dB) on the left side of the x-axis;
3. All samples are normalized to the mean, expressed in decibels (dB);
4. Each sample is assigned to a specific bin on the x-axis, ranging from 0 to 30 dB;
5.
The total number of samples falling into each x-axis bin, relative to the total number
of samples measured, is calculated and plotted as a percentage;
6. For greater precision, the y-axis is plotted in logarithmic format.
As can be seen in Figure 5, points 2 through 6 have been implemented in the block
written in Python and imported into GNU RADIO, called CCDF PLOT. This block performs
the CCDF calculation, displays the graph, and saves the data from the graph. Figure 6
shows the results. In order not to overload the displayed graph, the display and storage of
the mean, peak and crest values from the CCDF calculation were implemented separately.
Electronics 2023, 12, x FOR PEER REVIEW 11 of 25
Figure 6. Complete software implementation—graphical user interface.
3.3. Parameterization and Test System Implementation
In order for the proposed system to correctly measure all the parameters
implemented in the software, it is necessary to use the reference system to adjust them.
This process is called parameterization. In this case, parameterization of the software
refers to the fact that the software is adapted to the desired range of functions by seing
parameters in relation to the results displayed by the reference system. For this purpose,
the Rohde and Schwarz SMM100A signal generator was connected to one of the two RX
channels of the USRP platform via a coaxial cable. The USRP device was also connected
to the PC via fiber optic cable, where the application implemented in GNU Radio was
running (Figure 7).
Figure 7. Experimental test setup for seing parameters.
Figure 6. Complete software implementation—graphical user interface.
Figure 6shows the final result of the software implementation when analyzing a
wireless channel with a bandwidth of 40 MHz. On the main screen, the filtered wireless
channel is displayed in the frequency domain, and on the left side, the parameters such as
the Crest, Mean, Peak and Channel Power are also displayed. The second screen shows the
CCDF graph, with the amplitude above the mean signal value (in dB) plotted on the Ox
Electronics 2023,12, 2920 11 of 24
axis and the percentage of time the signal reached the corresponding amplitude level (in %)
on the Oy axis. The four most significant indicators are extracted as pairs of percentages
(10%, 1%, 0.1%, 0.01%)—amplitude above mean (dB) in the right area of the graph, as can
be observed in Figure 6.
3.3. Parameterization and Test System Implementation
In order for the proposed system to correctly measure all the parameters implemented
in the software, it is necessary to use the reference system to adjust them. This process is
called parameterization. In this case, parameterization of the software refers to the fact that
the software is adapted to the desired range of functions by setting parameters in relation
to the results displayed by the reference system. For this purpose, the Rohde and Schwarz
SMM100A signal generator was connected to one of the two RX channels of the USRP
platform via a coaxial cable. The USRP device was also connected to the PC via fiber optic
cable, where the application implemented in GNU Radio was running (Figure 7).
Electronics 2023, 12, x FOR PEER REVIEW 11 of 25
Figure 6. Complete software implementation—graphical user interface.
3.3. Parameterization and Test System Implementation
In order for the proposed system to correctly measure all the parameters
implemented in the software, it is necessary to use the reference system to adjust them.
This process is called parameterization. In this case, parameterization of the software
refers to the fact that the software is adapted to the desired range of functions by seing
parameters in relation to the results displayed by the reference system. For this purpose,
the Rohde and Schwarz SMM100A signal generator was connected to one of the two RX
channels of the USRP platform via a coaxial cable. The USRP device was also connected
to the PC via fiber optic cable, where the application implemented in GNU Radio was
running (Figure 7).
Figure 7. Experimental test setup for seing parameters.
Figure 7. Experimental test setup for setting parameters.
The test signals were verified by comparison with the Rohde and Schwarz FSVR7
real-time SA (frequency range 10 Hz–7 GHz). The R&S SMM100A vector signal generator
equipped with the SMM-K142 for the IEEE802.11ax standard was used for signal generation.
Parameterization was performed for a downlink channel with a bandwidth of 40 MHz
(the bandwidth was selected for comparison with the SA, which is limited to 40 MHz
real-time acquisition). The other set of parameters that define the 802.11ax standard are
highlighted in Table 2and Figure 8. To determine the correct parameters for the SDR
system, comparative measurements were performed.
Table 2. Description of the generated 802.11ax signal.
Link direction Downlink
Physical Layer Protocol Data Unit (PPDU) HE SU (transmitting to a single user)
Bandwidth 40 MHz
Symbol duration 12.8 µs
Guard interval duration 3.2 µs
Aggregate MAC Protocol Data Unit (A-MPDU) length 1028 bytes
Modulation schemes 1024 QAM (25% higher capacity than 802.11ac)
Data rate 286.76 Mbps
Bits per symbol 3900
Generated power level −30 dBm
Electronics 2023,12, 2920 12 of 24
Electronics 2023, 12, x FOR PEER REVIEW 12 of 25
The test signals were verified by comparison with the Rohde and Schwarz FSVR7
real-time SA (frequency range 10 Hz–7 GHz). The R&S SMM100A vector signal generator
equipped with the SMM-K142 for the IEEE802.11ax standard was used for signal
generation. Parameterization was performed for a downlink channel with a bandwidth of
40 MHz (the bandwidth was selected for comparison with the SA, which is limited to 40
MHz real-time acquisition). The other set of parameters that define the 802.11ax standard
are highlighted in Table 2 and Figure 8. To determine the correct parameters for the SDR
system, comparative measurements were performed.
Table 2. Description of the generated 802.11ax signal.
Link direction Downlink
Physical Layer Protocol Data Unit (PPDU) HE SU (transmiing to a single user)
Bandwidth 40 MHz
Symbol duration 12.8 µs
Guard interval duration 3.2 µs
Aggregate MAC Protocol Data Unit (A-MPDU)
length 1028 bytes
Modulation schemes 1024 QAM (25% higher capacity than 802.11ac)
Data rate 286.76 Mbps
Bits per symbol 3900
Generated power level −30 dBm
Figure 8. Seing parameters on the signal generator for an 802.11ax channel.
Laboratory testing ensures correct signal processing, which is achieved when the
signals transmied by the generator match in display value after the signal passes through
the SDR platform. The “Multiply” block was used to set the parameters. In the first step,
after seing the correction parameters, the amplitude response of the SDR system was
performed for the same test signal (described in Table 2), but for a dynamic range of the
input signal level between −30 dBm and −70 dBm. Comparative measurement results
between the SDR platform and the reference spectrum analyzer are shown in Table 3. It
can be seen that the SDR platform has a good response to the dynamic range of the channel
power being analyzed, with results comparable to those of the spectrum analyzer, the
differences being less than 1 dBm.
Table 3. Amplitude response for an 802.11ax channel: power variable between −70dBm and −30dBm;
bandwidth: 40 MHz.
Output Power Given by Signal
Generator R&S SMM100A [dBm]
Channel Power Measured
with R&S FSVR7 [dBm]
Channel Power Measured
with USRP X310 [dBm]
−70 −72.04 −72.17
Figure 8. Setting parameters on the signal generator for an 802.11ax channel.
Laboratory testing ensures correct signal processing, which is achieved when the
signals transmitted by the generator match in display value after the signal passes through
the SDR platform. The “Multiply” block was used to set the parameters. In the first step,
after setting the correction parameters, the amplitude response of the SDR system was
performed for the same test signal (described in Table 2), but for a dynamic range of the
input signal level between
−
30 dBm and
−
70 dBm. Comparative measurement results
between the SDR platform and the reference spectrum analyzer are shown in Table 3. It can
be seen that the SDR platform has a good response to the dynamic range of the channel
power being analyzed, with results comparable to those of the spectrum analyzer, the
differences being less than 1 dBm.
Table 3.
Amplitude response for an 802.11ax channel: power variable between
−
70dBm and
−
30dBm;
bandwidth: 40 MHz.
Output Power Given by Signal
Generator R&S SMM100A [dBm]
Channel Power Measured
with R&S FSVR7 [dBm]
Channel Power Measured
with USRP X310 [dBm]
−70 −72.04 −72.17
−60 −64.09 −64.90
−50 −54.20 −55.26
−40 −44.43 −45.11
−30 −34.20 −35.86
Under the same laboratory conditions (with the exception of the bandwidth variation),
the channel noise power was also analyzed by sequentially connecting a 50 ohms impedance
to both the RF input of the SDR platform and the reference SA. The measurement results are
shown in Table 4. We are interested in this parameter because the channel noise will have an
impact on the minimum power level, we can measure at a given time. The results obtained
show a difference of approximately 3 dB between the SDR system and the reference SA.
This means that the SDR platform is at least 3 dB less sensitive than the reference SA in its
ability to measure low-level signals.
Table 4. Noise power for 802.11ax channels with different bandwidths.
Channel Bandwidth [MHz] 20 40 80 160
Channel Power Noise measured with R&S FSVR7 [dBm] −
79.95
−
76.78
- -
Channel Power Noise measured with USRP X310 [dBm] −76.7 −73.6 −70.2 −67.7
The second step was to generate signals specific to the IEEE 802.11ax standard, for
CP measurements and statistical measurements CCDF, to see the efficiency of the SDR
Electronics 2023,12, 2920 13 of 24
instrument compared to the SA. The following settings of the SA were used for the channel
power measurements: Central frequency = 2400 MHz; SWT = 10.9 ms; N points = 860;
SPAN = 80 MHz; channel bandwidth = 40 MHz; RBW = 30 kHz, number of measurements:
2740 (equivalent to 10 802.11ax frames). The SA settings were implemented in such a way
as to ensure the accuracy of the measurements, especially with respect to sweep time and
number of pixels (interpixel time was calculated according to the symbol duration).
In contrast to CP measurements, the SWT is replaced by the number of samples to be
measured for CCDF measurements performed by the FSVR SA. Since only independent
samples affect the statistical measurement, the acquisition time (AQT) is automatically
determined and expressed as the ratio of the number of samples (N) to the resolution
bandwidth (RBW): AQT = N/RBW. Another restrictive setting of the SA is the limitation
of the resolution bandwidth in CCDF mode. This must be equal to the bandwidth of
the WLAN channel. Therefore, measurements were performed in the laboratory for a
maximum RBW of 40 MHz.
To set up the statistical measurement and obtain the CCDF curve (Figure 9), the SA
was configured with the following parameters: center frequency = 2400 MHz; acquisition
time of a CCDF measurement was set equal to the duration of a symbol specific to the
IEEE 802.11ax standard (12.8
µ
s); RBW equal to the bandwidth: 40 MHz; number of
measurements: 4284 (for the measurement of 10 frames). In addition, all calculations
for both CP and CCDF measurements were performed taking into account the PPDU
transmission time (PSDU and its preamble), which is limited to 5484 µs [38].
Electronics 2023, 12, x FOR PEER REVIEW 14 of 25
Figure 9. CCDF display on the spectrum analyzer.
Both CP and CCDF mode seings of the USRP instrument were made in comparison
with the analyzer. An application implemented in Phyton was used to extract channel
power and CCDF data from the analyzer (mean and peak power, crest factor and
percentage parameters above mean power for 10%, 1%, 0.1%, 0.01%). This application
runs on a computer so that the spectrum analyzer is accessed remotely and, depending
on the number of measurements and the time specified, returns the relevant information
resulting from the measurements.
Regarding the calculation of channel power, it should be noted that the vector signal
analyzer uses frequency domain evaluation based on the expression [26]:
𝐿 =10∙lg
∙
∙ 10
(1)
where L
ch
represents the channel power level expressed in dBm; B
ch
is the channel
bandwidth in Hz; B
NIF
is the intermediate frequency filter band expressed in Hz; P
i
represents the measured power value for pixel i.
In comparison, SDR platforms provide access to samples acquired in the time
domain, and the channel power calculation is performed according to the following
mathematical expression:
𝐿 =10∙lg
∙
[𝑆∙𝑆
] (2)
where L
ch
is the channel power reading in dBm, S
n
represents the current signal sample,
and S
n
′ is the complex conjugate of the current sample.
Accordingly, the channel power calculation is performed by averaging the product
of the current time domain sample and its complex conjugate (the number of samples is
between n
1
and n
2
) and applying the logarithmic operation to transform the final result
into dBm.
A critical component of an SDR platform is the analog-to-digital converter (ADC).
Converter characteristics that can limit the performance of SDR receivers are sample rate
and resolution. The sample rate limits the maximum bandwidth that can be processed in
real-time. For bandwidths of 160 MHz, an additional RF daughterboard is required, im-
plicitly increasing the price of the measurement system. The bandwidth upgrade has an
Figure 9. CCDF display on the spectrum analyzer.
Both CP and CCDF mode settings of the USRP instrument were made in comparison
with the analyzer. An application implemented in Phyton was used to extract channel
power and CCDF data from the analyzer (mean and peak power, crest factor and percentage
parameters above mean power for 10%, 1%, 0.1%, 0.01%). This application runs on a
computer so that the spectrum analyzer is accessed remotely and, depending on the
number of measurements and the time specified, returns the relevant information resulting
from the measurements.
Electronics 2023,12, 2920 14 of 24
Regarding the calculation of channel power, it should be noted that the vector signal
analyzer uses frequency domain evaluation based on the expression [26]:
Lch =10·lgBch
BNIF
·1
n2−n1
·∑n2
n110 Pi
10 (1)
where L
ch
represents the channel power level expressed in dBm; B
ch
is the channel band-
width in Hz; B
NIF
is the intermediate frequency filter band expressed in Hz; P
i
represents
the measured power value for pixel i.
In comparison, SDR platforms provide access to samples acquired in the time domain, and
the channel power calculation is performed according to the following mathematical expression:
Lch =10·lg1
n2−n1
·∑n2
n1Sn·S0
n(2)
where L
ch
is the channel power reading in dBm, S
n
represents the current signal sample,
and Sn0is the complex conjugate of the current sample.
Accordingly, the channel power calculation is performed by averaging the product
of the current time domain sample and its complex conjugate (the number of samples is
between n
1
and n
2
) and applying the logarithmic operation to transform the final result
into dBm.
A critical component of an SDR platform is the analog-to-digital converter (ADC).
Converter characteristics that can limit the performance of SDR receivers are sample rate
and resolution. The sample rate limits the maximum bandwidth that can be processed
in real-time. For bandwidths of 160 MHz, an additional RF daughterboard is required,
implicitly increasing the price of the measurement system. The bandwidth upgrade has
an additional cost compared to the basic version, SDR plus 40 MHz receiver board (price
11,000 Euro). If the basic bandwidth of the USRP platform is to be extended from 40 MHz
real-time to 160 MHz, the price increases by about 2000 Euro. For two 80 MHz TwinRX
channels (two highly synchronized receivers), the price of the basic version increases by
about 4600 Euro.
The other important parameter, resolution, affects the sensitivity (the level of the
received signal) and the dynamic range (the range of amplitudes that can be measured)
of the SDR receiver. To illustrate the importance of the converter resolution in terms of
bits, if it has one more bit, the dynamic range will improve by 6.02 dB (20lg2). Given a
sinusoidal signal with a certain amount of quantization noise, the maximum signal-to-noise
ratio (SNR) of the converter is defined by this relationship [39]:
SN R =(1.76 +6.02N)[dB](3)
where, N= number of bits of the analog-to-digital converter.
Since this measured signal-to-noise ratio is always lower than the theoretical one, the
effective bit number ENOB is used:
ENOB =SNRdB −1.76
6.02 [dB](4)
where SNR is the measured signal-to-noise ratio.
Another important aspect is that the measurement system needs high-performance
hardware in the processing part, especially for large bandwidths. For SDR boards that allow
these bandwidths, no packet loss is detected because the USRP platforms use different
types of FPGAs, ADCs and DACs with high dynamic range and the ability to transfer mea-
surements at high transfer rates based on 10 Gigabit Ethernet or PCIe Express connections.
The problem arises in the acquisition and data processing, and for this, the designer who
wants to implement such a system must consider a PC or server with powerful processing
capabilities according to the complexity of his application.
Electronics 2023,12, 2920 15 of 24
The SDR-based measurement system has additional features and limitations, which
are described below:
- Up to 160 MHz bandwidth;
- Wide frequency range: 10 MHz–6 GHz;
- RX noise figure (NF): 3–7 dB (10 MHz–6 GHz);
- Receive Gains Range: 0–31.5 dB;
-
FPGA Clock Rate: 200 MHz (below 1 GHz, it is necessary to reduce the daughterboard
clock rate to 20 MHz);
- High-performance 14-bit ADC (SNR = 86.04 dB);
- Streaming Bandwidth per Channel: 200 MS/s;
- Dual 10 Gigabit Ethernet—2x RX at 200 MSPs per channel;
- Digital signal processing (DSP) can take place in FPGA or/and a PC;
- PC DSP limitation in accordance with its performance.
3.4. System Validation in Realistic Wireless Environments
After parameterization, the designed system was applied to signals present in re-
alistic and challenging wireless environments. LTE-A, 5G and IEEE 802.11ax signals
emitted by a mobile communication device were chosen to demonstrate the system ca-
pabilities in performing CP and CCDF measurements for different channel types, band-
widths (20/40/80/100/160 MHz) and types of services carried in the channels. The
device used was a Huawei Mate xs2 mobile phone capable of supporting all three selected
communication standards.
For this validation phase, the measurements were performed over the air with the Om-
niLOG 70600 Aaronia manufactured antenna connected to the SDR RF channel. During the
measurements, the mobile device was placed in proximity (~40 cm) to the receiving antenna.
The channel center frequency assigned to the base station was 1750 MHz with 20 MHz
channel bandwidth for the LTE-A standard and 3580 MHz with 100 MHz channel band-
width for the 5G standard. The measurements were performed in a laboratory environment
without the exclusion of possible interfering signals.
The experimental CP calculation and CCDF statistical measurements targeted dif-
ferent services specific to the LTE-A and 5G standards. To include more diverse usage
profiles, the mobile device was connected to a base station and set to perform different
experimental scenarios:
•video call;
•voice over IP;
•file download;
•upload;
•video streaming.
WhatsApp, an application commonly used by smartphone users, was selected to
perform the video call and VoIP measurement scenarios; a large game was downloaded for
the download service; YouTube with 4K UHD 2160p fidelity was used for the streaming
scenario and a free speed-testing application was used for the upload service.
For 802.11ax-specific measurements, a Wi-Fi 6 router model TP-Link Archer AX10
was used for two experimental scenarios: downloading and uploading files on the same
Huawei mobile device. The Wi-Fi channels were configured in the 2.4 GHz Wi-Fi frequency
band and set to 40/80/160 MHz bandwidth. Other Wi-Fi networks were also visible
in the 2.4 GHz bandwidth, enabling the system validation in realistic electromagnetic
environments, with external sources of interference.
Electronics 2023,12, 2920 16 of 24
4. Results and Discussion
4.1. Comparison of the SDR Platform with the Spectrum Analyzer
In this section, the performance of the USRP X310 platform is presented in comparison
to the performance of the SA for an 802.11ax signal with 40 MHz channel bandwidth. Two
experimental scenarios were also carried out:
•
Both the SDR platform and the SA were set to run continuously for the same measure-
ment duration;
•
Measurements that have the same duration, but with an averaging performed each
0.5 s.
Figure 10 shows the comparative values of the channel power measured by both the
SA and the SDR as a boxplot of the channel power variation. In both continuous and 0.5 s
averaging cases there is a very narrow gap between the values measured by the SDR and
the SA. Most of the values in the whole band measured by the SDR show a very small
deviation from the median (<0.2 dB) compared to the reference analyzer.
Electronics 2023, 12, x FOR PEER REVIEW 17 of 25
Figure 10. Results from the SDR system and the SA: CP for an 802.11ax channel, 40 MHz channel
bandwidth.
The built-in CCDF function was used to measure the average signal strength and
peak values. These values are important for evaluating PAPR. This ratio is an important
parameter in wireless communication systems, as signals with a high PAPR require higher
power amplification, which can increase power consumption and cause distortion, reduc-
ing overall system efficiency.
Another parameter considered in measurements described by the CCDF curve is the
percentage of values above the mean for a given crest value. In wireless communications,
the CCDF curve can be used to analyze signal behavior under various channel conditions
such as filtering, compression, interference or fading. By calculating the probability that
the CCDF curve provides—10%, 1%, 0.1% and 0.01%—we can determine the probability
that the signal amplitude will exceed certain thresholds under these conditions.
Analyzing Figures 11 and 12, it can be seen that the mean values are distributed in a
larger interval in the case of the SA. Moreover, the absolute power levels are up to 0.4 dBm
higher for the system consisting of SDR and GNU Radio as compared to the SA. The dif-
ferent mean distribution profiles may be due to differences in the samples generated by
the ADCs of the two instruments or to differences in the software procedures. However,
the differences are not significant enough to change the general outcome of the measure-
ment. In the case of the peak values analysis, the differences were less than 0.8 dB.
Figure 10.
Results from the SDR system and the SA: CP for an 802.11ax channel, 40 MHz chan-
nel bandwidth.
The built-in CCDF function was used to measure the average signal strength and
peak values. These values are important for evaluating PAPR. This ratio is an important
parameter in wireless communication systems, as signals with a high PAPR require higher
power amplification, which can increase power consumption and cause distortion, reducing
overall system efficiency.
Another parameter considered in measurements described by the CCDF curve is the
percentage of values above the mean for a given crest value. In wireless communications,
the CCDF curve can be used to analyze signal behavior under various channel conditions
such as filtering, compression, interference or fading. By calculating the probability that
the CCDF curve provides—10%, 1%, 0.1% and 0.01%—we can determine the probability
that the signal amplitude will exceed certain thresholds under these conditions.
Analyzing Figures 11 and 12, it can be seen that the mean values are distributed in a
larger interval in the case of the SA. Moreover, the absolute power levels are up to 0.4 dBm
higher for the system consisting of SDR and GNU Radio as compared to the SA. The
different mean distribution profiles may be due to differences in the samples generated by
the ADCs of the two instruments or to differences in the software procedures. However, the
differences are not significant enough to change the general outcome of the measurement.
In the case of the peak values analysis, the differences were less than 0.8 dB.
Electronics 2023,12, 2920 17 of 24
Electronics 2023, 12, x FOR PEER REVIEW 18 of 25
Figure 11. CCDF mean and peak values as measured by the SDR platform and the SA for an 802.11ax
channel with 40 MHz bandwidth.
Figure 12. CCDF values from the SDR platform and SA: signal power above average for an 802.11ax
channel with 40 MHz bandwidth.
From analyzing the results, we observed that the measurements obtained from the
SDR platform are comparable to those of a high-performance SA, for both CP and CCDF
measurements. These results demonstrate that the SDR-based system can be used for
measurements outside the laboratory, where signal sources vary according to the service
provided.
4.2. LTE-A, 5G and IEEE 802.11ax Signals Measurements
Figure 13 shows the time evolution of the channel power during 30 s for the stream-
ing and VoIP test for LTE-A and 5G standards. In the case of 4G+, in the streaming scenario
represented by the blue color, it can be seen how the video is gradually loaded at different
Figure 11.
CCDF mean and peak values as measured by the SDR platform and the SA for an 802.11ax
channel with 40 MHz bandwidth.
Electronics 2023, 12, x FOR PEER REVIEW 18 of 25
Figure 11. CCDF mean and peak values as measured by the SDR platform and the SA for an 802.11ax
channel with 40 MHz bandwidth.
Figure 12. CCDF values from the SDR platform and SA: signal power above average for an 802.11ax
channel with 40 MHz bandwidth.
From analyzing the results, we observed that the measurements obtained from the
SDR platform are comparable to those of a high-performance SA, for both CP and CCDF
measurements. These results demonstrate that the SDR-based system can be used for
measurements outside the laboratory, where signal sources vary according to the service
provided.
4.2. LTE-A, 5G and IEEE 802.11ax Signals Measurements
Figure 13 shows the time evolution of the channel power during 30 s for the stream-
ing and VoIP test for LTE-A and 5G standards. In the case of 4G+, in the streaming scenario
represented by the blue color, it can be seen how the video is gradually loaded at different
Figure 12.
CCDF values from the SDR platform and SA: signal power above average for an 802.11ax
channel with 40 MHz bandwidth.
From analyzing the results, we observed that the measurements obtained from
the SDR platform are comparable to those of a high-performance SA, for both CP and
CCDF measurements. These results demonstrate that the SDR-based system can be used
for measurements outside the laboratory, where signal sources vary according to the
service provided.
4.2. LTE-A, 5G and IEEE 802.11ax Signals Measurements
Figure 13 shows the time evolution of the channel power during 30 s for the streaming
and VoIP test for LTE-A and 5G standards. In the case of 4G+, in the streaming scenario
represented by the blue color, it can be seen how the video is gradually loaded at different
timestamps, while in VoIP the signal is somewhat constant. The same can be observed
Electronics 2023,12, 2920 18 of 24
in 5G, wherein the VoIP experiment the signal is kept constant, while in streaming it is
observed how it is loaded at 4, 8, 24 and 27 s.
Electronics 2023, 12, x FOR PEER REVIEW 19 of 25
timestamps, while in VoIP the signal is somewhat constant. The same can be observed in
5G, wherein the VoIP experiment the signal is kept constant, while in streaming it is ob-
served how it is loaded at 4, 8, 24 and 27 s.
The differences observed in the CP measurements for the VoIP and streaming sce-
narios can be aributed to the application characteristics and capabilities of the two mo-
bile networks. In the case of the VoIP experiment, the signal remains relatively constant
also due to the real-time communication feature. The network maintains a constant chan-
nel strength to ensure stable and continuous voice transmission. On the other hand, in the
case of streaming, the channel power is not evenly distributed, the increase in channel
power is best observed in the case of 5G when the video load is high. Differences are ob-
served for streaming in LTE-A versus 5G due to the adaptation of data transmission to the
available bandwidth. By using the SDR-based measurement system, we can highlight the
different behavior of the LTE-A and 5G signals with respect to CP. These peculiarities are
due to the different channel bandwidths (20 MHz vs. 100 MHz) and modulation schemes
enabled by the two communication standards. Modulation schemes have varying power
requirements due to their inherent characteristics and the trade-offs they make between
data rate, spectral efficiency, and power efficiency [40]. Besides the bandwidth extension,
to achieve higher data rates, 5G employs more complex modulation and coding tech-
niques, which require additional power to accurately transmit and receive the signal.
(a) (b)
Figure 13. Time evolution for CP in LTE-A (a) and 5G (b). VoIP and streaming use cases.
In order to study and test the measurement system to its maximum, several experi-
ments were performed on the IEEE 802.11ax network. The scenario consisted of configur-
ing the router successively to 40/80/160 MHz and downloading a file for each bandwidth.
Figure 14 shows the power measurements over time for the three bandwidths. For the
scenario described, the channel performance was similar when using 40 MHz and 80 MHz
bandwidths with slightly higher CP values being measured for 80 MHz. The increase in
CP with channel bandwidth is more visible when a 160 MHz bandwidth was used and
the measured channel power was visibly higher. This can be explained by the fact that
signals with larger bandwidths tend to exhibit higher channel power requirements be-
cause, in a wider bandwidth, there are more available frequencies or subcarriers that need
to be powered to transmit the signal. It is important to note that while larger bandwidths
generally require higher power, advancements in technology and signal processing tech-
niques continue to improve the power efficiency of wideband communication systems.
These advancements aim to achieve higher data rates and improved performance while
minimizing power consumption [41].
Figure 13. Time evolution for CP in LTE-A (a) and 5G (b). VoIP and streaming use cases.
The differences observed in the CP measurements for the VoIP and streaming scenarios
can be attributed to the application characteristics and capabilities of the two mobile
networks. In the case of the VoIP experiment, the signal remains relatively constant also
due to the real-time communication feature. The network maintains a constant channel
strength to ensure stable and continuous voice transmission. On the other hand, in the case
of streaming, the channel power is not evenly distributed, the increase in channel power is
best observed in the case of 5G when the video load is high. Differences are observed for
streaming in LTE-A versus 5G due to the adaptation of data transmission to the available
bandwidth. By using the SDR-based measurement system, we can highlight the different
behavior of the LTE-A and 5G signals with respect to CP. These peculiarities are due to the
different channel bandwidths (20 MHz vs. 100 MHz) and modulation schemes enabled by
the two communication standards. Modulation schemes have varying power requirements
due to their inherent characteristics and the trade-offs they make between data rate, spectral
efficiency, and power efficiency [
40
]. Besides the bandwidth extension, to achieve higher
data rates, 5G employs more complex modulation and coding techniques, which require
additional power to accurately transmit and receive the signal.
In order to study and test the measurement system to its maximum, several experi-
ments were performed on the IEEE 802.11ax network. The scenario consisted of configuring
the router successively to 40/80/160 MHz and downloading a file for each bandwidth.
Figure 14 shows the power measurements over time for the three bandwidths. For the
scenario described, the channel performance was similar when using 40 MHz and 80 MHz
bandwidths with slightly higher CP values being measured for 80 MHz. The increase in
CP with channel bandwidth is more visible when a 160 MHz bandwidth was used and
the measured channel power was visibly higher. This can be explained by the fact that
signals with larger bandwidths tend to exhibit higher channel power requirements because,
in a wider bandwidth, there are more available frequencies or subcarriers that need to
be powered to transmit the signal. It is important to note that while larger bandwidths
generally require higher power, advancements in technology and signal processing tech-
niques continue to improve the power efficiency of wideband communication systems.
These advancements aim to achieve higher data rates and improved performance while
minimizing power consumption [41].
The next scenario aims to test all services for 4G+ and 5G networks. The measurements
analyzed comparatively refer to the channel power, i.e., the average (mean) and peak power
extracted from the CCDF curve. Figure 15 illustrates the measurement results and we can
observe significant differences in download and upload services between the two standards.
The higher values of CP, average power and peak power observed for the LTE-A standard
Electronics 2023,12, 2920 19 of 24
compared to the 5G network indicate that LTE-A faces more challenges in achieving efficient
data throughput performance as compared to the 5G system. This observation is valid for
data hunger applications like video calls, file upload and file download. Along with CP
measurement, the use of the CCDF measurement gives direct access to the main statistical
indicators of mean and peak power values. The temporal variability of the mean and peak
power together with the obtained CCDF curve can be used to derive specific usage profiles
depending on the enabled type of application, information that is extremely useful for
communication reconnaissance missions.
Electronics 2023, 12, x FOR PEER REVIEW 20 of 25
Figure 14. Time evolution of CP in 802.11ax—file download.
The next scenario aims to test all services for 4G+ and 5G networks. The measure-
ments analyzed comparatively refer to the channel power, i.e., the average (mean) and
peak power extracted from the CCDF curve. Figure 15 illustrates the measurement results
and we can observe significant differences in download and upload services between the
two standards. The higher values of CP, average power and peak power observed for the
LTE-A standard compared to the 5G network indicate that LTE-A faces more challenges
in achieving efficient data throughput performance as compared to the 5G system. This
observation is valid for data hunger applications like video calls, file upload and file
download. Along with CP measurement, the use of the CCDF measurement gives direct
access to the main statistical indicators of mean and peak power values. The temporal
variability of the mean and peak power together with the obtained CCDF curve can be
used to derive specific usage profiles depending on the enabled type of application, infor-
mation that is extremely useful for communication reconnaissance missions.
The difference between the peak and the mean signal value is the PAPR of the signal.
One can observe that even though both LTE-A and 5G signals use OFDM for accessing
the radio environment, our results indicate higher PAPR values for the 5G network in all
usage scenarios. This was theoretically expected and can be explained by the differences
in their underlying modulation schemes and transmission techniques. 5G networks often
utilize more complex modulation schemes, such as higher-order Quadrature Amplitude
Modulation (QAM), compared to LTE-A, which can result in a higher PAPR. In addition,
5G introduces new waveform technologies, such as filtered OFDM (f-OFDM) which aims
to further improve spectral efficiency and mitigate interference. This waveform often ex-
hibits higher PAPR than the traditional cyclic prefix-based OFDM used in LTE-A net-
works.
Figure 14. Time evolution of CP in 802.11ax—file download.
Electronics 2023, 12, x FOR PEER REVIEW 21 of 25
(a) (b)
Figure 15. Channel power boxplots (a), Mean and Peak power boxplots (b) for the tested 4G+ and
5G application service.
The values provided with the CCDF curve are very important for designing trans-
miers that are not built for the highest power that could ever occur, but for a power that
has a certain probability of occurring (e.g., 0.01%). These values are also shown in Figure
16, which shows the tail CCDF curves for all scenarios considered. It can be seen that in
the case of 5G measurements, the average power for 0.01% is exceeded by 12 dB compared
to 4G+ where the average power for the same probability is exceeded by 10 dB. The shape
of the CCDF curve is related to the power distribution of the signal. Signals with wider
bandwidths and higher data rates tend to have more complex power distributions. Some
signal types, such as Gaussian signals, have a well-defined CCDF that follows a known
mathematical function. The more the CCDF curve is shifted to the right, the more stochas-
tic the distribution, an effect that is more visible in the case of the 5G measured signals as
compared to the LTE-A signals.
(a) (b)
Figure 16. Signal power above average power for 4G+ (a) and 5G (b) application service tests—
boxplot representation.
The following figures (Figures 17 and 18) show the results of the statistical CCDF and
channel power measurements for download and upload scenarios at different band-
widths specific to the 802.11ax standard. The graphical results show the higher channel
Figure 15.
Channel power boxplots (
a
), Mean and Peak power boxplots (
b
) for the tested 4G+ and
5G application service.
The difference between the peak and the mean signal value is the PAPR of the signal.
One can observe that even though both LTE-A and 5G signals use OFDM for accessing
the radio environment, our results indicate higher PAPR values for the 5G network in all
usage scenarios. This was theoretically expected and can be explained by the differences
in their underlying modulation schemes and transmission techniques. 5G networks often
utilize more complex modulation schemes, such as higher-order Quadrature Amplitude
Modulation (QAM), compared to LTE-A, which can result in a higher PAPR. In addition,
5G introduces new waveform technologies, such as filtered OFDM (f-OFDM) which aims to
further improve spectral efficiency and mitigate interference. This waveform often exhibits
higher PAPR than the traditional cyclic prefix-based OFDM used in LTE-A networks.
Electronics 2023,12, 2920 20 of 24
The values provided with the CCDF curve are very important for designing transmit-
ters that are not built for the highest power that could ever occur, but for a power that has
a certain probability of occurring (e.g., 0.01%). These values are also shown in Figure 16,
which shows the tail CCDF curves for all scenarios considered. It can be seen that in the
case of 5G measurements, the average power for 0.01% is exceeded by 12 dB compared to
4G+ where the average power for the same probability is exceeded by 10 dB. The shape
of the CCDF curve is related to the power distribution of the signal. Signals with wider
bandwidths and higher data rates tend to have more complex power distributions. Some
signal types, such as Gaussian signals, have a well-defined CCDF that follows a known
mathematical function. The more the CCDF curve is shifted to the right, the more stochastic
the distribution, an effect that is more visible in the case of the 5G measured signals as
compared to the LTE-A signals.
Electronics 2023, 12, x FOR PEER REVIEW 21 of 25
(a) (b)
Figure 15. Channel power boxplots (a), Mean and Peak power boxplots (b) for the tested 4G+ and
5G application service.
The values provided with the CCDF curve are very important for designing trans-
miers that are not built for the highest power that could ever occur, but for a power that
has a certain probability of occurring (e.g., 0.01%). These values are also shown in Figure
16, which shows the tail CCDF curves for all scenarios considered. It can be seen that in
the case of 5G measurements, the average power for 0.01% is exceeded by 12 dB compared
to 4G+ where the average power for the same probability is exceeded by 10 dB. The shape
of the CCDF curve is related to the power distribution of the signal. Signals with wider
bandwidths and higher data rates tend to have more complex power distributions. Some
signal types, such as Gaussian signals, have a well-defined CCDF that follows a known
mathematical function. The more the CCDF curve is shifted to the right, the more stochas-
tic the distribution, an effect that is more visible in the case of the 5G measured signals as
compared to the LTE-A signals.
(a) (b)
Figure 16. Signal power above average power for 4G+ (a) and 5G (b) application service tests—
boxplot representation.
The following figures (Figures 17 and 18) show the results of the statistical CCDF and
channel power measurements for download and upload scenarios at different band-
widths specific to the 802.11ax standard. The graphical results show the higher channel
Figure 16.
Signal power above average power for 4G+ (
a
) and 5G (
b
) application service tests—
boxplot representation.
The following figures (Figures 17 and 18) show the results of the statistical CCDF
and channel power measurements for download and upload scenarios at different band-
widths specific to the 802.11ax standard. The graphical results show the higher channel
power level for download at higher bandwidths (most noticeable is the 160 MHz band-
width), while the upload area shows a relatively constant average/peak power for all three
bandwidths analyzed.
In Figure 17, we can observe a slight increase in the bandwidth for the peak values of
the 802.11ax measured signal. The wider bandwidth increases the number of subcarriers
used in the OFDM modulation scheme, contributing to higher PAPR as the peaks of
individual subcarriers may align and result in larger power excursions.
Figure 18 shows the CCDF values referring to the peak signal amplitudes for the
three bandwidths, both download and upload 802.11ax channels. For the download
measurements, the router performance was monitored and for the upload, the mobile
phone was monitored. It can be seen that in both situations the results present good
theoretical agreement for OFDM systems, with values of less than 5 dB for 10% and no
more than 12 dB for 0.01%.
Electronics 2023,12, 2920 21 of 24
Electronics 2023, 12, x FOR PEER REVIEW 22 of 25
power level for download at higher bandwidths (most noticeable is the 160 MHz band-
width), while the upload area shows a relatively constant average/peak power for all three
bandwidths analyzed.
(a) (b)
Figure 17. CP (a) and mean and peak signal power (b) for 802.11ax download and upload—boxplot
representation.
In Figure 17, we can observe a slight increase in the bandwidth for the peak values of
the 802.11ax measured signal. The wider bandwidth increases the number of subcarriers
used in the OFDM modulation scheme, contributing to higher PAPR as the peaks of indi-
vidual subcarriers may align and result in larger power excursions.
Figure 18 shows the CCDF values referring to the peak signal amplitudes for the three
bandwidths, both download and upload 802.11ax channels. For the download measure-
ments, the router performance was monitored and for the upload, the mobile phone was
monitored. It can be seen that in both situations the results present good theoretical agree-
ment for OFDM systems, with values of less than 5 dB for 10% and no more than 12 dB
for 0.01%.
(a) (b)
Figure 18. Signal power above mean power for 802.11ax download (a) and upload (b)—boxplot
represenation.
5. Conclusions
With the advent of modern communications technologies, the complexity of moni-
toring the electromagnetic spectrum has increased. The variability of signals in time,
Figure 17.
CP (
a
) and mean and peak signal power (
b
) for 802.11ax download and upload—
boxplot representation.
Electronics 2023, 12, x FOR PEER REVIEW 22 of 25
power level for download at higher bandwidths (most noticeable is the 160 MHz band-
width), while the upload area shows a relatively constant average/peak power for all three
bandwidths analyzed.
(a) (b)
Figure 17. CP (a) and mean and peak signal power (b) for 802.11ax download and upload—boxplot
representation.
In Figure 17, we can observe a slight increase in the bandwidth for the peak values of
the 802.11ax measured signal. The wider bandwidth increases the number of subcarriers
used in the OFDM modulation scheme, contributing to higher PAPR as the peaks of indi-
vidual subcarriers may align and result in larger power excursions.
Figure 18 shows the CCDF values referring to the peak signal amplitudes for the three
bandwidths, both download and upload 802.11ax channels. For the download measure-
ments, the router performance was monitored and for the upload, the mobile phone was
monitored. It can be seen that in both situations the results present good theoretical agree-
ment for OFDM systems, with values of less than 5 dB for 10% and no more than 12 dB
for 0.01%.
(a) (b)
Figure 18. Signal power above mean power for 802.11ax download (a) and upload (b)—boxplot
represenation.
5. Conclusions
With the advent of modern communications technologies, the complexity of moni-
toring the electromagnetic spectrum has increased. The variability of signals in time,
Figure 18.
Signal power above mean power for 802.11ax download (
a
) and upload (
b
)
—boxplot represenation.
5. Conclusions
With the advent of modern communications technologies, the complexity of mon-
itoring the electromagnetic spectrum has increased. The variability of signals in time,
space, frequency and amplitude, as well as the evolution of wireless communications,
require advanced spectrum measurements to evaluate and optimize the performance of
communications networks. Although traditional SAs are commonly used to monitor the
electromagnetic spectrum, they are associated with significant cost and reduced scalability.
In this context, SDR platforms offer an efficient and cost-effective alternative to implement
customized and reconfigurable measurement applications.
This paper focuses on practical aspects and aims to provide a cost-effective solution
for wideband signal monitoring using SDR platforms. These platforms offer several advan-
tages, including affordable acquisition costs, ease of implementation of a wideband signal
acquisition system, flexibility and versatility for developing custom solutions. Users can al-
ways adapt and customize applications to meet specific monitoring requirements/scenarios.
However, it is important to consider some of the drawbacks that have been identified for
SDR platforms. One of these drawbacks is the need for parameterization (adjustment of the
measurement system using a reference device) to ensure accurate results. In addition, the
Electronics 2023,12, 2920 22 of 24
cost of the SDR platform may increase depending on the need to add an RF daughterboard
to accommodate the higher bandwidth of the channel being monitored.
The novelty of the proposed approach lies in the design of an SDR-based measurement
system that performs real-time spectrum monitoring on wide frequency bands (up to
160 MHz) and enables both CP and CCDF measurement functions. The two measurement
functions are software implemented in GNU Radio by designing customized blocks and
integrated into a graphical user interface. CP measurements play a crucial role in assessing
signal quality, interference analysis, system optimization and spectrum management. The
use of the CCDF measurement function represents a valuable tool for efficiently capturing
fast signal variations and direct retrieval of main statistical parameters. By plotting the
CCDF curves, we can observe and analyze the differences in signal power levels across
various probability levels. This provides a deeper understanding of signal behavior and
helps optimize the performance of communication systems in dynamic and challenging
wireless environments.
The proposed monitoring system implemented on the USRP X310 platform, was
tested under laboratory conditions and its performances were compared to those of a
high-performance SA. After validation, the SDR measurement system was used to measure
both CP and CCDF of signals present in real electromagnetic environments. The signals
specific to LTE-A, 5G and the IEEE 802.11ax standards emitted for different usage profiles
were considered, highlighting the system’s potential to capture and analyze a wide range of
signals. By performing the measurement campaign, we have demonstrated the capabilities
of the measurement system to perform real-time measurements on broadband channels
(up to 160 MHz for IEEE 802.11ax). Moreover, we demonstrated the usability of CP and
CCDF measurement functions in highlighting differences in the usage profiles between the
three modern communication standards.
Implementing CP and CCDF applications on the USRP X310 platform and compar-
ing the results with a high-performance spectrum analyzer (SA) represents a significant
achievement. The fact that the results obtained by the SDR system were comparable to
those of a state-of-the-art SA highlights the capabilities and potential of SDR platforms
for spectrum monitoring and analysis. Successfully performing these measurements in
real-time for broadband channels, where few instruments are capable, further validates the
effectiveness and versatility of the proposed monitoring system.
Author Contributions:
Conceptualization, M.
S
,
., A.S. and P.B.; Methodology, M.
S
,
., A.S. and P.B.;
Software, E.
S
,
. and M.
S
,
.; Validation, M.
S
,
., A.S. and E.
S
,
.; Formal analysis, P.B.; Investigation, M.
S
,
.
and E.
S
,
.; Writing—original draft preparation, M.
S
,
.; Writing—review and editing, M.
S
,
., A.S. and E.
S
,
.;
Visualization, P.B. and A.S.; Supervision, A.S. and P.B.; Project administration, M.
S
,
. and A.S. All
authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Balapuwaduge, I.A.; Li, F.Y. Cellular Networks: An Evolution from 1G to 4G. In Encyclopedia of Wireless Networks; Springer: Cham,
Switzerland, 2018. [CrossRef]
2. Mahmud, H. Cellular Mobile Technologies (1G to 5G) and Massive MIMO. Int. J. Sci. Res. 2019,8, 929–937. [CrossRef]
3.
Chataut, R.; Akl, R. Massive MIMO Systems for 5G and beyond Networks—Overview, Recent Trends, Challenges, and Future
Research Direction. Sensors 2020,20, 2753. [CrossRef]
4.
Jimaa, S.A.; Chai, K.K.; Chen, Y.; Alfadhl, Y. LTE-A an overview and future research areas. In Proceedings of the IEEE 7th
International Conference on Wireless and Mobile Computing, Networking and Communications, Washington, DC, USA, 10–12
October 2011; pp. 395–399. [CrossRef]
5.
Frank, I.; Ike, D.U.; Jimi, A.; Kayode, O. Multiple-Access Technology of Choice in 3GPP LTE. Indones. J. Electr. Eng. Inform.
2013
,1,
69–77. [CrossRef]
6. Vaigandla, K.K.; Venu, D.N. BER, SNR and PAPR Analysis of OFDMA and SC-FDMA. GIS Sci. J. ISSN 2021,8, 970–977.
7.
Al-Fuhaidy, F.A.; Al-Sofy, K.A.; Alkamali, F.S. Discrete Sine Transform Based OFDMA System for Wireless Broadband Communi-
cations. AASCIT Commun. 2019,6, 13–21.
Electronics 2023,12, 2920 23 of 24
8.
Jawhar, Y.A.; Audah, L.; Taher, M.A.; Ramli, K.N.; Shah, N.S.; Musa, M.; Ahmed, M.S. A Review of Partial Transmit Sequence for
PAPR Reduction in the OFDM Systems. IEEE Access 2019,7, 18021–18041. [CrossRef]
9.
Saghir, A.; Bhatti, B.A.; Yawar, S.J.; Hassan, A.; Siddiqui, F.M.; Shakeel, A. Analysis of LTE-Advanced Uplink System for the Effect
of Carrier Frequency Offset and Ratio of Peak to Average Power. J. Xi’an Shiyou Univ. 2022,8, 242–252.
10.
Zayani, R.; Shaiek, H.; Roviras, D. PAPR-Aware Massive MIMO-OFDM Downlink. IEEE Access
2019
,7, 25474–25484. [CrossRef]
11.
Xing, Z.; Liu, K.; Rajasekaran, A.S.; Yanikomeroglu, H.; Liu, Y. A Hybrid Companding and Clipping Scheme for PAPR Reduction
in OFDM Systems. IEEE Access 2021,9, 61565–61576. [CrossRef]
12.
Gupta, A.; Jha, R.K. A Survey of 5G Network: Architecture and Emerging Technologies. IEEE Access
2015
,3, 1206–1232. [CrossRef]
13.
Siddiqi, M.A.; Yu, H.; Joung, J. 5G Ultra-Reliable Low-Latency Communication Implementation Challenges and Operational
Issues with IoT Devices. Electronics 2019,8, 981. [CrossRef]
14. Mane, S. 5G Communications & Networks. Int. J. All Res. Educ. Sci. Methods 2022,10, 261–268.
15.
Franci, D.; Coltellacci, S.; Grillo, E.; Pavoncello, S.; Aureli, T.; Cintoli, R.; Migliore, M.D. Experimental Procedure for Fifth
Generation (5G) Electromagnetic Field (EMF) Measurement and Maximum Power Extrapolation for Human Exposure Assessment.
Environments 2020,7, 22. [CrossRef]
16.
Mazar (Madjar), H. RF Human Hazards: Implementing 5G for Good: Does EMF Matter? In Proceedings of the ITU Regional
Forum for Europe: 5G Strategies, Policies and Implementation, Warsaw, Poland, 22–23 October 2020.
17.
Muhammad, S.; Zhao, J.; Refai, H.H. An Empirical Analysis of IEEE 802.11ax. In Proceedings of the 2020 International Conference
on Communications, Signal Processing, and their Applications, Sharjah, United Arab Emirates, 16–18 March 2021; pp. 1–6.
[CrossRef]
18.
Tutelian, S.; Bankov, D.; Shmelkin, D.; Khorov, E.M. IEEE 802.11ax OFDMA Resource Allocation with Frequency-Selective Fading.
Sensors 2021,21, 6099. [CrossRef] [PubMed]
19.
Yang, H.; Deng, D.; Chen, K. Performance Analysis of IEEE 802.11ax UL OFDMA-Based Random Access Mechanism. In
Proceedings of the GLOBECOM 2017—2017 IEEE Global Communications Conference, Singapore, 4–8 December 2017; pp. 1–6.
[CrossRef]
20.
ITU-R SM.2039-2013 Spectrum Monitoring Evolution. Available online: https://www.itu.int/rec/R-REC-SM.2039-0-201308-I/en
(accessed on 23 May 2023).
21.
Bechet, P.; Miclaus, S.G.; Bechet, A.C. Improving the Accuracy of Exposure Assessment to Stochastic-like Radiofrequency Signals.
IEEE Trans. Electromagn. Compat. 2012,54, 1169–1177. [CrossRef]
22.
Pfammatter, D.; Giustiniano, D.; Lenders, V. A Software-Defined Sensor Architecture for Large-Scale Wideband Spectrum
Monitoring. In Proceedings of the 14th International Conference on Information Processing in Sensor Networks, Seattle, WA,
USA, 14–16 April 2015; pp. 71–82. [CrossRef]
23.
Perotoni, M.B.; Ferreira, L.; Maniçoba, A. Low-Cost Measurement of Electromagnetic Leakeage in Domestic Appliances Using
Software-Defined Radios. Rev. Bras. Ensino Física 2022,44, e20220009. [CrossRef]
24.
Sârbu, A.; Migliore, M.D.; Sorecau, E.; Sorecau, M.; Miclaus, S.G.; Bechet, P. SDR-Enabled Multichannel Real-Time Measurement
System for In Situ EMF Exposure Evaluation. Electronics 2022,11, 2670. [CrossRef]
25.
Sorecau, M.; Sorecau, E.; Sârbu, A.; Helbet, R.; Bechet, P. Electromagnetic spectrum monitoring of LTE channels based on SDR
portable sensor: Preliminary analysis. IOP Conf. Ser. Mater. Sci. Eng. 2022,1254, 012009. [CrossRef]
26.
Helbet, R.; Bechet, P.; Monda, V.; Miclaus, S.; Bouleanu, I. Low-cost sensor based on SDR platform for TETRA signals monitoring.
Sensors 2021,21, 3160. [CrossRef]
27.
Cui, X.; Liu, K.; Liu, Y. Novel Linear Companding Transform Design Based on Linear Curve Fitting for PAPR Reduction in OFDM
Systems. IEEE Commun. Lett. 2021,25, 3604–3608. [CrossRef]
28.
Characterizing Digitally Modulated Signals with CCDF Curves, Application Note, Agilent Technologies Literature. Available
online: https://www.testunlimited.com/pdf/an/5968-6875E.pdf (accessed on 18 October 2022).
29.
Sârbu, A.; Bechet, A.C.; Balan, T.C.; Robu, D.; Bechet, P.; Miclaus, S.G. Using CCDF Statistics for Characterizing the Radiated
Power Dynamics in the Near Field of a Mobile Phone Operating in 3G+ and 4G+ Communication Standards. Measurement
2019
,
134, 874–887. [CrossRef]
30.
Zakaria, H.; Sarhan, A.E. Performance Analysis of Multicarrier Code Division Multiple Access Transmission Techniques using
BPSK Modulation. Int. J. Eng. Res. Technol. 2015,4, 1266–1273.
31.
Ronte, S.; Arakawa, S. Development of Multi-Channel APD Measuring Apparatus, Anritsu Technical Review. Available
online: https://dl.cdn-anritsu.com/ja-jp/test-measurement/reffiles/About-Anritsu/R_D/Technical/E-20/20_06.pdf (accessed
on 18 October 2022).
32.
Han, S.H.; Lee, J.H. An Overview of Peak-To-Average Power Ratio Reduction Techniques FOR Multicarrier Transmission. IEEE
Wirel. Commun. 2005,12, 56–65. [CrossRef]
33.
Carcangiu, S.; Fanni, A.; Montisci, A. A Closed Form Selected Mapping Algorithm for PAPR Reduction in OFDM Multicarrier
Transmission. Energies 2022,15, 1938. [CrossRef]
34.
Sengar, S.; Bhattacharya, P.P. Performance Improvement in OFDM System by PAPR Reduction. Signal Image Process. Int. J.
2012
,3,
157–169. [CrossRef]
35.
Angrisani, L.; Langella, A.; Vadursi, M. New Digital Signal-Processing Approaches for Measuring Power CCDF Curves. In
Proceedings of the 13th Symposium on Measurements for Research and Industrial Applications, Taormina, Italy, 12–15 June 2004.
Electronics 2023,12, 2920 24 of 24
36.
Main Features of the USRP X310 Platform. Available online: https://www.ettus.com/all-products/x310-kit/ (accessed on
18 May 2023).
37. Available online: https://wiki.gnuradio.org/index.php?title=Main_Page (accessed on 18 May 2023).
38.
Sharon, O.; Alpert, Y. Scheduling Strategies and Throughput Optimization for the Uplink for IEEE 802.11ax and IEEE 802.11ac
Based Networks. Wirel. Sens. Netw. 2017,9, 250–273. [CrossRef]
39.
Grayver, E. Implementing Software Defined Radio; Springer: Berlin/Heidelberg, Germany, 2013; ISBN 978-1-4419-9332-8. [CrossRef]
40.
Brown, C.; Vigneron, P.J. Power and Spectrally Efficient Communications: A Comparison of Linear and Nonlinear Modulation Schemes;
Dianat, S.A., Zoltowski, M.D., Eds.; SPIE: Bellingham, WA, USA, 2013. [CrossRef]
41.
Abdu, T.S.; Kisseleff, S.; Lagunas, E.; Chatzinotas, S.; Ottersten, B. Energy Efficient Sparse Precoding Design for Satellite
Communication System. In Proceedings of the IEEE 96th Vehicular Technology Conference, London, UK, 26–29 September 2022.
[CrossRef]
Disclaimer/Publisher’s Note:
The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.
