2017 11th European Conference on Antennas and Propagation (EUCAP)

UAV-based Antenna Measurements: Scan Strategies

Fabio Paonessa1, Giuseppe Virone1, Pietro Bolli2, Andrea M. Lingua1,3
1
Istituto di Elettronica e di Ingegneria dell’Informazione e delle Telecomunicazioni (IEIIT), Consiglio Nazionale delle Ricerche
(CNR), Torino, Italy, [Link]@[Link]
2
Osservatorio Astrofisico di Arcetri (OAA), Istituto Nazionale di Astrofisica (INAF), Firenze, Italy
3
Dipartimento di Ingegneria dell’Ambiente, del Territorio e delle Infrastrutture (DIATI), Politecnico di Torino, Torino, Italy

Abstract—In the recent years, the authors developed a flight duration, which already represents a limiting factor for
system for the characterization of the radiation pattern of VHF battery-powered UAVs.
and UHF antennas by means of a test-source mounted on a Exploiting the recent developments of the UAV
micro Unmanned Aerial Vehicle (UAV). So far, the adopted
technology, more complex flight strategies with variable
scan strategies typically consisted of two orthogonal straight
paths at constant height from ground to obtain E- or H-plane height can be planned in order to both optimize the above
cuts. In this paper, more complex scan strategies are presented mentioned parameters and obtain information over a planar
to perform two-dimensional pattern measurements over planar or curved surface rather than linear cuts. In this case, the
and curved surfaces. purpose is to perform several scans with a time-effective
Index Terms—antenna measurements, unmanned aerial strategy and combine the data, obtaining a two-dimensional
vehicles, UAV. representation of the AUT pattern.
The next section presents some of the scan strategies
I. INTRODUCTION recently employed by the authors to perform two-
dimensional pattern measurements. Such examples of
In the recent years, UAV-mounted radio sources have implemented flight trajectories were measured using the
been exploited by researchers to perform in-situ antenna onboard GPS unit. The corresponding measured antenna
pattern measurements [1]–[5]. In particular, the authors patterns will be shown at the conference.
developed a system to characterize the radiation pattern of
VHF and UHF ground-based antennas for radio astronomy
observations (see Fig. 1). This UAV-based far-field test-
source has been successfully used to characterize both single
elements [6]–[8] and arrays [9] [10] in the frequency range
from 45 MHz to 650 MHz.
The typical measurement strategy consists of a rectilinear
constant-height flight trajectory lying in the E- or H-plane of
the antenna radiation pattern (Fig. 2). Moreover, since the
flying test source is linearly-polarized [11], co-/cross-polar
measurements can be performed by setting the proper UAV RF generator
orientation in the horizontal plane (yaw angle) [12]. The craft
can also hover at the zenith of the Antenna Under Test dipole
(AUT) and rotate slowly around the vertical axis to
investigate the AUT polarization properties [8]. Fig. 1. The UAV-mounted radio source, which consists of an RF
Such linear/punctual scan strategies only provide partial generator and a dipole antenna.
information on the AUT radiation pattern. However, they are
simple and reliable from the point of view of the flight
dynamics and accuracy of execution. The corresponding 100
AUT pattern data are obtained by post-processing the
z (m)

50
received power at the AUT port according to [13].
It should be noted that, in a rectilinear flight, the distance 0
200
between the test-source and the AUT becomes larger for
increasing values of zenith angle (scan angle from zenith). 100
AUT
The increasing path loss with respect to the zenith angle will x (m)
0
-200
slightly affect the required receiver dynamic range. -100
0
-100
Furthermore, the long flying distance required to reach high -200 100 y (m)
200
zenith angles generally leads to a quite long (non-optimum)
Fig. 2. Typical flight trajectory to obtain E- or H-plane cuts. The extension
is ±180 m at 100 m height, i.e. ±60° of zenith angle variation. Take-off and
landing paths not shown.

978-88-907018-7-0/17/$31.00 ©2017 IEEE 1303 #1570318319

 2017 11th European Conference on Antennas and Propagation (EUCAP)

II. RASTER SCAN STRATEGIES

30
A. Cartesian raster
The Cartesian raster in Fig. 3a consists of several
rectilinear parallel flight paths performed at a constant height 20
and with constant UAV orientation. Such orientation is

z (m)
usually parallel or perpendicular to the direction of the 10
segments, providing two orthogonal field components.
However, it requires quite large power consumption and
flight time because of the many veers the craft has to 0
perform, reducing the average speed. 20
Such a flying strategy has already been used in the far- 10
field of an array in order to map specific details of the 0 AUT
x (m)
-10 -20
radiation patterns such as the main beam. Moreover, it has 0 -10
-20 10
also been used as a near-field scan strategy: the experimental 20 y (m)
case of Fig. 3a has been performed at 25 m height inside a
(a)
square area with extension of ±20 m both along x and y, i.e.
39° of zenith angle extension. The spacing between adjacent
paths is 0.75 m, which gives a total flight length of about 150
2400 m. As a first step, this strategy will be used to verify
amplitude-only near-field simulation data. Further research 100
work is being carried out in order to obtain the phase
z (m)
measurement as well. In this way, the far-field pattern could 50
be derived from such a near-field distribution. This evolution
would enable the characterization of very large installations 0
for which the required far-field distance would prevent a 150
100
small UAV to be exploited. The concept of UAVs as near- 50 AUT
field positioners has already been reported in [14]. 0
x (m) -50 -150
-50 -100
-100 0
B. Radial raster -150 100 50
150
y (m)
The radial raster shown in Fig. 3b develops on a planar
(b)
surface. However, it can be implemented with curved
segments as well. It consists of several rectilinear constant-
height paths having radial direction with respect to the 100
vertical axis. The UAV yaw angle is kept constant within a
75
single segment, i.e. parallel or perpendicular to the radial
direction. In this way, either the θ- or φ- component of the
z (m)

50
radiation pattern can be detected, respectively. It should be
noted that additional post-processing is still required in order 25
to obtain the two components of the AUT pattern from the 0
measured data. 100
The flight efficiency, in terms of covered area vs. time, is 50
AUT
higher than the Cartesian raster. However, the sampling 0
density is higher close to the zenith and reduces as the x (m)
-50 -100
-50
distance increases. In the case of Fig. 3b, the flight height is -100 50
0
100
100 m and the circular area given by the flight envelope has y (m)
a radius of 150 m, which corresponds to approximately 56° (c)
of zenith angle. The azimuthal angular pitch is 22.5° and the
total flight length is 3400 m.
Fig. 3. Measured flight trajectories of (a) a Cartesian raster, (b) a radial
raster and (c) an azimuthal 3D raster. Take-off and landing paths not
C. Azimuthal 3D raster shown.
The azimuthal raster (Fig. 3c) consists of several circular
concentric paths that are performed at different heights in raster. Moreover, although the flight envelope is three-
order to maintain a constant distance from the AUT. This dimensional, a single circular path has a constant height,
feature will lead to both a constant path loss and minimum which provides high flight accuracy.
flight duration. It should be noted that the azimuthal Differently from [15], the UAV orientation (yaw) should
sampling density is more uniform with respect to the radial be set as parallel (perpendicular) to the flight direction (speed

1304 #1570318319
 2017 11th European Conference on Antennas and Propagation (EUCAP)

vector) in order to detect either the φ- (θ-) component. It frequency source,” Experimental Astronomy, vol. 39, issue 2, pp.
405-421, June 2015.
should be noted that the UAV horizontal orientation
[11] G. Virone, et al., “Antenna Pattern Measurement with UAVs:
continuously varies also within a single circular path. This Modeling of the Test Source,” European Conference on Antennas and
behavior could be not straightforward to implement with Propagation (EuCAP), Davos, Switzerland, Apr. 11-15 2016.
some pilot platforms. Subsequent post-processing will be [12] F. Paonessa, et al., “Antenna Pattern Measurement with UAVs: Cross
required in order to obtain the AUT pattern. Polarization Performance,” 36th ESA Antenna Workshop on
In the case of Fig. 3c, the distance between the AUT and Antennas and RF Systems for Space Science, Oct. 6-9 2015,
Noordwijk, the Netherlands.
the test-source is 100 m during the whole flight. The surface
[13] G. Virone, et al., “Antenna pattern measurements with a flying far-
envelope extends to 60° of zenith angle with 15° steps and a field source (hexacopter),” in IEEE International Conference on
total flight length of 1750 m. Antenna Measurements and Applications (CAMA), Antibes Juan-les-
Pins, France, Nov. 16-19 2014.
III. CONCLUSION [14] T. Fritzel, H.-J. Steiner, J. Habersack, H. Schippers, J.-G. Ferrante,
“Antenna Pattern Measurements of Full-Size Air Vehicles with an
Due to the limited accuracy of the onboard GPS-based Airborne Near-field Test Facility (ANTF)” European Test &
Telemetry Conference, ETTC 2005, Toulouse, France.
navigation controller and the effects of external disturbances
(e.g. wind), the real flight trajectories can differ from the [15] D. Jacobs, “Report on ECHO measurements of ORBCOMM dipoles
in Green Bank,” Aug. 28 2015.
planned ones by a few meters. Such a discrepancy is usually
slowly fluctuating (see path of Fig. 3a).
The GPS accuracy is not sufficient to compute the AUT
pattern with the desired level of accuracy. For this reason,
the flight trajectory is also measured with an additional
Differential Global Navigation Satellite System (DGNSS),
which provides centimeter-level accuracy. The discrepancy
between the paths measured with the GPS and the DGNSS is
generally consistent with a slow-varying error of few meters
rather than a random variation.

REFERENCES
[1] G. Virone, et al., “Antenna Pattern Verification System Based on a
Micro Unmanned Aerial Vehicle (UAV),” IEEE Antennas and
Wireless Propagation Letters, vol.13, pp. 169-172, Jan. 2014.
[2] C. Chang, et al., “Beam calibration of radio telescopes with drones,”
Publications of the Astronomical Society of the Pacific, vol. 127, no.
957, 2015.
[3] F. Üstüner, et al., “Antenna radiation pattern measurement using an
unmanned aerial vehicle (UAV),” General Assembly and Scientific
Symposium (URSI GASS), 2014 XXXIth URSI, Beijing, 2014, pp. 1-
4.
[4] L. Washburn, E. Romero, C. Johnson, B. Emery, C. Gotschalk, “New
applications for autonomous aerial vehicles (drones) in coastal
oceanographic research”, Ocean Sciences Meeting 21-26 Febraury
2016, New Orleans, Louisiana, USA.
[5] A. Nelles, et al, “Calibrating the absolute amplitude scale for air
showers measured at LOFAR”, JINST 10 (2015) no.11, P11005
arXiv:1507.08932.
[6] F. Paonessa, et al., “UAV-based pattern measurement of the
SKALA,” in IEEE Antennas and Propagation Society International
Symposium (AP-S/URSI), Vancouver, Canada, July 19-25 2015.
[7] F. Paonessa, et al., “VHF/UHF Antenna Pattern Measurement with
Unmanned Aerial Vehicles,” in IEEE International Workshop on
Metrology for Aerospace (MetroAeroSpace), Florence, Italy, June 22-
23 2016.
[8] P. Bolli, et al., “Antenna pattern characterization of the low-frequency
receptor of LOFAR by means of an UAV-mounted artificial test
source,” in SPIE Ground-based and Airborne Telescopes VI,
Edinburgh, Scotland, United Kingdom, June 26 – July 1 2016.
[9] P. Bolli, et al., “From MAD to SAD: the Italian experience for the
Low Frequency Aperture Array of SKA1-LOW”, Radio Science, vol.
51, issue 3, pp. 160–175, Mar. 2016.
[10] G. Pupillo, et al., “Medicina Array Demonstrator: calibration and
radiation pattern characterization using a UAV-mounted radio-

1305 #1570318319