IEEE Geoscience and Remote Sensing Magazine - June 2022 - 70

that can be used for accurate positioning is precise point positioning.
Its main drawback is the long convergence time,
limiting its use for real-time positioning [37].
Concerning vertical or height accuracy, GNSS inA
PART OF GPR'S SUCCESS
LIES IN ITS ABILITY TO
OVERCOME THE
LIMITATIONS OF METAL
DETECTORS AND
MAGNETOMETERS, WHICH
ARE UNABLE TO DETECT
NONMETALLIC TARGETS.
formation is based on a reference surface, such as an
ellipsoid (e.g., the World
Geodetic System 1984) that
does not represent local variations
of the terrain profile.
That is, GNSS height corresponds
to the height over a
reference surface (ellipsoid
or geoid), not over the terrain.
Thus, when large areas
are scanned, GNSS height
information is not accurate
enough for GPR processing.
Therefore, additional sensors,
such as laser rangefinders,
have to be placed on board
the UAV to retrieve the ground profile. To illustrate this,
a comparison among the ground profile obtained using
GNSS-RTK information, a laser rangefinder sensor, and
radar altimeter is shown in [38, Fig. 4].
GPR SUBSYSTEM
The GPR subsystem is composed of a radar module connected
to a set of antennas that sends the radio-frequency signals
generated by the radar module toward the ground and
collects the electromagnetic waves reflected on the ground
and in the objects buried in it. Raw GPR measurements can
be sent to the ground station in real time or stored on board
the UAV (to be downloaded after the flight). The processing
of GPR measurements is usually conducted in the ground
station or in a workstation due to the computational burden
associated with GPR processing techniques.
In GPR systems, there is a tradeoff between range resolution
and penetration depth. Range or depth resolution is
proportional to the frequency bandwidth of the GPR system.
In this regard, the use of higher frequencies enables
broader bandwidths. Conversely, the penetration depth of
electromagnetic waves in the soil depends on the working
frequency of the GPR as well as the soil-constitutive parameters.
To illustrate this, [39, Fig. 1] provides an estimation
of the expected penetration depth as a function of the frequency
and soil moisture content. For example, at a frequency
of 3 GHz, the penetration depth ranges from 0.5 m (dry
soils, soil moisture of 1%) to a few centimeters for wet soils
(soil moisture of 30%). In addition to this, the choice of the
working frequency band is related to the radar module and
the kind of antennas to be mounted on board the UAV.
In general, UAV-based GPR systems can be classified into
two main groups, based on the kind of radar technology.
Several implementations of UAV-based GPR systems have
opted for continuous wave (CW) radars, either frequencymodulated
CW (FMCW) or stepped-frequency CW (SFCW)
70
[40]. Their main advantage is simplicity in terms of hardware
implementation. Other UAV-based GPR prototypes make
use of time-domain radar modules, either impulse-like
or transmitting pseudorandom signals. These radars send
mostly ultrawideband (UWB) signals and they might provide
a better signal-to-noise ratio than CW radars [41], [42].
Antennas are one of the major challenges when designing
and implementing UAV-based GPR systems. Antenna
size and weight are closely related to its working frequency
band: the lower the frequency, the larger (and heavier) the
antenna is. Thus, planar or horn-like antennas are good candidates
for these airborne systems thanks to their balance
among size, weight, and radiation performance. Unlike
ground-based GPR systems, cavity-backed spiral antennas
are not suitable for UAV-based GPR prototypes as they are
designed to match air-ground interface impedance, so their
performance in air-launched GPR systems is worsened.
COMMUNICATIONS SUBSYSTEM
The communications subsystem comprises at least the radio
control, consisting of a transmitter module within the
UAV remote controller, and a receiver module on board the
UAV. Typical frequencies of operation for the R/C are 433,
868, and 2,400-2,500 MHz. UAV R/C systems usually take
advantage of frequency-hopping, spread-spectrum techniques
to reduce interferences with other in-band radio
transmissions. Aside from the commands used to control
flight of the UAV, it is typical to transmit telemetry data
using this radio link. Alternatively, a dedicated data link
between the ground station and the UAV to transmit telemetry
data and the GPR measurements can be deployed. For
example, a wireless local area network operating in the 2.4and/or
5.8-GHz frequency bands can be used.
DESIGN AND IMPLEMENTATION APPROACHES
Payload integration on board the UAV can be conducted
under two different approaches. On the one hand, the
payload can be designed to be fully independent from
the UAV platform. This design methodology is illustrated
in Figure 2(a), where the airborne segment is divided
into two independent parts: the payload and UAV subsystems.
Under this approach, the payload can (must, in
some cases) include its own positioning subsystem for
GPR measurements georeferring. On the other hand, the
UAV and the payload can be designed and implemented
altogether, resulting in the architecture depicted in Figure
2(b). This architecture requires direct access to the
UAV's controller.
The main advantage of an independent payload with
respect to a payload integrated within the UAV is the
possibility of placing it in different UAV platforms and
models, requiring little modification of the software.
However, it still requires designing and manufacturing
an appropriate interface to adapt the payload to the UAV's
frame. Conversely, a payload integrated within the UAV
makes synchronization among the different subsystems
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE JUNE 2022
IEEE Geoscience and Remote Sensing Magazine - June 2022

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - June 2022
