117
IEEE SIGNAL PROCESSING MAGAZINE | September 2019 |
research stands to benefit greatly from the marriage of knowl-
edge accumulating in the communications literature and state-
of-the-art radar techniques. This article presents an overview
of DFRC systems from the information-embedding perspec-
tive, and discusses the various techniques and implementation
strategies that define the state of the art. We describe different
approaches to DFRC based on radar beampattern and radar-
waveform modulation, and we consider both phased-array and
multiple-input, multiple-output (MIMO) configurations, along
with uplink and downlink communications. The article con-
cludes by highlighting several challenges in DFRC systems
and outlining some possible future research directions.
DFRC configuration and signal model
To develop DFRC systems, researchers needed to devise sig-
naling strategies and modulations of the radar pulse and beam
that would lead to the integrated operation and improved use
of the finite RF spectrum. With this goal in mind, researchers
recognized that communication signals emitted from a radar
platform may convey radar signal and target characteristics
to other cooperating radars. To create a unified aperture and
bandwidth system in an RF-restricted environment, it is desir-
able to embed such information into radar pulses. The infor-
mation data rate is determined by the radar pulse repetition
frequency (PRF), whether the system uses a phased-array or
MIMO configuration, and the permissible incremental chang-
es in radar waveform structure and bandwidth.
In downlink communications, information is transmitted
from the DFRC platform toward one or more communication
users (Figure 2). The essence of downlink communications is
to embed messages into the radar emissions, preferably with-
out disturbing the radar operation. Unless otherwise stated, we
assume that the communication symbol duration equals the
pulse repetition interval (PRI) of the radar. To illustrate, we
consider the baseband transmit signal vector in two common
types of radar configurations: single-input, multiple-output
(SIMO) and MIMO.
SIMO radar
Consider a radar system with a linear transmit array com-
prising
antennas. The bandwidth and total transmit power
budget available to the DFRC system are denoted as
and
respectively. For a SIMO radar, the
#
baseband trans-
mit signal vector during the
th radar pulse can be defined as
(;
),tPt
SIMO
=
)
(1)
where
and
denote the fast time and pulse number, respec-
tively, ()
*
$ is the conjugate operation,
the unit-norm transmit
beamforming weight vector, and
z
is the radar waveform.
The radar signal
z
is assumed to have unit energy, i.e.,
() ,tdt 1
T
2
;;z =
z
where
z
is the waveform duration.
MIMO radar
Let
m
z
f=
be a predesigned set of orthogo-
nal waveforms that satisfy the condition
() ()ttdt
m
T
m
zz =
)
z
l
,
d -
l
where
$d
denotes the Kronecker delta func-
tion. During the
th pulse, the MIMO radar baseband trans-
mit signal vector can be expressed as a linear combination of
the individual orthogonal waveforms as
(;
) (),t
M
P
t
M
P
tswW
*
m
M
m
MIMO
zxz
=
/
(2)
where
is the
#
transmit beamforming weight vec-
tor associated with the mth orthogonal waveform,
_z
(),, ()
M
T
1
f
is the vector of orthogonal waveforms,
and
()
T
$
stands for the transpose operation. The
#
transmit beamforming weight matrix ,,Ww w
M1
f_
is
assumed to be normalized such that
,Mtr WW
H
=
with
$
being the trace of a square matrix and
()
H
$
the Hermi-
tian transpose.
In the transmit signal model (1), it is assumed that the trans-
mit beamforming weight vectors
and transmit waveform
z
satisfy the transmit beampattern and range-Doppler resolution
as mandated by the SIMO radar. Similarly, the transmit beam-
forming matrix
and vector of orthogonal waveforms
z
in (2) are assumed to be optimized to satisfy the requirements
mandated by the MIMO radar. In practice, perfectly orthogo-
nal waveforms with common spectral content are not realizable.
Instead, several techniques for the design and realization of
waveforms with low cross-correlations are reported in the lit-
erature (see [19] and references therein). Typically, the operating
parameters of the radar need to remain fixed within a coherent
processing interval (CPI). Communications, being secondary to
the primary radar function of the system, can be incorporated
by modulating the transmit beampattern, the radar waveforms,
or both. Several schemes for embedding information into radar
emissions have recently been reported. In principle, those that
minimize the impact on the radar operation can be grouped into
three categories: beampattern modulation, IM, and fast-time
Communication
User
Communication
User
Target
Radar
DFRC Platform
FIGURE 2. A diagram of a DFRC system.
Authorized licensed use limited to: Xi'an Univ of Posts & Telecom. Downloaded on August 19,2023 at 04:24:32 UTC from IEEE Xplore. Restrictions apply.