You are on page 1of 11

IET Radar, Sonar & Navigation

Review Article

Development of frequency diverse array radar ISSN 1751-8784


Received on 15th May 2017
Revised 12th October 2017
technology: a review Accepted on 28th October 2017
E-First on 30th November 2017
doi: 10.1049/iet-rsn.2017.0207
www.ietdl.org

Abdul Basit1,2 , Wasim Khan1, Shafqatullah Khan3, Ijaz Mansoor Qureshi4


1School of Communication and Information Engineering, University of Electronic Science and Technology of China, Chengdu, People's Republic

of China
2Department of Electrical Engineering, International Islamic University, Islamabad, Pakistan
3Advanced RF & Microwave Research Group, Department of Communication Engineering, Faculty of Electrical Engineering, Universiti

Teknologi Malaysia, 81310 Skudai, Johor, Malaysia


4Department of Electrical Engineering, Air University, Islamabad, Pakistan

E-mail: abdulbasit@iiu.edu.pk

Abstract: Electronic beam steering with greater efficiency is a vibrant feature of a phased array antenna, but for all the range
cells, it is fixed at a specific angle. To mitigate this problem, frequency diverse array (FDA) antenna was proposed. This study
presents a review on the development of FDA technology in radar and navigation applications. FDA is different from a
conventional phased array antenna radar in a sense that it uses a small frequency offset across the array, which helps to
generate a range, angle and time-dependent beam pattern. This pattern assures the energy transmission towards the desired
angle and range cell. In addition, this study also focuses the research on getting range-angle uncoupled beam patterns along
with diverse hybrid cognitive FDA design, available in the literature, for improved radar performance.

1 Introduction 22]. Furthermore, unequal inter-element spacing may increase the


aperture of array without increasing the employed number of
Phased array is a most common arrangement of arrays with antenna elements, this helps in thinning or broadening of beam-
multiple transmit antennas. It has been used initially in late 1930s width and reduces SLL with a reduced number of elements [23]. A
[1]. Nevertheless, the researchers are still investigating towards variety of algorithms have been applied in literatures to suppress
improvements in phased array technology [2–4]. Phased array interferences using unequally spaced phased arrays [24, 25].
antennas consist of a group of antennas, which are combined in Although phased arrays have many advantages, but some of its
such a manner that the relative phase among the elements is varied drawbacks in many applications are listed here. One of the major
in a specific pattern to achieve a constructive and destructive disadvantage is their high manufacturing cost. A more directed
interference at desired locations. This is meant to achieve a beam beam requires a greater number of antenna elements and electronic
pattern having high directivity in the desired direction, while phase shifters, which are quite expensive [26]. Another
generating side lobes in undesired directions [5]. The high gain disadvantage is that in phased arrays, exceeding inter-element
offered by phased arrays facilitates in detecting and tracking weak distance beyond half wavelength results in the appearance of
targets with small cross section, while suppressing strong grating lobes in the visible regions that deteriorates the radar
interferences from other directions [6]. The relative phased tracking performance [27]. Another disadvantage is that a phased
between the elements is varied using electronic phase shifters. In array antenna produces a power maximum at a fixed angle for all
order to control the side lobe levels (SLLs), an appropriate ranges. Therefore, it is very difficult to suppress range-dependent
amplitude distribution is applied across a phased array [7]. clutter and to localise multiple targets having same direction but
The phased array radars (PARs) having the properties of different ranges using phased arrays. Although many techniques
electronic steering, high gain and low side lobes, narrow beam exist that can suppress a range-dependent interference such as
width, multiple beams, variable scan rates and interference STAP techniques [28, 29] but they increase the computational cost
cancellation etc. have benefitted variety of radar and satellite and complexity. Therefore, the desire for more advanced array
applications [8]. These antennas have much more flexibility and antennas such as multitask radar [30] to radio astronomy [31] and
advantages compared to a conventional reflector antenna. pursuit to localise and track multiple targets in more than one
However, the high cost of phase shifters is one of the limiting dimension has come up with a novel beam steering technique
factors [9]. exploiting frequency diverse array (FDA) [32].
Phased array antennas allow electronic beam steering, multi- The frequency diversity was originally exploited to get
function radar operations [10], computation of complex weights to independent measurements of radar cross section (RCS) [33]. In
form a transmit-receive beam pattern for suppression of multiple order to do so, a single antenna was used to transmit frequency
interferences [11, 12], detection a moving target in the presence of diverse pulses sequentially. In contrast, an FDA radar is multiple
a clutter using space–time adaptive processing (STAP) [13], easy element radar like PAR. However, it is different from a
estimation of direction of arrivals using multiple signal conventional phased array antenna radar in a sense that it uses a
classification (MUSIC) algorithm [14, 15] and estimation of signal linearly increasing frequency offset across the array, which helps to
parameters by rotational invariance techniques algorithm [16] generate a range, angle and time-dependent transmit/ receive beam
along with the computation of target ranges exploiting time delay pattern [34].
between the transmitted and received signal peaks [17]. Moreover, The FDA is quite different from multiple input multiple output
PARs exhibit flexibility to select narrow and wide band array (MIMO) [35] radar as it transmits overlapped signals closely
structure with the choice of amplitude taper in transmit/receive spaced in frequencies, while MIMO, in some cases, employs
modules [18, 19] to get better null depths and thus improved orthogonal signals from widely separated antennas (i.e. widely
signal-to-interference and noise ratio (SINR) at the receiver [20– spaced MIMO) over multiple independent paths to provide spatial

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 165
© The Institution of Engineering and Technology 2017
2 Phased array radar
A uniform linear phased array architecture [45] is shown in Fig. 1.
An array of N elements is organised in a straight line with a
uniform inter-element spacing d, where w. is the respective weights
of each element, while θ is the direction of target. An N element
array transmits single waveform towards a point target from its
antennas scaled by complex weights w0, w1, …, wN − 1 as shown in
Fig. 1. The PAR array factor is given as [46]
N−1
AF θ = ∑ wn e jn νdsinθ

(1)
n=0

where ν = 2π/λ represents the wave number, while λ denotes the


Fig. 1  Uniform linear phased array with N elements wavelength, d is inter-element distance and θ is desired direction.
The vector form of this array is given as

AF θ = wHa θ (2)

here . H denotes the conjugate transpose of a vector, w is the


weight vector and a θ is the array steering vector which are given
as [47]
T
w = w0, w1, …, wN − 1 (3)

T
aθ = 1 e jvdsinθ e j2vdsinθ … e j N − 1 vdsinθ (4)

here . T represents the transpose of a vector. With uniform


weights, i.e. wn = 1, the array factor may be expressed as [48]

sin (N /2)vdsin θ
AF θ = (5)
Nsin (1/2)vdsin θ

Fig. 2  Synthesised beam pattern using a conventional beam former The beam pattern is generally used to represent the radiation
N = 15, θ0 = 40∘, νd = π pattern of an array, which shows the space distribution of
transmitted power. The beam pattern for the PAR antenna is given
diversity [36] and, in some cases, uses multiple waveforms to as [45]
provide waveform diversity from closely located antennas (i.e. 2
collocated MIMO) [37]. Likewise, it is also different from PPAR θ = AF θ 2
= wHa θ (6)
orthogonal frequency division multiplexing (OFDM) [38] as on
contrary to an FDA, an OFDM technique uses orthogonal sub- It is apparent from (6) that a beam pattern towards a desired target
carriers [39]. Moreover, an FDA is also different from the position can be generated by applying a suitable weight vector
conventional frequency scanning arrays [40], where each element known as beam former. It can be either adaptive to compute a
uses the same frequency at a given time [41]. weight vector adaptively for optimal performance or non-adaptive
Different from phased arrays, the FDA beam pattern provides a where weights are fixed to get high directivity in the desired
global maximum, as well as, a number of local maxima at diverse direction. The two commonly used adaptive beam formers are
angle and ranges values [42]. This ability can be utilised for minimum variance distortion less response and linear constrained
detecting multiple targets having same directions, but different minimum variance beam formers [20]. On the other hand, non-
range values. Moreover, the apparent scan angle of an FDA beam adaptive beam formers compute weights with uniform or non-
pattern is different from the nominal scan angle [32] and its uniform amplitude distribution [49–51]. The non-uniform
transmit/receive array factor is dependent upon on both the range distribution includes binomial, triangular and Dolph–Chebyshev
and angle [43]. Therefore, precise beam steering alike phased distributions while uniform weight distribution is known as
arrays is not sufficient. Moreover, its apparent scan angle can be conventional beam former [48].
outside of the real beam space that provides additional degrees of These conventional weights are generally computed as
freedom to perform multiple operations simultaneously [44]. w = a θ0 , where θ0 is the desired direction. Therefore, the beam
In this paper, we give a review on the development of FDA pattern of a PAR using conventional beam former is expressed as
radar technology i.e. from PAR to FDA and till now. We also [51]
highlight the potential research areas, in which the research is
going on. Moreover, we present technical challenges and future N−1 2

∑ e jn νd sinθ − sinθ0
2
directions. Additionally, we appeal to the antenna and wireless PPAR θ = wHa θ =
propagation research community to contribute in terms of more n=0
(7)
publications on FDA research and development. sin (N /2)νd sin θ − sin θ0 2

The remaining sections are organised as follows: Section 2 =


sin (1/2)νd sin θ − sin θ0
introduces the basic PAR technology to make its connection with
FDA. Section 3 discusses the basic FDA radar scheme, its
generated beam pattern analysis, present current research on FDA Fig. 2 shows the synthesised beam pattern of a PAR using a
with non-uniform frequency offset and cognitive FDA research. conventional beam former weight vector. The maxima of the beam
Section 4 discusses several future research problems in FDA radar pattern is towards θ = θ0 as apparent from (7) and rejects all the
technology. Finally, the conclusion is made in Section 5. interferences outside the illuminated spatial sector.

166 IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175
© The Institution of Engineering and Technology 2017
3 FDA radar 2πΔ f r
i.e. 2πΔ f t − + νdsin θ = 2πm, m = 0,
c
Fig. 3 shows a linear FDA structure having N antenna elements. (15)
An FDA uses a small frequency offset Δ f compared to the ± 1, ± 2, …
fundamental frequency f 0, across the array elements [52]. The
input frequency at nth antenna element is given as [53] where ν = 2π/λ.
Considering (15), there exist enormously many pairs of range
f n = f 0 + nΔ f , n = 0, 1, …, N − 1 (8) and angle values for a fixed time t = t0, which cause multiple peaks
in the range dimension. Similarly, keeping two parameters fixed
where f 0 is a fundamental frequency and Δ f is a frequency offset changes the FDA pattern into a periodic function of the third
value. Each antenna element transmits a signal waveform sn t parameter [54]. Figs. 5a–c show the periodicity of an FDA beam
given as pattern. The fundamental periods of the FDA pattern w.r.t. the
range, time and angle are c/(Δ f ), 1/(Δ f ) and (2π/νd), respectively
sn t = exp( − j2π f nt), n = 0, 1, …, N − 1 (9) [54]. Therefore, the periodicity property of an FDA pattern can be
utilised to generate power maxima at desired range-angle position
The distance between the antenna elements is d = (λ/2), where pairs by controlling Δ f across the array antenna elements.
λ = (c/ f 0). The range-angle-dependent FDA beam pattern allows the radar
The array factor of an FDA has been derived in [54]. Therefore, system to illuminate a desired spatial sector, while suppressing the
adapting from (1) to the notations used in this paper, we can write range-dependent clutter and interferences [42]; thus, resulting in
improved SINR. On the other hand, this range angle coupling may
N−1 degrade the range and angle estimation performance of an FDA,
rn
AF t, Δ f , r, θ = ∑ wn × r1n exp

− j2π f nt −
λn
(10) therefore, in [57] a range-angle-coupled beam pattern for improved
n=0 performance has been explored with frequency diverse chirp
signals. Likewise, an FDA structure, for bi-static radars, has been
where rn = r0 + ndsin(θ) denotes a path difference and λ denotes introduced in [58, 59] for improved estimation performance.
wavelength. Assuming a point target in space, the generated FDA Furthermore, an FDA design, with transmit sub-apertures, has been
beam pattern PFDA with uniform weights, i.e. wn = 1 can be written proposed in [60, 61] for improved range angle estimation of targets
as [54, 55] and improved receive beampattern performance. Similarly, for
improved range-angle localisation, a double pulse FDA radar
N−1 2 system (i.e. pulses with zero and non-zero frequency increments)
r
PFDA t, Δ f , r, θ = ∑ r1n exp − j2π f nt − λnn (11) was proposed in [62], which used PAR for direction detection and
n=0 FDA for target range localisation.
Furthermore, a joint range-angle estimation using MIMO radar
For a far field target, i.e. rn ≃ r0 and considering f 0 ≫ Δ f , the with FDA has been utilised for target localisation in range-
pattern can be approximated as [55] ambiguous scenarios along with getting an improved receive
beampattern [63]. As the transmit/receive FDA pattern is range-
N−1 2 ambiguous and range dependent, a vertical FDA approach has been
exp jγ
PFDA t, Δ f , r, θ ≃
r0 ∑ exp − jnψ (12) proposed to separate range-ambiguous clutter in vertical frequency
n=0 domain and then achieving clutter suppression in each separated
region using STAP methods [64]. Another adaptive receive
where γ = − 2π f 0(t − (r0 /c)) and beamforming for MIMO-FDA has been proposed in [65] to
ψ = 2πΔ f t − ((2πΔ f r)/c) + νdsin θ . suppress the range-dependent interferences for improved target
The beam pattern can also be expressed as (adapting from (5)) localisation. Furthermore, a doubled pulsed MIMO-FDA has been
proposed in [66] for improved target localisation. Another MIMO-
FDA radar transmitter has been designed to solve a Cramer–Rao
PFDA t, Δ f , r, θ lower bound minimisation problem using convex optimisation to
sin (N /2) 2πΔ f t − ((2πΔ f r)/c) + νdsin θ 2 localise targets via beam-spaced-based MUSIC algorithm [67]. An
= (13) MIMO-FDA radar design with a deceptive jamming suppression
sin (1/2) 2πΔ f t − ((2πΔ f r)/c) + νdsin θ
sin (N /2)ψ 2 method has been proposed in [68], where the difference between
= true and false target has been analysed and deceptive jamming
sin (1/2)ψ
suppression has been studied.
The FDA generated beam pattern expression in (13) is not merely a As the FDA beam pattern is time-variant, therefore, a pulsed
function of angle θ but also a function of time t , target range r FDA has been proposed in [69] to generate a quasi-static beam
pattern for enhancing the flexibility of beam scanning and its signal
and frequency offset value Δ f . From (13), the following
processing. For an improved detection performance of an FDA, a
observations can be concluded [56]
nested FDA concept was introduced in [70] to utilise more degrees
of freedom to improve the probability of target detection.
(i) With Δ f = 0, the FDA steering vector is alike a linear phased
Moreover, a nested FDA architecture was utilised in [71] for
array steering vector.
estimating the target direction of arrival with improvement. Due to
(ii) With a constant Δ f , an FDA beam pattern changes as a the coupling in range-angle responses, another sub-array scheme
function of target range. Additionally, with a fixed r, it becomes a where each array has a different frequency offset, has been
function of frequency offset. presented in for improved joint range-angle estimation for FDA
radar target [72]. A random FDA radar has been presented in [73],
Figs. 4a and b show the comparison between phased array and where each sub array has been assigned a random carrier frequency
FDA generated beam patterns. It is noteworthy that several and two different types of signal processing algorithms were
maxima are generated in the range dimension only. Using (13), the proposed, i.e. matched filtering and compressive sensing-based
magnitude of any maxima can be achieved as [42] algorithm for getting uncoupled range-angle FDA beam pattern. A
flattop FDA beam pattern problem has been efficiently solved
ψ = 2πm, m = 0, ± 1, ± 2, … (14) based on second-order cone programming using the inner point
method, where the array is divided into small sub arrays, with a
distinguished frequency offset for each array, to get improved radar
performance [74]. In [75], authors proposed a subarray-based FDA
framework, which obtains a range–angle-decoupled beam pattern

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 167
© The Institution of Engineering and Technology 2017
Fig. 3  Linear FDA structure with N antennas

Fig. 4  Comparison between beam patterns generated by PAR and FDA


(a) PAR generated beam pattern d = λ/2 , N = 15, f 0 = 10 GHz , (b) FDA generated beam pattern d = λ/2, N = 15, f 0 = 10 GHz, Δ f = 450 Hz, t = (1/Δ f ) = 2.2 ms

using multiple non-overlapped subarrays with different frequency Furthermore, along with linear array and planar array-based
shifts that exist in pairs. They proposed a novel beam pattern patents [87–89], two FDA patents [90, 91] have been issued
synthesis method based on semidefinite relaxation to merely focus vitalising the range-dependent characteristics of an FDA. In [90],
the transmit energy in a desired range–angle region. initially radio frequency signals are applied to a power divider
An FDA beam pattern with energy focusing on two- network. Moreover, an increasing frequency offset is applied
dimensional section of the surveillance region has been presented across all spatial channels, which introduces a new beam pattern
in [76] for improved target localisation performance. Moreover, a that focuses in different directions as a function of range. Similarly,
two-dimensional low-side lobe transmit beam pattern synthesis for in [91], radio frequency signals are applied to a transmit/receive
FDA radar has been presented in [77] for improved detection module which have been generated by a plurality of independent
performance. An FDA design for forward looking ground moving waveform generators to get a range-angle-dependent beam pattern.
target indication advantages has been analysed in [78], while The frequency offset plays a vital role to improve the
multipath characteristics of FDA over a ground plane have been performance of FDA radar therefore, researchers have shown their
examined in [79]. The usage of a linear frequency modulated keen interest to investigate different ways for its proper selection.
continuous waveform for FDA and its mathematical design have Consequently, FDA radars with small inter-element frequency
been proposed in [80]. offset [92, 93] have been used to generate range-dependent pattern.
The pursuit, to get more improved FDA receive beam pattern, On the contrary, large frequency offset [43, 55, 94] has been
has resulted in different types of FDA receiver structures, some of exploited to achieve independent echoes from the target for
those have been proposed in [81]. In FDA receiver array structures, improved localisation performance.
erroneous phase shifts and erroneous frequency computation Some useful investigations about exploring the ways to apply
methods that lead to inaccuracy in direction of arrival and the frequency offset across an FDA have been made in [62, 95] for
decreased SINR need a special attention of researchers. In [81], the improved range-angle localisation performance. In [66], a double
author emphasised, firstly, on band limited coherent receiver FDA pulse MIMO-FDA radar has been presented with a double pulse
architecture that ensures no cross talk at the element levels and, technique to manipulate frequency offset to improve range-angle
secondly, full band coherent receiver FDA architecture that ensures target localisation performance. Moreover, an FDA with an
maximum signal from the target location. For improved received adaptive frequency offset selection scheme was proposed in [96] to
beam pattern performance, an FDA in a multi-mission multi-mode improve localisation performance. Likewise, an optimal frequency
system has been originally proposed in [44]. Moreover, the concept offset selection for MIMO-FDA radar was presented in [97], where
of code diversity with an FDA through frequency diverse chirps the optimal frequency offset was selected by maximising output
was presented in [57] for improved receive beamforming. SINR in each coherent interval. Similarly, an adaptive computation
Likewise, an improved joint range-angle estimation, using MIMO- of frequency offset at each scan based on the receiver feedback has
FDA radar, has been presented in [82], while a range-angle been proposed in [98] for improved radar performance. Likewise,
decoupled receive beam pattern with sub-array-based FDA has in [99], an FDA with a time-dependent frequency offset was
been presented in [75]. Moreover, the deceptive jamming proposed to get improved beam pattern for a given range and
suppression at a receiver using MIMO-FDA has been presented in direction. A null steering technique using suitable frequency offsets
[68], while range ambiguous clutter suppression for air borne for an FDA radar has been presented in [100] for improved
FDA-STAP radar has been presented in [83]. The planer FDA interference cancellation.
receiver architectures have been proposed in [84]. The FDA Moreover, the characteristics of an FDA radar beam pattern by
concept has also been applied towards synthetic aperture radar changing frequency offset, inter-element spacing and array
systems for achieving high-resolution imaging of targets [85, 86]. elements, have been investigated in [101] using the finite

168 IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175
© The Institution of Engineering and Technology 2017
Fig. 5  Periodicity of FDA generated beam-pattern in (a) range dimension (b) time dimension and (c) angle dimension having parameters
d = λ/2, N = 15, f 0 = 10 GHz, Δ f = 10 kHz
(a) t0 and θ0 fixed, (b) ro and θ0 fixed, (c) ro and t0 fixed

difference time-domain method. An FDA radar with time- Consequently, the non-uniformly spaced FDA radar was
dependent transmit weights have been presented in [102] for introduced in [55] for getting range-dependent receive beam
getting an improved target localisation and interference pattern for improved radar performance. Similarly, a non-uniformly
suppression performance. Similarly, a time-invariant beam pattern spaced FDA concept was introduced in [55], where an accurate
synthesis for transmitter side via weight design has been presented physical placement of transmitter and receiver was required to
in [103] for improved suppression of range-dependent achieve better SINR. To reduce this difficulty, a non-uniformly
interferences. The impacts of frequency offset errors on an FDA spaced FDA with an inter-element spacing proportional to its
and FDA-MIMO have been investigated in [104, 105], which are wavelength has been proposed in [95], which generates a non-
caused due to imperfect electronic devices and thus resulting in repeating beam pattern for a range-angle imaging of targets to
degradation of overall generated FDA beam patterns. improve SINR. The frequency offset and carrier frequency for this
An FDA radar with uniform inter-element frequency offset scheme cannot be altered in real time as it requires physical
generates a range-angle coupled beam pattern, which exhibits relocation of antennas. Similarly, in [95] a non-uniform FDA
maxima at multiple range and angle values [32]. Certainly, these transmitter has been proposed to generate a range-dependent beam
multiple maxima property of an FDA beam pattern are undesirable pattern for achieving an improved range-angle imaging. Therefore,
due to the facts that it allows the interferences located at any of the for generating range-angle uncoupled beam pattern and avoiding
maxima to affect the target returns. Moreover, it decreases the the change in physical sensor positions, researchers put their focus
maximum reachable range and does not facilitate to distinguish on the second option, i.e. uniformly spaced FDA with non-uniform
between targets that are very close in range. Therefore, researchers frequency offset.
are finding different methods apply non-uniform frequency offset Fig. 6 shows a uniformly spaced FDA antenna with non-
across an FDA to mitigate these disadvantages. uniform frequency offsets, having symmetrical elements around the
origin. This symmetry is with respect to frequency offsets. The
3.1 FDA radar with non-uniform frequency offset inter-element distance is d and total number of elements is
N = 2P + 1, here P is the maximum number of elements on each
An FDA generates a range and angle coupled beam pattern due to side of array from origin, while wn is the transmitter weights.
synchronous linearly changing frequency offset and element The frequency input at nth element is given as
spacing, which causes problem in parameter estimation. To
generate an uncoupled range-angle beam pattern, this f n = f 0 + Δ f n, n = 0, ± 1, …, ± P (16)
synchronisation needs to be disturbed by either using (i) non-
uniformly spaced array with uniform frequency offset or (ii)
here, f 0 is the fundamental frequency, Δ f n is the frequency offset
uniformly spaced array with non-uniform frequency offset [106].
The earlier deals with the physical positions of array sensors, while of nth element with reference to f 0, and is given by
the latter merely varies the non-uniform frequency offset across the
array. Δ f n = αnΔ f (17)

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 169
© The Institution of Engineering and Technology 2017
Fig. 6  Frequency diverse array structure with non-uniform frequency offsets

αn is a set of non-uniform frequency offset coefficients. These a r, α = e− j (2π ⋅ α−PΔ f ⋅ r)/ c , …, e− j (2π ⋅ α−1Δ f ⋅ r)/ c , 1,
coefficients are either integers or fractional values, but non- (24)
e j (2π ⋅ α1 ⋅ Δ f ⋅ r)/ c , …, e j (2π ⋅ αPΔ f ⋅ r)/ c
uniform.
It can be noted that these non-uniform frequency offsets only affect
3.2 Non-uniform FDA beamforming
the range steering vector, i.e. a r, α . The FDA beam pattern is a
Consider a far field target in space, the phase of signals transmitted function of angle θ and range (r), therefore, the following
by the middle (reference) and adjacent elements is given by [56] observations are in line [80]:

2π f 0 2π f 1 (i) If we take Δ f = 0 , then it reduces to a conventional PAR


Ψ0 = r and Ψ1 = r (18)
c 0 c 1 pattern.
(ii) If the values of αn are taken uniform, a conventional FDA beam
here, c is the speed of light, while r0 and r1 are the target distances pattern is generated.
from reference and adjacent elements, respectively. The frequency (iii) The non-uniform frequency offset does not affect the angle
f 1 is given as f 1 = f 0 + Δ f 1, while Δ f 1 = α1Δ f and dimension of beam pattern as there is no αn in a θ vector.
r1 = r0 − dsin θ, where θ is the target direction. The phase
difference of these signals is given as ^
If a desired target position is at θ0, r^0 , then the weights for steered
2πdsin θ α Δ f r0 beam pattern to achieve a maximum gain, can be given as
Ψ1 − Ψ0 = − f 0 + α1Δ f − 1 (19)
c dsin θ ^
αnΔ f r^0 ndsin θ0
wT , n = exp − j2π f n + , n = 0, ± 1, …,
Similarly, the phase difference between the reference and Pth c c
element can be written as (25)
±P
2πdsin θ αPΔ f r0
Ψ M − Ψ0 = − f 0P + αPPΔ f − (20)
c dsin θ
where wT , n is a weight of nth transmitter element. The array factor
since f 0 ≫ αPΔ f and r0 ≫ αPdsin θ , hence the term containing
of an FDA system is given by
their product, i.e. ((2παpPΔ f dsin θ)/c) may be neglected. In case
of far filed targets, i.e. rn ≃ r, the steering vector a θ, r, α can be P
2π f 0ndsin θ 2π ⋅ αnΔ f ⋅ r
taken as AF = ∑ wnexp − j
c

c
(26)
n= −P

a θ, r, α = The beam pattern towards the desired target can be approximated


as the magnitude squared of the array factor and is given by
e j ((2π f 0Pdsinθ)/ c) − ((2π ⋅ α−PΔ f ⋅ r)/ c) , …, e j ((2π f 0dsinθ)/ c) − ((2π ⋅ α−1Δ f ⋅ r)/ c) ,
j((2π f 0dsinθ)/ c) − ((2π ⋅ α1 ⋅ Δ f ⋅ r)/ c)
1, e− , …, e− j ((2π f 0Pdsinθ)/ c) − ((2π ⋅ αPΔ f ⋅ r)/ c)
PFDA r0, θ0, α
^ 2 (27)
P f 0nd sin θ − sinθ0 αnΔ f r − r^0
(21) ≃ ∑ exp − j2π −
n= −P
c c
It can be represented in vector form, given as
Originally in [55], an FDA with non-uniform frequency offsets has
a θ, r = a θ ⊙ a r, α (22) also been proposed and simulations were carried out to indicate the
modified shape of beam pattern in terms of null depths and
where placements. After that a uniform linear array with logarithmically
increasing inter-element frequency offset has been presented in
a θ = e j (2π f 0Pdsinθ)/ c , …, e j (2π f 0dsinθ)/ c , 1, e− j (2π f 0dsinθ)/ c , …, [107], which generates a single-maximum range-angle uncoupled
(23) beam pattern for an arbitrary value of frequency offset and also
e− j (2π f 0Pdsinθ)/ c suppresses interferences significantly. Similarly, a range-bin-based
MIMO-FDA with logarithmic frequency offset has been presented
and in [108] for detecting multiple targets present in different range
bins. Additionally, the performance analysis of MIMO-FDA with
logarithmic frequency offset in terms of improved SINR, detection
probability and Cramer–Rao lower bounds (CRLB) on range-angle

170 IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175
© The Institution of Engineering and Technology 2017
Fig. 7  Comparison of the beam pattern generated by
(a) Conventional FDA [32], (b) Log-FDA [107], (c) Multilog-FDA [113] d = λ/2, N = 15, f 0 = 10 GHz, Δ f = 4000 Hz, θ0 = 20∘, r0 = 2 and 5 m

estimation has been presented in [109]. Moreover, transmit-receive 120] frequency diverse coprime array radars with coprime
beamforming for MIMO log-FDA was presented in [110], it avoids frequency offset have been proposed for multi target localisation,
the periodicity of conventional FDA using non-uniform frequency which showed its superiority in terms of DOA-range resolution,
offsets and thus achieves improved uncoupled beam pattern. localisation accuracy, and the number of resolvable targets.
Similarly, an MIMO-FDA radar design with non-uniform Similarly, these coprime arrays have been utilised in MIMO-FDA
frequency offset have been presented in [111] for generating a radars for improving the number of resolvable targets and
pencil beam or flat-top beam, according to the situation, to detect improved localisation performance [121]. In [106], a Costas-
and track multiple targets. In [112], the values of frequency offset sequence modulated frequency offset have been utilised to produce
as well as sensor positions were optimally computed using particle random-like energy distribution in the transmit beam pattern and
swarm optimisation algorithm to generate a pencil beam with low- thumbtack transmit-receive beam pattern. It helps in improving the
side lobes using an FDA. range-angle estimation of target and can be used for low
Moreover, a convex-multi-log FDA has been proposed in [113] probability of intercept (LPI) radar applications. Fig. 7 shows the
to generate a dot-shaped FDA beam pattern. Despite the increase in comparison of FDA generated beam patterns using conventional
computational complexity due to multi-log structure, it showed frequency offset, logarithmically increasing frequency offset and
improved performance in terms of transmit energy focusing, side multilog frequency offset. The non-uniform frequency offsets help
lobe suppression and array resolution. In [114], non-uniform to generate a range-angle uncoupled beam pattern when compared
frequency offset has been used for optimising the FDA transmit with conventional FDA radar.
beam pattern using genetic algorithm. This technique also Additionally, different types of geometries have been
generated a dot-shaped beam pattern to get improved range-angle investigated for FDA radars. The beam pattern analysis of planer
target localisation performance. Moreover, genetic algorithm-based FDA design has been presented in [122], where it was shown that
non-uniform frequency offsets have been exploited in [115] to planer FDA provides better directionality and higher gain when
detect and localise multiple targets with same direction but compared with conventional FDA radar. Moreover, the circular
different ranges. An FDA radar with non-uniform frequency offset FDA geometries were deeply studied in [123], where theory, basic
based on mu-law formulae have been presented in [116], which analysis and beam steering for uniform circular FDA was presented
generated improved uncoupled beam pattern and showed improved to get the advantages of better spatial resolutions. Moreover, non-
radar performance in terms of SINR, probability of detection and uniform frequency offsets based on tangent hyperbolic function
CRLB estimation performance. were applied to this uniform circular FDA, where beam patterns of
An FDA with non-uniformly increasing integer frequency three different configurations of an FDA can be generated, just by
offsets has been utilised in [117] to achieve shaper beam pattern for adjusting a single function parameter of the tangent hyperbolic
improved energy focusing towards the target. Similarly, square function in order to get enhanced system detection capability and
increasing and cubic increasing non-uniform frequency offsets for signal-to-interference plus noise ratio [124]. Likewise, elliptical
an FDA have been utilised in [118] to show improved performance geometry for an FDA, i.e. elliptical FDA radars with uniform and
in both focusing transmit energy and target localisation non-uniform frequency offsets have been investigated in [125] for
performance, especially in the range dimension. Similarly, in [119, getting range selective beam patterns with decreased average SLL.

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 171
© The Institution of Engineering and Technology 2017
Fig. 8  Generic information flow of a cognitive FDA radar

The non-uniform frequency offsets were obtained using tangent FDA receiver and feedback link is used for improved radar
hyperbolic functions. performance.
Since, the current FDA radar designs are often confined to Taking the essence from adaptive CRs [142, 143], a hybrid
adaptive receivers, they are unable to sense the environment cognitive PAR [144] was presented to enhance the overall radar
through transmitters. Therefore, researchers’ pursuit to develop performance in terms of LPI beam forming. Stepping forward
some intelligent radar system, adaptive in both the receiver and the towards the FDA radar technology, an FDA radar for selecting the
transmitter, paved a way for an intelligent radar system known as frequency offset, cognitively, was presented in [96]. Moreover, a
hybrid cognitive FDA radars. frequency offset calculation method based on the receiver feedback
was proposed in [98] to improve the transmit energy focusing
3.3 Hybrid cognitive FDA radar performance in the target position. Moreover, cognitive null
steering using a cognitive FDA was proposed in [100], to improve
The research towards an intelligent radar, which can adapt itself to the interference suppression performance. A symmetric cognitive
the statistical changes of environment and unknown time varying FDA radar design with non-uniform frequency offsets along the
scenarios, has been one of the dynamic research areas since last array was proposed in [116], to generate a signal maximum sharp
decade [126]. To address this issue, modern radar design and wide beam pattern according to the requirement, received as
technologies have been dominated not only by signal processing feedback from receiver, for enhanced target detection, SINR and
and control techniques alone, but also with the hybridisation of CRLB-based estimation performance. Similarly, a cognitive FDA
both to meet numerous milestones [127]. These advancements in beam pattern synthesis for detecting multiple targets with same
the field of radar design have addressed diverse radar problems, direction but different ranges was presented in [115], which used
such as range estimation, RCS measurement, speed estimation, and genetic algorithm to compute the non-uniform frequency offset for
direction estimation [128, 129] quite effectively to enhance the employed FDA radar based on the receiver feedback.
radar detection and tracking performance. Likewise, an intelligent Improved moving target tracking performance using a cognitive
dynamic system like a human brain that can have the properties of FDA stealth radar was presented in [145, 146], which not only
problem solving, decision making, memory, learning and improved the radar tracking and localisation performance but also
perception may outperform all the conventional systems. enhanced the radar LPI beamforming performance. A robust
Therefore, the idea of cognitive radar (CR) was firstly coined by adaptive beamforming for fast moving target detection using an
Haykin [130], which had four essential points of CR, i.e. Bayesian FDA-STAP radar was presented in [147]. Moreover, a cognitive
filtering or Kalman filters and its variants [131–134] in the receiver FDA radar design to achieve increased SINR for improved target
for target tracking [135–137], dynamic programming at the tracking was presented in [148]. A range-angle-dependent
transmitter, memory and global feedback, which acts a facilitator cognitive beamforming using an FDA was presented in [149],
of intelligence. which used the receiver feedback to control the overall beam
The receiver senses the radar environment by continuously generation towards the target predicted position. Moreover, a
interacting with it to get the necessary information regarding cognitive MIMO-FDA for improved LPI performance was
targets/interceptors. This information is sent as a feedback from the presented in [150], to achieve the advantages of MIMO radar and
receiver to the transmitter. The transmitter of cognitive radar CR. A cognitive FDA with situational awareness was presented in
contains the intelligent signal processing block which allows to [151] to utilise the combined advantages of CR due to its feedback
change its properties for transmitting an appropriate beam pattern control loop and FDA range-angle-dependent beam forming.
in accordance with the statistical variations in the environment. Table 1 shows a comparison of three types of FDA radars
This cycle is known as perception-action cycle and keeps on discussed in this paper.
repeating itself for improved performance [138–141]. A generic Although a lot of research is going on in the field of FDA radar,
information flow of a cognitive FDA radar is given in Fig. 8, which yet there is a lot of room for new methods to be researched for
consists of signal processing blocks on both transmitter and improved radar performance. Below are given some future
receiver side to enhance the radar performance even more. directions for FDA radar.
Moreover, an FDA transmitter senses the environment using the

172 IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175
© The Institution of Engineering and Technology 2017
Table 1 Comparison among three types of FDA radar systems
Range–angle Computation Interference Detection Tracking On-the-fly
coupled/decoupled complexity suppression performance performance decision making
beam pattern (intelligence)
conventional FDA range–angle coupled reasonable good good good no
beam pattern
FDA radar with non- range–angle decoupled less improved improved improved no
uniform frequency beam pattern
offset
hybrid cognitive FDA Both more very improved very improved very improved yes
radar

4 Future work many advantages of FDA radar technology, yet the room for
improvement is always there.
Although FDA radar technology has many promising advantages
and it has turned a new leaf in the radar technology, but still there
are several FDA radar problems which need to be addressed. 6 References
[1] Milne, K.: ‘Phased arrays in radar’. IEE Tutorial Meeting on Phased Array
i. Along with a linear array geometry, different types of other Radar, 1989, pp. 1–2
geometries (some have been reported in this paper) for an FDA [2] Zrnic, D.S., Melnikov, V.M., Doviak, R.J., et al.: ‘Scanning strategy for the
multifunction phased-array radar to satisfy aviation and meteorological
radar can be used to get the full advantage of an FDA radar in needs’, IEEE Geosci. Remote Sens. Lett., 2015, 12, (6), pp. 1204–1208
terms of spatial, spectral and temporal aspects of the FDA. [3] Yoshikawa, E., Ushio, T., Kawasaki, Z., et al.: ‘MMSE beam forming on fast-
Optimal geometry for FDA radar needs to be investigated. scanning phased array weather radar’, IEEE Trans. Geosci. Remote Sens.,
2013, 51, (5), pp. 3077–3088
ii. Non-uniform frequency offsets applied along the FDA radar [4] Spasojevic, Z., Dedeo, S., Jensen, R.: ‘Dwell scheduling algorithms for
have many advantages, as reported in this paper, but there is a phased array antenna’, IEEE Trans. Aerosp. Electron. Syst., 2013, 49, (1), pp.
need for getting some closed form expressions to get optimised 42–54
non-uniform frequency offset to be applied along the FDA for [5] Mouhamadou, M., Vaudon, P., Rammal, M.: ‘Smart antenna array patterns
synthesis: null steering and multi-user beamforming by phase control’, Prog.
improved radar performance. Electromagn. Res., 2006, 60, pp. 95–106
iii. Waveform optimisation for an FDA radar needs to be further [6] Rocca, P., Haupt, R.L., Massa, A.: ‘Interference suppression in uniform linear
investigated for improved FDA radar performance. The arrays through a dynamic thinning strategy’, IEEE Trans. Antennas Propag.,
optimised waveform needs to be designed keeping in mind the 2011, 59, (12), pp. 4525–4533
[7] Harrington, R.F.: ‘Sidelobe reduction by nonuniform element spacing’, IRE
constraints of total bandwidth, number of transmit and receive Trans. Antennas Propag., 1961, 9, (2), pp. 187–192
FDA elements, transmit signal peak amplitude and minimum [8] Von Aulock, W.H.: ‘Properties of phased arrays’, Proc. IRE, 1960, 48, (10),
SLLs. This constrained optimisation problem should be further pp. 1715–1727
investigated. [9] Fourikis, N.: ‘Phased array-based systems and applications’ (Wiley, 1997)
[10] Brookner, E.: ‘Major advances in phased arrays: part I’, Microw. J., 1997, 40,
iv. Cognitive radar properties hybridised with FDA radar have (5), pp. 288–294
some great advantages (as reported in this paper), yet there is [11] Ghouz, H.H.M., Elghany, F.I.A., Qutb, M.M.: ‘Adaptive space-time
of a need a full investigation of hybrid cognitive FDA designs processing for interference suppression in phased array radar systems. II.
Tracking radar’. Seventeenth National Radio Science Conf., 2000, NRSC
for improved performance. May be more use of neural ‘2000, 2000, pp. B9/1–B9/7
networks and deep learning algorithms can be used to sense the [12] Allen, B., Ghavami, M.: ‘Adaptive array systems: fundamentals and
radar environment more precisely for improved cognitive FDA applications’ (John Wiley & Sons, 2006)
radar performance. The memory, a CR property, which keeps [13] Brown, R.D., Schneible, R., Wicks, M.C., et al.: ‘STAP for clutter
suppression with sum and difference beams’, IEEE Trans. Aerosp. Electron.
the information about the terrain and the radar environment, Syst., 2000, 36, (2), pp. 634–646
may be used in a more efficient way for improved radar [14] Schmidt, R.O.: ‘Multiple emitter location and signal parameter estimation’,
performance. These cognitive properties combined with FDA IEEE Trans. Antennas Propag., 1986, 34, (3), pp. 276–280
radar with non-uniform frequency offset can further improve [15] Rao, B.D., Hari, K.V.S.: ‘Performance analysis of root-MUSIC’, IEEE Trans.
Acoust., Speech, Signal Process., 1989, 37, (12), pp. 1939–1949
FDA radar performance and needs to be investigated [16] Roy, R., Kailath, T.: ‘ESPRIT-estimation of signal parameters via rotational
thoroughly. invariance techniques’, IEEE Trans. Acoust., Speech, Signal Process., 1989,
v. The optimal array processing algorithms and techniques, which 37, (7), pp. 984–995
are not only confined in receiver now (due to CR properties), [17] Carter, G.C.: ‘Time delay estimation for passive sonar signal processing’,
IEEE Trans. Acoust., Speech, Signal Process., 1981, 29, (3), pp. 463–470
need to be investigated even more. The CR properties, the non- [18] Agrawal, A.K., Holzman, E.L.: ‘Beamformer architectures for active phased-
uniform frequency offsets, learning about and sensing the radar array radar antennas’, IEEE Trans. Antennas Propag., 1999, 47, (3), pp. 432–
environment has extensively increased the computational 442
complexities not only on the receiver side but also on the [19] Willey, R.E.: ‘Space tapering of linear and planar arrays’, IRE Trans.
Antennas Propag., 1962, 10, (4), pp. 369–377
transmitter side, therefore optimal array signal processing [20] Khan, S.U., Qureshi, I.M., Zaman, F., et al.: ‘Null placement and sidelobe
algorithms are to be investigated. suppression in failed array using symmetrical element failure technique and
hybrid heuristic computation’, Prog. Electromagn. Res. B, 2013, 52, pp. 165–
184
5 Conclusion [21] Haupt, R.L.: ‘Phase-only adaptive nulling with a genetic algorithm’, IEEE
Trans. Antennas Propag., 1997, 45, (6), pp. 1009–1015
PAR has been already in its mature form, while research [22] Akdagli, K.: ‘Null steering of linear antenna arrays using a modified tabu
community is focusing towards FDA and CRs. Moreover, hybrid search algorithm’, Prog. Electromagn. Res., 2001, 33, pp. 167–182
designs of CR with PAR and FDA radars have promised significant [23] Ishimaru, A., Chen, Y.-S.: ‘Thinning and broadbanding antenna arrays by
performance improvement when compared with stand-alone PAR unequal spacings’, IEEE Trans. Antennas Propag., 1965, 13, (1), pp. 34–42
[24] Chun, T.F., Zamora, A., Lei, B.J., et al.: ‘An interleaved, interelement phase-
and FDAs. Additionally, FDA radar with non-uniform frequency detecting/phase-shifting retrodirective antenna array for interference
offset is also a new field to achieve and a control a range-angle reduction’, IEEE Antennas Wirel. Propag. Lett., 2011, 10, (10), pp. 919–922
uncoupled FDA beam pattern for improved performance. The [25] Tokan, F., Güneş, F.: ‘Interference suppression by optimising the positions of
performance improvement of FDA radar designs has been shown selected elements using generalised pattern search algorithm’, IET Microw.
Antennas Propag., 2011, 5, (2), pp. 127–135
in terms of enhanced SINR, improved detection probability, better [26] Brookner, E.: ‘Phased array radars-past, present and future’. RADAR 2002,
range and angle resolution, enhanced transmit energy focusing, 2002, pp. 104–113
better energy efficiency, improved tracking performance and [27] Ender, J.H.G., Wilden, H., Nickel, U., et al.: ‘Progress in phased-array radar
improved CRLB on range and angle estimates. Although, there are applications’. Microwave Symp. Digest, 2005 IEEE MTT-S Int., 2005, p. 4

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 173
© The Institution of Engineering and Technology 2017
[28] Li, J., Liao, G., Griffiths, H.: ‘Range-dependent clutter cancellation method in [65] Gao, K., Shao, H., Cai, J., et al.: ‘Frequency diverse array MIMO radar
bistatic MIMO-STAP radars’. 2011 IEEE CIE Int. Conf. on Radar (Radar), adaptive beamforming with range-dependent interference suppression in
2011, vol. 1, pp. 59–62 target localization’, Int. J. Antennas Propag., 2015, 2015, pp. 1–10
[29] Weiner, M.M.: ‘Proposed algorithm for sequential implementation of adaptive [66] Khan, W., Qureshi, I.M., Basit, A., et al.: ‘A double pulse MIMO frequency
antennas and receivers [letters to the editor]’, IEEE Antennas Propag. Mag., diverse array radar for improved range-angle localization of target’, Wirel.
2009, 51, (5), pp. 161–162 Pers. Commun., 2015, 82, (4), pp. 2199–2213
[30] Murthy, V., Pillai, U., Davis, M.E.: ‘Waveforms for simultaneous SAR and [67] Gao, K., Wang, W.-Q., Cai, J.: ‘Frequency diverse array and MIMO hybrid
GMTI’. 2012 IEEE Radar Conf. (RADAR),, 2012, pp. 51–56 radar transmitter design via Cramér–Rao lower bound minimisation’, IET
[31] Oliveri, G., Caramanica, F., Massa, A.: ‘Hybrid ADS-based techniques for Radar Sonar Navig., 2016, 10, (9), pp. 1660–1670
radio astronomy array design’, IEEE Trans. Antennas Propag., 2011, 59, (6), [68] Xu, J., Liao, G., Zhu, S., et al.: ‘Deceptive jamming suppression with
pp. 1817–1827 frequency diverse MIMO radar’, Signal Process., 2015, 113, pp. 9–17
[32] Antonik, P., Wicks, M.C., Griffiths, H.D., et al.: ‘Frequency diverse array [69] Xu, Y., Shi, X., Xu, J., et al.: ‘Range-angle-dependent beamforming of pulsed
radars’. IEEE National Radar Conf. – Proc., 2006, pp. 215–217 frequency diverse array’, IEEE Trans. Antennas Propag., 2015, 63, (7), pp.
[33] Barton, D.K.: ‘Frequency agility and diversity’, vol. 6 (Artech House on 3262–3267
Demand, 1977) [70] Zhu, C., Wang, W.-Q., Chen, H., et al.: ‘Detection performance analysis of
[34] Antonik, P.: ‘An investigation of a frequency diverse array’ (University nested frequency diverse array radar’. 2015 16th Int. Radar Symp. (IRS),
College London, UK, 2009) 2015, pp. 700–705
[35] Wang, W.-Q., Cai, J.: ‘MIMO SAR using chirp diverse waveform for wide- [71] Zhu, C., Wang, W.-Q., Chen, H., et al.: ‘Target direction-of-arrival estimation
swath remote sensing’, IEEE Trans. Aerosp. Electron. Syst., 2012, 48, (4), pp. using nested frequency diverse array’. 2015 Int. Conf. Estimation, Detection
3171–3185 and Information Fusion (ICEDIF), 2015, pp. 200–203
[36] Haimovich, A.M., Blum, R.S., Cimini, L.J.: ‘MIMO radar with widely [72] Huang, L., Li, X., Gong, P.-C., et al.: ‘Frequency diverse array radar for target
separated antennas’, IEEE Signal Process. Mag., 2008, 25, (1), pp. 116–129 range-angle estimation’, COMPEL: Int. J. Comput. Math. Electr. Electron.
[37] Li, J., Stoica, P.: ‘MIMO radar with colocated antennas’, IEEE Signal Eng., 2016, 35, (3), pp. 1257–1270
Process. Mag., 2007, 24, (5), pp. 106–114 [73] Liu, Y., Ruan, H., Wang, L., et al.: ‘The random frequency diverse array: a
[38] Ding, Y., Zhang, J., Fusco, V.: ‘Frequency diverse array OFDM transmitter new antenna structure for uncoupled direction-range indication in active
for secure wireless communication’, Electron. Lett., 2015, 51, (17), pp. 1374– sensing’, IEEE J. Sel. Topics Signal Process., 2017, 11, (2), pp. 295–308
1376 [74] Xu, Y., Shi, X., Li, W., et al.: ‘Flat-top beampattern synthesis in range and
[39] Wang, W.-Q.: ‘Mitigating range ambiguities in high-PRF SAR with OFDM angle domains for frequency diverse array via second-order cone
waveform diversity’, IEEE Geosci. Remote Sens. Lett., 2013, 10, (1), pp. 101– programming’, IEEE Antennas Wirel. Propag. Lett., 2016, 15, pp. 1479–1482
105 [75] Xu, Y., Shi, X., Xu, J., et al.: ‘Range–angle-decoupled beampattern synthesis
[40] Johansson, F.S., Josefsson, L.G., Lorentzon, T.: ‘A novel frequency-scanned with subarray-based frequency diverse array’, Digit. Signal Process., 2017,
reflector antenna’, IEEE Trans. Antennas Propag., 1989, 37, (8), pp. 984–989 64, pp. 49–59
[41] Vazquez, C., Garcia, C., Alvarez, Y., et al.: ‘Near field characterization of an [76] Shao, H., Chen, H., Li, J.: ‘Transmit energy focusing in two-dimensional
imaging system based on a frequency scanning antenna array’, IEEE Trans. sections with frequency diverse array’. 2015 IEEE China Summit and Int.
Antennas Propag., 2013, 61, (5), pp. 2874–2879 Conf. Signal and Information Processing (ChinaSIP), 2015, pp. 104–108
[42] Wang, W.-Q.: ‘Range-angle dependent transmit beampattern synthesis for [77] Gao, K., Chen, H., Shao, H., et al.: ‘A two-dimensional low-sidelobe transmit
linear frequency diverse arrays’, IEEE Trans. Antennas Propag., 2013, 61, beampattern synthesis for linear frequency diverse array’. 2015 IEEE China
(8), pp. 4073–4081 Summit and Int. Conf. Signal and Information Processing (ChinaSIP), 2015,
[43] Wang, W.Q.: ‘Phased-MIMO radar with frequency diversity for range- pp. 408–412
dependent beamforming’, IEEE Sens. J., 2013, 13, (4), pp. 1320–1328 [78] Baizert, P., Hale, T.B., Temple, M.A., et al.: ‘Forward-looking radar GMTI
[44] Antonik, P., Wicks, M.C., Griffiths, H.D., et al.: ‘Multi-mission multi-mode benefits using a linear frequency diverse array’, Electron. Lett., 2006, 42,
waveform diversity’. 2006 IEEE Conf. on Radar, 2006, p. 3 (22), pp. 1311–1312
[45] Balanis, C.A.: ‘Antenna theory: analysis and design’, vol. 1 (John Wiley & [79] Cetintepe, C., Demir, S.: ‘Multipath characteristics of frequency diverse
Sons, 2005) arrays over a ground plane’, IEEE Trans. Antennas Propag., 2014, 62, (7), pp.
[46] Brookner, E.: ‘Phased-array radars’, Sci. Am., 1985, 252, (2), pp. 94–102 3567–3574
[47] Fenn, A.J., Temme, D.H., Delaney, W.P., et al.: ‘The development of phased- [80] Eker, T., Demir, S., Hizal, A.: ‘Exploitation of linear frequency modulated
array radar technology’, Physics, 2000, 12, (2), pp. 321–340 continuous waveform (LFMCW) for frequency diverse arrays’, IEEE Trans.
[48] Van Trees, H.L.: ‘Optimum array processing: part IV of detection, estimation, Antennas Propag., 2013, 61, (7), pp. 3546–3553
and modulation’ (Wiley, New York, 2002) [81] Jones, A.M.: ‘Frequency diverse array receiver architectures’ (Wright State
[49] Van Trees, H.L.: ‘Detection, estimation, and modulation theory, optimum University, 2011)
array processing’ (John Wiley & Sons, 2004) [82] Xu, J., Liao, G., Zhu, S., et al.: ‘Joint range and angle estimation using MIMO
[50] Widrow, B., Stearns, S.D.: ‘Adaptive signal processing’, vol. 1 (Prentice-Hall, radar with frequency diverse array’, IEEE Trans. Signal Process., 2015, 63,
Englewood Cliffs, NJ, 1985), 491pp (13), pp. 3396–3410
[51] Ho, T.V., McWhirter, J.G., Nehorai, A., et al.: ‘Radar array processing’, vol. [83] Xu, J., Zhu, S., Liao, G.: ‘Range ambiguous clutters suppression for airborne
25 (Springer Science & Business Media, 2013) FDA-STAP radar’, IEEE J. Sel. Topics Signal Process., 2015, 9, (8), pp.
[52] Huang, J., Tong, K.-F.: ‘Frequency diverse array with beam scanning feature’. 1620–1631
2008 IEEE Antennas and Propagation Society Int. Symp., 2008, pp. 1–4 [84] Jones, A.M., Rigling, B.D.: ‘Planar frequency diverse array receiver
[53] Huang, J., Tong, K.F., Baker, C.: ‘Frequency diverse array: simulation and architecture’. Radar Conf. (RADAR), 2012 IEEE, 2012, pp. 145–150
design’. Antennas & Propagation Conf., 2009. LAPC 2009, Loughborough, [85] Farooq, J., Temple, M.A., Saville, M.A.: ‘Application of frequency diverse
2009, pp. 253–256 arrays to synthetic aperture radar imaging’. 2007 Int. Conf. Electromagnetics
[54] Secmen, M., Demir, S., Hizal, A., et al.: ‘Frequency diverse array antenna in Advanced Applications, ICEAA'07, 2007, pp. 447–449
with periodic time modulated pattern in range and angle’. IEEE National [86] Farooq, J., Temple, M.A., Saville, M.A.: ‘Exploiting frequency diverse array
Radar Conf. – Proc., 2007, pp. 427–430 processing to improve SAR image resolution’. 2008 IEEE Radar Conf.,
[55] Sammartino, P.F., Baker, C.J., Griffiths, H.D.: ‘Frequency diverse MIMO RADAR 2008, 2008
techniques for radar’, IEEE Trans. Aerosp. Electron. Syst., 2013, 49, (1), pp. [87] Collinson, D.L., Rougas, J.A.: ‘Integrated antenna radar system for mobile
201–222 and transportable air defense’. Google Patents, 19 August 2003
[56] Wang, W.-Q., Shao, H., Cai, J.: ‘Range-angle-dependent beamforming by [88] Schaubert, D.H., Farrar, F.G., Hayes, S.T., et al.: ‘Frequency-agile,
frequency diverse array antenna’, Int. J. Antennas Propag., 2012, 2012, pp. 1– polarization diverse microstrip antennas and frequency scanned arrays’.
10 Google Patents, 04 January 1983
[57] Higgins, T., Blunt, S.D.: ‘Analysis of range-angle coupled beamforming with [89] Chan, J.K.: ‘Planar phased array antenna assembly’. Google Patents, 14
frequency-diverse chirps’. 2009 Int. Waveform Diversity and Design Conf., September 1999
2009, pp. 140–144 [90] Wicks, M.C., Antonik, P.: ‘Frequency diverse array with independent
[58] Sammartino, P.F., Baker, C.J.: ‘Developments in the frequency diverse bistatic modulation of frequency, amplitude, and phase’. Google Patents, 15 January
system’. 2009 IEEE Radar Conf., 2009, pp. 1–5 2008
[59] Sammartino, P.F., Baker, C.J.: ‘The frequency diverse bistatic system’. 2009 [91] Wicks, M.C., Antonik, P.: ‘Method and apparatus for a frequency diverse
Int. Waveform Diversity and Design Conf., 2009 array’. Google Patents, 31 March 2009
[60] Wang, W.-Q.: ‘Subarray-based frequency diverse array radar for target range- [92] Antonik, P., Wicks, M.C., Griffiths, H.D., et al.: ‘Range dependent
angle estimation’, IEEE Trans. Aerosp. Electron. Syst., 2014, 50, (4), pp. beamforming using element level waveform diversity’. Proc. Int. Waveform
3057–3067 Diversity and Design Conf., 2006, pp. 1–4
[61] Wang, W.Q., So, H.C.: ‘Transmit subaperturing for range and angle [93] Yong-guang, C., Yun-tao, L., Yan-hong, W., et al.: ‘Research on the linear
estimation in frequency diverse array radar’, IEEE Trans. Signal Process., frequency diverse array performance’. 2010 IEEE 10th Int. Conf. Signal
2014, 62, (8), pp. 2000–2011 Processing (ICSP), 2010, pp. 2324–2327
[62] Wang, W.Q., Shao, H.: ‘Range-angle localization of targets by a double-pulse [94] Sammartino, P.F., Baker, C.J., Griffiths, H.D.: ‘Range-angle dependent
frequency diverse array radar’, IEEE J. Sel. Topics Signal Process., 2014, 8, waveform’. Radar Conf., 2010 IEEE, 2010, pp. 511–515
(1), pp. 106–114 [95] Wang, W.-Q., So, H.C., Shao, H.: ‘Nonuniform frequency diverse array for
[63] Xu, J., Liao, G., Zhu, S., et al.: ‘Joint range and angle estimation using MIMO range-angle imaging of targets’, IEEE Sens. J., 2014, 14, (8), pp. 2469–2476
radar with frequency diverse array’, IEEE Trans. Signal Process., 2015, 63, [96] Shao, H., Li, J., Chen, H., et al.: ‘Adaptive frequency offset selection in
(13), pp. 3396–3410 frequency diverse array radar’, IEEE Antennas Wirel. Propag. Lett., 2014, 13,
[64] Xu, J., Liao, G., So, H.C.: ‘Space-time adaptive processing with vertical pp. 1405–1408
frequency diverse array for range-ambiguous clutter suppression’, IEEE [97] Xiang, Z., Chen, B.: ‘Optimal frequency increment selection in frequency
Trans. Geosci. Remote Sens., 2016, 54, (9), pp. 5352–5364 diverse multiple-input–multiple-output radar’, IET Radar Sonar Navig., 2016,
10, (8), pp. 1431–1438

174 IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175
© The Institution of Engineering and Technology 2017
[98] Basit, A., Qureshi, I.M., Khan, W., et al.: ‘Cognitive frequency offset [122] Xu, Y., Shi, X., Xu, J., et al.: ‘Beampattern analysis of planar frequency
calculation for frequency diverse array radar’. 2015 12th Int. Bhurban Conf. diverse array’, Int. J. RF Microw. Comput. Aided Eng., 2015, 25, (5), pp. 436–
Applied Sciences and Technology (IBCAST), 2015, pp. 641–645 444
[99] Khan, W., Qureshi, I.M.: ‘Frequency diverse array radar with time-dependent [123] Saeed, S., Qureshi, I.M., Khan, W., et al.: ‘An investigation into uniform
frequency offset’, IEEE Antennas Wirel. Propag. Lett., 2014, 13, pp. 758–761 circular frequency diverse array (UCFDA) radars’, Remote Sens. Lett., 2015,
[100] Saeed, S., Qureshi, I.M., Basit, A., et al.: ‘Cognitive null steering in 6, (9), pp. 707–714
frequency diverse array radars’, Int. J. Microw. Wirel. Technol., 2015, pp. 1–9 [124] Saeed, S., Qureshi, I.M., Khan, W., et al.: ‘Tangent hyperbolic circular
[101] Shin, J., Choi, J.-H., Kim, J., et al.: ‘Full-wave simulation of frequency frequency diverse array radars’, J. Eng., 2016, 1, (1)
diverse array antenna using the FDTD method’. 2013 Asia-Pacific Microwave [125] Saeed, S., Qureshi, I.M., Khan, W., et al.: ‘Elliptical frequency diverse array
Conf. Proc. (APMC), 2013, pp. 1070–1072 radars: uniform and non-uniform frequency offsets’, Remote Sens. Lett., 2017,
[102] Han, S., Fan, C., Huang, X.: ‘Frequency diverse array with time-dependent 8, (5), pp. 478–487
transmit weights’. 2016 IEEE 13th Int. Conf. Signal Processing (ICSP), 2016, [126] Singh, V.K., Kankanhalli, M.S.: ‘Adversary aware surveillance systems’,
pp. 448–451 IEEE Trans. Inf. Forensics Sec., 2009, 4, (3), pp. 552–563
[103] Shao, H., Li, X., Wang, W.Q., et al.: ‘Time-invariant transmit beampattern [127] Haykin, S.: ‘Adaptive radar signal processing’ (John Wiley & Sons, 2007)
synthesis via weight design for FDA radar’. 2016 IEEE Radar Conf. [128] Nathanson, F.E., Reilly, J.P., Cohen, M.N.: ‘Radar design principles-Signal
(RadarConf), 2016, pp. 1–4 processing and the environment’, NASA STI/Recon Tech. Rep. A, 1991, 91, p.
[104] Gao, K., Chen, H., Shao, H., et al.: ‘Impacts of frequency increment errors on 46747
frequency diverse array beampattern’, EURASIP J. Adv. Signal Process., [129] Skolnik, M.I.: ‘Radar handbook’ (McGraw-Hill, 1970)
2015, 2015, (1), pp. 1–12 [130] Haykin, S.: ‘Cognitive radar: a way of the future’, IEEE Signal Process.
[105] Gao, K., Shao, H., Chen, H., et al. ‘Impact of frequency increment errors on Mag., 2006, 23, (1), pp. 30–40
frequency diverse array MIMO in adaptive beamforming and target [131] Kalman, R.E.: ‘A new approach to linear filtering and prediction problems’, J.
localization’, Digit. Signal Process. Rev. J., 2015, 44, (1), pp. 58–67 Fluids Eng., 1960, 82, (1), pp. 35–45
[106] Wang, Z., Wang, W.-Q., Shao, H.: ‘Range-azimuth decouple beamforming for [132] Julier, S.J., Uhlmann, J.K.: ‘New extension of the Kalman filter to nonlinear
frequency diverse array with Costas-sequence modulated frequency offsets’, systems’. AeroSense'97, 1997, pp. 182–193
EURASIP J. Adv. Signal Process., 2016, 2016, (1), p. 124 [133] Wan, E., Van Der Merwe, R.: ‘The unscented Kalman filter for nonlinear
[107] Khan, W., Qureshi, I.M., Saeed, S.: ‘Frequency diverse array radar with estimation’. The IEEE 2000 Adaptive Systems for Signal Processing,
logarithmically increasing frequency offset’, IEEE Antennas Wirel. Propag. Communications, and Control Symp., AS-SPCC, 2000, pp. 153–158
Lett., 2015, 14, pp. 499–502 [134] Grewal, M.S., Andrews, A.P.: ‘Kalman filtering: theory and practice with
[108] Khan, W., Qureshi, I.M., Basit, A., et al.: ‘Range-bins-based MIMO MATLAB’ (John Wiley & Sons, 2014)
frequency diverse array radar with logarithmic frequency offset’, IEEE [135] Zhu, L., Su, W., Wu, S., et al.: ‘Phased-array radar beam tracking and
Antennas Wirel. Propag. Lett., 2016, 15, pp. 885–888 predicting’. 5th Int. Conf. Wireless Communications, Networking and Mobile
[109] Khan, W., Qureshi, I.M., Basit, A., et al.: ‘Performance analysis of MIMO- Computing, 2009, WiCom ‘09, 2009, pp. 1–4
frequency diverse array radar with variable logarithmic offsets’, Prog. [136] Daum, F.E., Fitzgerald, R.J.: ‘Decoupled Kalman filters for phased array radar
Electromagn. Res. C, 2016, 62, pp. 23–34 tracking’, IEEE Trans. Autom. Control, 1983, 28, (3), pp. 269–283
[110] Khan, W., Qureshi, I.M., Basit, A., et al.: ‘Transmit/received beamforming for [137] Skolnik, M.I.: ‘Radar handbook’ (McGraw-Hill, 2008)
MIMO log-frequency diverse array radar’. 2016 13th Int. Bhurban Conf. [138] Guerci, J.R.: ‘Cognitive radar: a knowledge-aided fully adaptive approach’.
Applied Sciences and Technology (IBCAST), 2016, pp. 689–693 2010 IEEE Radar Conf., 2010, pp. 1365–1370
[111] Wang, Y., Huang, G., Li, W.: ‘Transmit beampattern design in range and angel [139] Haykin, S., Xue, Y., Setoodeh, P.: ‘Cognitive radar: step toward bridging the
domains for MIMO frequency diverse array radar’, IEEE Antennas Wirel. gap between neuroscience and engineering’, Proc. IEEE, 2012, 100, (11), pp.
Propag. Lett., 2017, 16, pp. 1003–1006 3102–3130
[112] Lan, L., Liao, G., Xu, J., et al.: ‘Range-angle pencil-beamforming for non- [140] Haykin, S., Fuster, J.M.: ‘On cognitive dynamic systems: cognitive
uniformly distributed array radar’, Multidimens. Syst. Signal Process., 2017, neuroscience and engineering learning from each other’, Proc. IEEE, 2014,
pp. 1–20 102, (4), pp. 608–628
[113] Dai, J., Xiong, J., Chen, H., et al.: ‘Dot-shaped range-angle beampattern [141] Haykin, S.: ‘Cognitive dynamic systems: radar, control, and radio [point of
synthesis for frequency diverse array’, IEEE Antennas Wirel. Propag. Lett., view]’, Proc. IEEE, 2012, 100, (7), pp. 2095–2103
2016, 15, pp. 1703–1706 [142] Haykin, S.: ‘Adaptive radar: evolution to cognitive radar’. IEEE Int. Symp.
[114] Xiong, J., Wang, W.-Q., Chen, H.: ‘Frequency diverse array transmit Phased Array Systems and Technology, 2003, 2003, pp. 613–613
beampattern optimization with genetic algorithm’, IEEE Antennas Wirel. [143] Guerci, J.R., Guerci, R.M., Ranagaswamy, M., et al.: ‘CoFAR: cognitive fully
Propag. Lett., 2017, 16, pp. 469–472 adaptive radar’. IEEE Radar Conf., 2014, 2014, pp. 984–989
[115] Basit, A., Qureshi, I.M., Khan, W., et al.: ‘Beam pattern synthesis for a [144] Basit, A., Qureshi, I.M., Khan, W., et al.: ‘Hybridization of cognitive radar
cognitive frequency diverse array radar to localize multiple targets with same and phased array radar having low probability of intercept transmit
direction but different ranges’. 2016 13th Int. Bhurban Conf. Applied beamforming’, Int. J. Antennas Propag., 2014, 2014, pp. 1–10
Sciences and Technology (IBCAST), 2016, pp. 682–688 [145] Wang, W.-Q.: ‘Adaptive RF stealth beamforming for frequency diverse array
[116] Basit, A., Qureshi, I.M., Khan, W., et al.: ‘Cognitive frequency diverse array radar’. 2015 23rd European Signal Processing Conf. (EUSIPCO), 2015, pp.
radar with symmetric non-uniform frequency offset’, Sci. China Inf. Sci., 1158–1161
2016, 59, (10), p. 102314 [146] Wang, W.-Q.: ‘Moving-target tracking by cognitive RF stealth radar using
[117] Basit, A., Qureshi, I.M., Malik, A.N., et al.: ‘Beam sharpening of a range- frequency diverse array antenna’, IEEE Trans. Geosci. Remote Sens., 2016,
angle-dependent pattern using non-uniform symmetric but integer frequency 54, (7), pp. 3764–3773
offset’. 2016 Int. Conf. Intelligent Systems Engineering (ICISE), 2016, pp. [147] Xu, J., Liao, G., Huang, L., et al.: ‘Robust adaptive beamforming for fast-
232–235 moving target detection with FDA-STAP radar’, IEEE Trans. Signal Process.,
[118] Gao, K., Wang, W.-Q., Cai, J., et al.: ‘Decoupled frequency diverse array 2017, 65, (4), pp. 973–984
range–angle-dependent beampattern synthesis using non-linearly increasing [148] Wang, Z., Wang, W.Q., Xiong, J.: ‘Cognitive target tracking using FDA radar
frequency offsets’, IET Microw. Antennas Propag., 2016, 10, (8), pp. 880–884 for increased SINR performance’. 2016 IEEE Radar Conf. (RadarConf),
[119] Qin, S., Zhang, Y.D., Amin, M.G.: ‘Multi-target localization using frequency 2016, pp. 1–4
diverse coprime arrays with coprime frequency offsets’. 2016 IEEE Radar [149] Basit, A., Qureshi, I.M., Khan, W., et al.: ‘Range–angle-dependent
Conf. (RadarConf), 2016, pp. 1–5 beamforming for cognitive antenna array radar with frequency diversity’,
[120] Qin, S., Zhang, Y.D., Amin, M.G., et al.: ‘Frequency diverse coprime arrays Cogn. Comput., 2016, 8, (2), pp. 204–216
with coprime frequency offsets for multi-target localization’, IEEE J. Sel. [150] Huang, L., Gao, K., He, Z., et al.: ‘Cognitive MIMO frequency diverse array
Topics Signal Process., 2016, PP, (99), p. 1 radar with high LPI performance’, Int. J. Antennas Propag., 2016, 2016, pp.
[121] Qin, S., Zhang, Y.D., Amin, M.G.: ‘Multi-target localization using frequency 1–11
diverse coprime arrays with coprime frequency offsets’, IEEE Trans. Signal [151] Wang, W.Q.: ‘Cognitive frequency diverse array radar with situational
Process., 2015, 63, (6), pp. 1377–1390 awareness’, IET Radar Sonar Navig., 2016, 10, (2), pp. 359–369

IET Radar Sonar Navig., 2018, Vol. 12 Iss. 2, pp. 165-175 175
© The Institution of Engineering and Technology 2017

You might also like