Short-ranae surveillance

radar systems
by C. J. Baker and B. D. Trimmer

Small, light-weight, coherent radar systems have been successfully exploited in the
military domain for many years. Recent advances in both technology and signal
processing techniques are enabling the production of more versatile systems aimed a t
a wider range of applications spanning both the military and civil markets. In this
paper the fundamental design concepts underpinning this class of radar are briefly
introduced t o place in context current and emerging technical developments. The
relationships between these developments, an increasing range of future potential
applications and the areas of most significant technical challenge are discussed.

1 Introduction Thirdly, radar is inherently coherent and consequently

backscatter from a distributed scene will exhibit
In the modern military environment it is universally constructive and destructive intederence, leading to the
recognised that accurate, timely and reliable knowledge of well-known ‘speckle’phenomenon. This manifests itself as
enemy activity is a vital ingredient in the provision of scene-induced multiplicative noise.
overall situational awareness, potentially malting the These characteristics result in the radar returns being
difference between success and failure of operations. presented to an operator as a single bright spot on a
The detection and location of movement of, amongst display. This may lead to the return froin a vehicle, €or
others, vehicles, men, enemy fire and ‘the fall of shot’ are example, being indistinguishable lrom other bright spots
key situation awareness constituents leading to an on the display caused by similar returns from other man-
understanding of enemy dispositions, actions and intent. made objects or the surrounding terrain. However, the
The availability of this class of information on an all- deployment pattern of a group of vehicles and other radar
weather, day-and-night basis in a hostile and extremely derived data such as their speed and direction of travel may
complex environment can only be provided by radar make their identity and intent clear to a trained operator. In
systems. This advantage together with the ability to view order to aid the interpretation task the raw radar data is
relatively large areas makes radar a vital component within processed into a form making its assimilation as simple as
a commander’s overall portfolio of surveillance and target possible. Overall the radar designer must have a deep
acquisition assets. Indeed, soon after its initial invention understanding of hardware, processing algorithms,
radar was used for military surveillance of the battlefield display of processed data, the role of the radar operator and
and this continues to the present day. the operating environment and all o€ their interactions if
These advantageous characteristics of radar systems applications are to be addressed successfully.
have been exploited in numerous other applications. The underlying design principles which determine
Examples which fall into the category ‘short-range detection and classification performance, location
surveillance’ include harbour surveillance, border control, accuracies and overall effectiveness ol short-range
traftic monitoring, airport and building security. surveillance radar systems are briefly introduced in the
Humans are used to viewing scenes at optical next section. This is followed by a review of past and
frequencies and our natural abilities to interpret context current battlefield radars, highlighting the proliceration of
and content are extraordinarily good. However, for radar small light-weight systems. This background places
the situation is somewhat more complicated for three main current and likely future developments in a context which
reasons. Firstly, radar operating frequencies are much clearly demonstrates their importance and leads naturally
lower (up to 100 GHz) and result in objects appearing to the increasingly comprehensive capability and widening
relatively smooth with little or no textural information. range of applications to which these systems are being put.
Secondly, radar resolutions are typically of the order ol
10 m by a few hundred metres and consequently the 2 Basic design principles
detected backscatter will include contributions from both
objects of interest and their background; there is therefore All radar systems work on the principle of transmitting and
an effective loss of definition and of contextual information. receiving electromagnetic radiation, which may take the

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000 181

Table 1: Typical battlefield radar requirements of 10 km. Using techniques such as monopulse or azimuth
matched filtering this accuracy can be improved by
Detection ranges approximately an order of magnitude. There are many
Man walking 3 km
excellent introductory radar textbooks available that
Moving vehicle 10 km establish and describe these relationships (e.g. References
Fall of shot (artillery) 8 km 1-3). Fig. 1 shows a very simplified schematic diagram of
Range accuracy 20 m the major components comprising a radar system. Note
Bearing accuracy 0.3' that a reference signal is used to retain a knowledge of
Environment phase by comparison of the outgoing transmission and
Rain up t o 4 mm/h subsequent incoming reception.
Heavy clutter > - 10 dBm2/m2 The overall design of a radar system to meet a particular
application must take into account the complex
Weight 30 kg
interactions between the characteristics of the transmitter,
Volume man-portable loads
receiver, antenna, target, environment, size, weight, power
Power consumption 30 W consumption, affordability and role of the operator. All of
these design aspects need to be carefully considered
form of a train of pulses or may be a continuous together in order to provide the most cost-effective
transmission and reception. The presence of a target is systems. Next, the design of an example radar system is
determined from the amount of incident radiation that is considered against the requirements of a hypothetical
reflected back in the direction of the radar receiver and this man-portable battlefield surveillance radar as in the outline
is a complex function of radar specification, target typej specification listed in Table 1. This will highlight the
characteristics and environment. The range to the target is significance of the role of the parameters described above
a function of the round-trip transit time from transmitter to in system design options.
target and back to receiver. The cross-range resolution and The technical specification of a radar system, at the
accuracy are functions of the radar's beamwidth. simplest level, is expressed by the well known radar
Used together these measurements enable a target or equation, which can be used to indicate the high-level
object of interest to be detected and its position to be system parameters and the effects of their interactions.
located in a convenient co-ordinate system. By examining The radar equation for noise-limited detection may be
the returns from moving targets over a period of time, written as:
radial velocity can be calculated. Usually this is achieved in
modern radars by measurement of the phase histoiy of the R4= PG2o A z
radar returns as this can be related directly to radial ( 4 7 ~LSkTBF
)~
velocity. As stated above, target location accuracy in the
cross-range dimension is a function of beamwidth. For a where R is the detection range, Pis the peak power, G is the
1.5" beam and no additional processing, this implies a antenna gain, o i s the radar cross-section of the target, A is
location accuracy of approximately 160 m at a radar range the wavelength, S is the minimum detectable signal, L
represents the system losses and kTBFis the noise power
Table 2: Typical system parameters in bandwidth B and with noise figure F.
Clcarly, to dct(vt a target oi a Rivcn cross-section at a
Radar system parameter Value
chosen range t l i c w are various u)nibin:itiotis of powcr,
Average transmitted power 1w wavelength and antenna gain that have to be optimised
Transmitter duty radio 0.1 against the constraints of external factors. One example,
Receiver noise figure 5.0 dB for a man-portable system that shows this in a simple way,
is the contrasting needs to propagate through adverse
System losses 7.0 dB
weather whilst maintaining a small and manageable
Elevation beam loss 1.O dB
antenna size. The former will limit the upper frequency
Scan rate 36 degls
whilst the latter will tend to impose a limit on the lower
Antenna size 1 x 0.5 m
frequency.
Antenna gain 38.5 dB A further example is provided by the way in which a
Azimuth 3dB beamwidth 1.5" predetermined antenna gain may be realised in practice for
Elevation 3dB beamwidth 3.0" the chosen application. This will be subject to external
Wavelength a t 15 GHz 0.02 m factors such as coverage optimisation, terrain masking, the
Dwell time at 15 GHz 0,0625 s likelihood of third party detection and jamming. For
example it may be advantageous to use a fan beam that is
System losses breakdown narrower in azimuth than elevation so that the radial extent
Frequency weighting and scalloping 2.0 dB of the area of interest (as projected onto typical terrain) can
Receiver range matching and scalloping 2.0 dB be increased to provide an extended field of coverage. This
Detection thresholding 1.O dB will also improve cross-range location accuracy, which is a
In-service degradation 2.0 dB
direct function of azimuth beam width, Alternative
applications may demand a wider azimuth beam, which

182 ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000

can scan an area more quickly, or a
pencil beam, which concentrates
energy onto a smaller area of
intersected ground.
oscillator and power transmit
By optimising the principal radar modulator amplifier
design parameters against a well
understood requirement in this way the
high-level design goals of the radar may reference
signal
be outlined. This can then be used as
preparation for iurther, more detailed
refinement. Table 2 shows a receive
representative set of the principal
parameters able to meet the
requirement for a man-portable
battlefield surveillance radar system
meeting the requirements in Table 1.
The relative importance and
significance of the parameters in design
optimisation can now be seen. For
example, consider the case of two
classes of targets - moving vehicles Fig. 1 Schematic diagram of radar system
and moving people - which are
assumed to have arbitrary radar cross-sections of 10 m2 3 Systems
and 0.5 m', respectively. The radar equation can now be
used to examine performance characteristics, such as the From its beginnings in the early 1940s the development oi
influence o i frequency on detection range. It is relatively short-range surveillance radar has been mainly aimed at
straightforward to calculate range and cross-range providing battlefield targeting information from a man-
accuracies, unambiguous Doppler bands and Doppler portable platform for lorward artillery observers. Notable
resolutions. radars have included the ZB298, which entered service in
Consider the calculation of the optimum operating the UK in the late 1960s. It operates in the I and J bands
frequency, for which some typical results are summarisecl using largely solid-state technology with the exception of
in Table 3. The optimum frequency is shown for detection the magnetron power source. It is man-portable in two
range, range and cross-range accuracy and Doppler segments with the radar head tripod-mounted and the
ambiguity. The frequency range over which the control system able to be located up to 20 m away. The
performance falls by 6 dB is also shown, which in turn system displays target returns on a gallium phosphide
reduces the range by approximately 30%and doubles the screen and additionally has an audio output used by the
measurement errors. A frequency of around 15 GHz operator to aid detection and act as a crude means of
probably offers the best compromise all-round classification. The ZB298 has been incorporated into two
performance but is clearly not optimal in all cases. For tracked armoured personnel carriers, the FV 103 and the
example, at longer ranges or in adverse weather FV 432 with the antenna mounted on a telescopic mast.
conditions, lower frequencies would be more suitable, or, il A similar system, the AN/PPS-15 radar, is in service in
ranges of approximately 7 kin or less are required, this the USAand worldwide.'I'here are currently many varieties
could be achieved at a higher frequency, around 35 GHz, of battlefield surveillance radar. Anumber (not all) of these
which might have size and weight advantages. are listed in Table 4, which shows the enormous variety of

Table 3: Optimum operating frequencies showing a 6 dB performance spread

Detection range Optimum 15 15 11 8

6 dB spread 4-35 4-35 2 - 20 2-16
Range accuracy Optimum 15 15 11 8
6 dB spread 4-35 4-35 2 - 20 2-16
Cross-range accuracy Optimum 35 35 15 12
6 dB spread 12 - 35 1 1 -35 7-25 6 - 19
Doppler ambiguity Spread (+IO0 km/h) 0 - 32 0 - 16 0 - 32 0 - 16

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000 183

Table 4: Battlefield and short-range ground surveillance radars. Detection ranges are given for large shells (e.g. 155 mm).
Data is taken from open literature sources.

Name Country Radiated Weight, Frequency Azimuth Azimuth Azimuth Detection range, km Minimum
power kg bandt bandwidth, coverage, accuracy, detect
deg deg Men Vehicle FOS* velocity,

*FOS = fall of shot

t I Band: 8-10 GHz. J Band: 10-20 GHz. B Band: 250-500 MHz. C Band: 0.5-1 GHz.

extant systems and widely ranging radar specifications. It against a typical ground environment. Fig. 2 shows the
is worth noting that it is often quite sinal1 changes in the MSL'AR equipment as a typical instance of the man-
operational requirements for these systems that have portable ground surveillance radar.
resulted in very significant diflerences in the final design. MSTAR uses a solid-state power amplifier and retains
Table 4 illustrates that although there is a clear correlation phase information Cor subsequent coherent processing.
between weight, power and detection range it is not The radar detects men and vehicle targets with a location
necessarily the case that the heavier and the more accuracy oi approximately 0.3", enables classificationfrom
powerful the radar the longer the detection range. A sound the Doppler signature (presented as an audio signal to the
understanding of the radar design fundamentals and the operator) and is able to detect and correct artillery fire.
electromagnetic backscattering properties of the target MSTARhas alow peak power to minimise hostile detection
and environment can result in much more efiicient, by electronic surveillance measures and operates from a
compact designs resulting in greater all-round operational standard field battery4.
versatility and effectiveness. A note of caution should also One of the most important aspects of this class of radar
be sounded as these specifications do not necessarily tell design is the required simplicity of the display to the
the whole story and other characteristics, such as usability, operator. In general, he or she is under high stress and
cost, reliability and technology vintage, also need careful cannot assimilate complex information presentations. Figs
consideration. 3 and 4 show typical MSTAR displays, representative of the
Advances in technology beyond that incorporated in the displays available lrom this general class ol radar system.
ZB298 have allowed the capability of later generation radar Fig. 3 shows the area surveillance mode, where some
systems to improve further and have led to manportable targets have been detected at the edge of the scanned
surveillance radar systems such as MSTAR (Moving and sector. Fig. 4 shows the display in acquisition mode;
Stationary Target Acquisition and Kecognition), currently effectively it is an enlarged section of the surveillance
in service with the British Army and IUSIT (Radar display using a 'B Scope' format and showing the trails
d'Acquisition et de Surveillance dans les Intervalles from two moving targets that have been observed over a
Terrestres) in service with the French Army. period of time. Note the absence of extraneous data that
MSI'AK is a pulse Doppler scanning radar system could distract the operator.
optimised for the detection of moving men and vehicles Current developments in research and technology are

184 ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000

Fig. 2 MSTAR deployed: (a) man-portable and (6)vehicle-mounted roles

resulting in systems with an ever-increasing radar Technology

capability with ever-reducing sizes and weights. Not only
can this lead to an increase in the capability of man-portable The design of radar systems Tor short-range surveillance
battlefield surveillance radar systems but it is also enabling applications needs to strike a balance between leading-
systems to be hosted on a greater range of platform types edge technology and application-driven requirements. For
and a greater range oi applications to be addressed. many applications weight, volume, power consumption,
However, the key to the successlul realisation of any performance, ease of use and cost are the principal design
system lies in optimisation of the radar performance in the

technology topic. box

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000 185

drivers that severely constrain the radar designer’s generated electromagnetic power to the areas of interest
options. For military systems low probability of intercept ancl is the first point ol reception of the backscattered
by hostile ESM (Electronic Support Measures) systems energy. Current systems use a mechanically scanned dish
and electronic protection measures is also important. It will antenna that is relatively simple to fabricate and is efficient
be seen in this section how technology advances are now and easy to operate. However, it has the disadvantage that
increasing the freedom of the radar designer from these the beam shape is fixed and therefore not always optimised
constraints. for purpose. The pointing of the beam is slow and not very
agile and the size and shape of the antenna make it
Transmitters and receivers cumbersome and physically vulnerable in many
All radar systems have to generate and amplify an applications. Many o l these disadvantages may be
electromagnetic source prior to transmission and it is in overcome with phased-array antennas, which can be
the design of power amplifierswhere most recent progress operated either passively or actively.
has been made. There has been an evolving trend to higher There is much research ongoing into phased arrays and
power devices (>lo W peak in high J-band, with they are beginning to find application in larger systems.
corresponding advances scaled to the other radar bands) The single biggest difficulty to be overcome in order to be
in smaller packages with higher efficiencies. The able to use phased arrays in small radars is cost. Current
waveform generator and power amplifier combination of phased arrays are prohibitively expensive. Conventional
tomorrow will also be capable of enhanced agility and will phased arrays, although reducing in cost, seem destined to
be able to host a wider variety of waveform types which can be out of reach of the class of radar considered here. There
be tailored to a given application. For example, this may is research ongoing into low-cost phased arrays that on
enable the duty cycle to be varied such that when coupled maturity would revolutionise the capability of short-range
with power management techniques the system is able to surveillance radar systems. These will enable nonlinear
minimise the likelihood of intercept by ESM systems at all scanning so that, for example, dwell times on targets of
ranges by automatically controlling the output power to interest could be increased whilst still maintaining an all-
suit the operating conditions. CW (continuous wave) round scanning capability. Laboratory prototypes have
operation may also be easily supported resulting in been fabricated, however much research is still required
significant potential cost savings that accompany this before they will be mature enough for production systems.
simpler radar design approach.
Future envisaged amplifiers will be compatible with Power generation
substantial increases in the resolution of the radar, which If the radar is to be self contained then a source of power
will be enhanced by more than an order of magnitude to will be required, most likely a battery if the radar is to be
less than 1m. This offers the potential for improved target used remotely from a mains-like power source. Battery
detection and clutter rejection and additionally opens up technology is advancing, with longer lifetimes available
one method for the classificationof targets (discussed later from smaller, lighter weight packages largely developed
in this paper). The sensitivity of receivers (coupled with for a wide range of applications.
improvements in receiver signal processing) has also The advances in battery technology as applied to this
improved notably, resulting in overall system losses being type of radar can be seen in the 24 V ‘Clansman’-typepower
reduced by several decibels, Together these source suitable for an MSAR-like radar. The typical Ni-
improvements enable the detection range of systems to be Cad form has been augmented with lithium or nickel-
extended by up to a factor of two without weight, volume or metal-hydride based disposable and rechargeable units.
cost penalties. Indeed, future systems may be envisaged Typical parameters of available military batteries are
with overall weights of considerably less than 30 kg. These shown in Table 5. Quite clearly there has been a factor of
advances instantly translate into significant improvements two improvement in power available per unit weight,
in the capability of current systems. Coupled with particularly if the user is prepared to consider a disposable
corresponding improvements in signature understanding battery.
and processing they will allow future systems to increase This is a very active field of engineering and significant
greatly the versatility of short-range surveillance radars. improvements are expected over the next few years as the
commercial imperatives of computer and automotive
Antennas systems force battery development.
The radar antenna is a key component that directs the It is also important to manage power with the utmost

Table 5 : Military battery parameters

Capacity, Ah Weight, kg Size (L x W x H), inches

186 ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000

care to ensure that it isn’t being used when not needed and hardware. These and other signal processing issues are
that processing algorithms are designed in a way that considered next.
makes them as power eMcient as possible. This is an
excellent example of the importance of taking a holistic Signal processing
view of the radar system design to optimise performance in
all operational conditions. Ever increasing signal processing power from ever more
efficient devices has been a major technology trend for
Display technology and human computer interaction many years and is set to continue for the foreseeable
Display technology is advancing at a considerable rate, future, driven again by the requirements of the mass
largely driven by the demands of the mass consumer consumer market. The consumer market may not,
market. Research prototypes of light-weight, low-power- however, progress the processing power per watt as fast as
consumption high-resolution flat screens have been it increases processing power per chip. A telling example is
developed and there is the possibility of ultra-thin liquid the increasing need for on-chip fans as the power
crystal display (LCD) screens that could be rolled or folded consumption of the PC processor rapidly expands.
for ease of storage and carriage. Clearly these forms of There is some hope for improvements in this area via
display have to be tailored to the operating environment to laptop PC developments, but the pace of development of
ensure that they can be viewed in all likely conditions. The processing per watt can be illustrated by the example of an
way in which data is displayed and the method by which update to an existing radar, currently using ten year old
the operator interacts with the radar system are also key technology. In this case the processing available (for the
factors in ensuring that the system meets its overall design same power consumption) by using all of the modern
goals. This may become simplified through the adoption of technological advances has only increased by a factor of
PC standards such as Windows NT, although it is not clear five. This contrasts with the processing power per chip,
that this necessarily provides a route to the most efficient which has increased over the same period by a factor
system performance, nor will the commercial human between 30 and 100. Clearly, when prime power is at a
interface necessarily be well suited to operation in extreme premium, this issue is of fundamental importance.
conditions (for instance, operation in the pouring rain, at Of much more importance is the move from function-
night, in a foxhole, and under attack may not be the best specific hardware to general-purpose processing.
time to be finding the right submenu or fighting the Although this seems at first to be just another way of
mouse). achieving the required process, it does offer a qualitatively
The battlefield of the future will exploit digital different approach.
technology and next-generation system designs will need In this new scheme a pre-process operating at the signal
to consider carefully electronic forms of reporting and processing rate can cue a much more complex algorithm,
communicating with the commander. It will also be but at the much lower information rate. Critically, because
important to examine interoperability issues, particularly of the availability oilarge memories accessible to the signal
as most future conflicts are likely to be coalition based. processor devices, these ‘cued’ complex algorithms can
now work on the raw data stored while the cueing
Implications of technology advance algorithms operated.
Overall, the above advances in technology are enabling This ability to mix algorithm types on a single
future radar systems to function at longer ranges in processing engine allows the designer to use much more
smaller, lighter more cost-efficient packages offering a sophisticated radar algorithms without greatly increasing
much more versatile set of operating modes. For example, the power consumption, size or weight.
the man-portable radar of tomorrow will be lighter, easier
to carry and able to operate at longer ranges but able to Target detection
manage its radiated power so that the likelihood of hostile As discussed earlier, in the principal mode of operation
intercept is minimised. Further it will be able to adjust its of current systems targets are distinguished from their
operating parameters so that they are automatically background by virtue of differences in the magnitude and
tailored to the task in hand. An example might be using a radial Doppler velocity of the backscatter. When the radial
lower resolution wide-area surveillance mode to velocity of the target with respect to the background is
automatically cue a high-resolution classification mode high, this is not too demanding. However the radial
towards detected targets of interest. Naturally these radar velocity of many targets of interest will naturally be low, for
attributes translate into advantages in vehicle mounting example a man walking or a vehicle moving close to a
where real estate for sensors is at a premium but demands cross-radial direction. Consequently the resulting Doppler
on performance are equally severe. The likely future velocity may be comparable with the background,
performance of these radar systems also makes them particularly if the latter consists of wind-blown vegetation.
prime candidates for mounting in either manned or If this is the case complex adaptive detection schemes
unmanned air platforms. The advantages are that the field must be employed that take account of the unpredictable
of regard is greatly enhanced and also that the forward and ill-behavecl nature of the background clutter whilst,
motion of the platlorm can be used to add a synthetic simultaneously, positively exploiting the characteristics of
aperture imaging mode to the current moving-target target signatures, the overall objective being to maximise
detection mode without major changes to the basic radar target detection whilst minimising false alarms. Fig. 5.

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000 187

Fig. 5 Example of adaptive
thresholding: (a) real data;
(b)targets found using
constant thresholds;
(c) targets found using
adaptive thresholds (CFAR -
constant false alarm rate)

shows a comparative example of fixed and adaptive resolutions the nature of the clutter returns will change.
thresholding. In the fixed case the threshold removes all The energy backscattered will be reduced but its nature
the clutter but conceals two targets with radial velocities will become progressively more [Link] requires
close to that of the clutter. However, in the adaptive case, a detailed understanding so that detection algorithms can
where prior knowledge of clutter behaviour is embodied in be adjusted accordingly. Various empirical and statistical
the algorithm, all the targets are detected with no false models have been proposed. One of these, the compound
alarms. This leads to a less confusing radar display, and form of the K-distribution, has been successfully used to
allows the operator to concentrate their attention on describe the amplitude and correlation properties of
evaluating targets rather than being distracted by false clutter5, However, when man-made clutter dominates a
alarms. resolution cell the K-distribution model breaks down and
It is well recognised that the behaviour of the clutter is a an alternative must be used. The fact that more than one
function of both environment and radar specification. As model is needed to represent all lorms of clutter illustrates
future systems may be capable of operating at higher radar the complexity of the operating environment.

Fig. 6 Time records of

radar targets

188 ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000

Understanding this
environment and target
signatures is one of the
major challenges for the
future. Overall detection
performance at high
resolutions will generally
improve as there will be
less clutter in a resolution
cell competing with the
target. However there
will come a point where
the target will start to
traverse resolution cells
during the period of the
processing interval and
this will cause its energy
to be smeared out;
consequently detection
will not be enhanced.
The detailed relationship
governing the detect-
ion performance under
these conditions is
also the subject of on-
going research but the
performance of prototype
systems is extremely
encouraging.

Target classification
There can be little
doubt that if detected
targets could be classified
into type, such as men
and vehicles, the utility of
the information would
be greatly enhanced, Fig. 7 Spectrograms (Doppler frequency against time): (a) of man walking; (b) of tank
enabling the battlefield
commander to form a much better understanding of may be built into automatic or semi-automatic algorithms,
enemy dispositions. The utility would be improved still either to make the classification decision or to cue the
further if the type of vehicle, for example, could also be operator to targets 01greatest interest. Recent research
discerned. Having detected a target of interest one method has taken this form of classification a stage further.
of classification is to use the magnitude of the time-varying To understand the machine view of the target returns we
backscatter to modulate an audio system and provide an must view them with the processing available to the
acoustic representation of the target Doppler signal to the machine. In general the Doppler radar sees the target as a
operator. Experience with ZB298 and MSTAR show this to series of Doppler spectra against time. This representation
be effective in certain conditions but overall the of typical targets of interest is shown in Fig. 7. The two
performance is variable and the relationship between spectrograms clearly show the difference between a radar
signature and classification is ill understood. return from a man walking and a vehicle (in this case a
Doppler classification: Fig. 6 shows two examples of the tank).
time-varying intensity (as might be used to modulate a Of particular interest for our understanding of this form
speaker) of a single resolution cell of a target return over a of classification is the periodic nature of the returns from
period of a few seconds. It is possible to observe gross men (Fig. 7a), representing the movement of arms, legs
movement of the target as evidenced by the modulation etc.; the returns from a vehicle (Fig. 7b) have much less
and it is this that the human listener hears and uses to periodicity, representing a relatively smooth motion. In
make the classification decision. The experienced addition to the longer term modulation behaviour of the
operator is able to make quite subtle distinctions between backscattered radiation there is important target-related
differing target modulations. Research work is attempting information contained in the shorter-term fluctuations. For
to identify the key classification criteria in order that they example, reflections from wheeled vehicles may exhibit a

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL. AUGUST 2000 189

characteristic Doppler signature, enabling them to be the smaller land-based targets provide a further challenge
distinguished from other targets, such as tracked vehicles. due lo the greater variability of target types and
Machine classifiers working on short-term Doppler background environments. Early research results are
classification have been shown to be capable of encouraging but algorithms are far from being either
distinguishing classes of vehicle, and capable of achieving optimal or robust and target signature understanding is
this with very short bursts of information (such as could be still a relatively immature science with much further work
gathered as the scan of a radar passes a target -a few tens required. For example, little research has yet been carried
of milliseconds). out to establish the trade-offs between algorithm
High range resolution classification: Another technique performance, processing demands and power
for target classificationis to increase the spatial resolution consumption. The role of the radar operator is key in
by increasing the radar bandwidth. If the radar resolution overall systems operation and must be included in
is sufficiently high it is possible that the target can be performance trade-offs.
classified spatially. The simplest approach is to form a Target classificationis a statistical technique, and future
‘range profile’, which is in effect a one-dimensional target systems are likely to employ multiple methods to classify
signature formed from a series of sequential range gates. It targets in an attempt to make the results more robust.
is one dimensional in the sense that the second dimension Current developments in Doppler and high range
is bounded by the antenna beamwidth, which is relatively resolution techniques will complement each other in
large and therefore provides little or no additional future ground surveillance radar systems.
information. The range profile will contain backscatter
from the different parts of the target under interrogation. If Imaging for target detection and classification
the nature of the profile is characteristic then it is possible II there is relative motion (or angular change) between
to perform classification. The simplest method would be to the radar and the target or area of interest then an image
build a library set of ‘known’ targets and compare the can be formed by synthesising a virtual antenna that
profile under test with the library sets. Library sets can be results typically in a cross-range resolution that matches
compiled either from measurements or from simulation. that of the radial direction. Two of the more common forms
Pattern recognition techniquesfor classification:Standard of this technique are known as ISAR (inverse synthetic
pattern recognition or matched-filtering techniques can be aperture radar) and SAR (synthetic aperture radar).
used to perform the classificationfor either the Doppler or In ISAR relative motion is provided via the velocity or
high range resolution technique. For Doppler rotation of the target. However, if the motion of the target is
classification, the simpler techniques appear adequate to unknown this results in defocusing and uncertainty in
achieve an acceptable and robust performance. the scaling of the image. Techniques aimed at
For improved performance in high range resolution removing these imaging errors have been developed
systems, more advanced neural network or genetic for ship targets but their successful application to
programming approaches can be employed. This has been smaller land targets is still in its infancy.
demonstrated successfully for maritime applications but An obvious extension to the military utility offered by
this class of radar is to mount it on an
airborne platform (such as an unmanned
aerial vehicle or a helicopter), which affords
a much less inhibited view of the battlefield.
In addition, the forward motion of an
airborne platform can be used to synthesise
a large antenna and achieve high resolution
in two dimensions. This is known as SAR
and typically the direction of view is
perpendicular to that of flight of the air
platform. The attributes of the radar
systems described here lend themselves
ideally to this form of imaging. The
techniques for producing a correctly
focused image are relatively mature and
have been demonstrated to resolutions of a
few metres; in principle it is possible to form
an image with a resolution of the order of 1
m or less. Fig. 8 shows an example SAR
image. At very high resolutions an image of
a target can be formed as a basis for
classification. This is the subject of intense
worldwide research and is likely to lead to
techniques that can be used reliably and
Fig. 8 Example of SAR imagery effectively in operational systems.

190 ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL AUGUST 2000

 Airborne short-range surveillance radar systems can be
used in both moving-target and imaging modes to provide Chris Baker graduated lrom the
a more comprehensive surveillance capability. Indeed they University of Hull in 1980 with a
can be used simultaneously so that the moving targets can first class honours degree in
Applied Physics. After completing a
be overlaid on the SAR image, thus providing a valuable PhD in laser physics (also at Hull)
context in which interpretation decisions can be made. Not he joined DERA (formerly RSRE)
surprisingly there are penalties to be paid, largely in the in 1984. Initially, Chris worked in
maritime radar, pioneering pulse
form o l cxtra system complexity and consequently in cost. by-pulse analysis techniques aimed
For example, the rate at which data is produced, at characterising electromagnetic
particularly by imaging radars, is somewhat prodigious backscatter from the sea surface. In
1990he was appointed leader of the
(up to several hundred megabits per second) and a means Battlefield surveillance research
of communicating this to the ground for exploitation must group and played a key role in the
form part of the overall system. This of course also implies development of airborne radar techniques, including ultra-high-
resolution SAR, SL4P and interferometry. This culminated in the
a ground segment and operations will require mission highly successful MODprocurement programme ASTOR. Chris is
planning and a much higher level of all-round currently head of short-range radar researc lising in
sophistication and support. Clearly the increased seeker sensors and surveillance and target ac systems.
Chris is a Fellow of the IEE, is chairman of professionalgroup E15
capability olYered has to be carefully balanced against the (Radar, Sonar and Navigation) and is a member of the organising
increased complexity and system costs. Nevertheless the committee for RADAR 2002
potential for using small lightweight coherent radars to
Address: PE 304, DERA Malvern, St. Andrews Rd., Malvern,
perform many functions is considerable and likely to grow Worcs. WR14 3PS, UK.
in the near future. Implicit reductions in production costs Email: cbaker@[Link];[Link]
should increase further the range ol applications and
Barry Trimmer graduated with a
accelerate developments in this important category of degree in Physics from Warwick
radar. Applications o l SAR imaging include 3-D mapping, Universityin 1978and was awarded
flood management, oil exploration, volume estimation, a Master’s degree in Astronomy by
Sussex University in 1979. He
treaty verification, surveying, urban planning, ocean joined the radiation laboratory at
monitoring and coastal erosion monitoring. EM1 Electronics, Hayes, working
on radar antennas for the
Searchwater radar and naval ESM
systems. During the 198Os, he
5 Summary developed RF and system
modelling within EMI, leading to
system design of ground
In this paper the essential design parameters of short- surveillance and weapon locating
range battlefield surveillance radars have been outlined
and examples of operational systems given. Against this
backdrop current and likely future technological and
signal processing advances have been described pointing
the way to future system developments. Rapid advances
are impacting almost every aspect ol the system and are IEE and a chartered engin
quickly increasing the capability ol man-portable
battlefield radars whilst simultaneously finding application Electronics, Manor Royal, Crawley, West Sussex,R H l O
in vehicle and air platform based surveillance. The quality Email: [Link]@[Link]
of information about target type and status is likely lo
enable the operator to perform increasingly wide ranging
and sophisticated tasks, so enabling the battlefield
commander to gain a more comprehensive understanding References
of his situation. This trend is set to continue. As costs
continue to be driven downwards it is likely that additional 1 SKOLNIK, M. I.: ‘Introduction to radar systems’ (McGraw Hill,
1980)
commercial applications, from harbour surveillance to the
2 S I M S O N , G. W..: ‘An introduction to airborne radar’ (SciTech,
protection of buildings and other lacilities requiring
1998,2ndedn.)
security measures, will emerge. The range of applications 3 NATHANSON, I;. E.: ‘Ibadar design principles’ (McGraw Hill,
clearly increases greatly with the ability to form SAR 1969)
imagery, although airborne systems will be accompanied 4 WATTS, S.,‘I’IUMMER, E., PRIESTLEY, B., and BAKEK, C. J.:
by higher manufacturing and operating costs. ‘Battlefield surveillance radars’ (Defence Systems
Internalional, 1996)
5 WARD, K. I)., BAKER, C. J., and WATTS, S.: ‘Maritime
Acknowledgments survcillance radar part 1: Radar scattering k o m the ocean
surface’, IEE Proc. F, Radar Signal Process, 1990,137,pp.51-62
The authors would like to thank Professor Miltc Dean and
Dr. Simon Watts for their helpful comments and assistance OIEE:2000
during the writing of this paper. Received 11th Januaiy 2000

ELECTRONICS & COMMUNICATION ENGINEERING JOUKNAL AUGUS1‘ 2000 191