RADAR

INTRODUCTION TO LECTURES ON RADAR

These 5 lectures of the Advanced Communications module cover an introduction to the basic
principles and applications of radar.

In Section 1, we shall start off with a discussion of the various applications of radar before
describing briefly the principles of operation of radar.

In Section 2, the Radar Equation will be introduced. This is an approximate, but useful,
method to calculate the range over which you can transmit and receive signals. We will need
to take into account various factors such as the antenna gains, transmitter power, losses etc....
The need to use transponders and beacons will be discussed which is relevant to Secondary
Radar which will be described later on.

In Section 3, the basic principles of Primary Radar will be described and the problems
encountered in determining the position and speed of the target of interest will be discussed.

In Section 4, Secondary Radar will be described ; this type of radar is used extensively in Air
Traffic Control, and emphasis will be given on the communication that takes place between
the ground based radar and the aircraft's transponder.

Finally, in Section 5, a discussion of other applications of radar will be briefly described.

The emphasis in these lectures is in carrying out calculations to choose the radar parameters
(eg : transmitter power, pulse repetition frequency) for a given application.

1. INTRODUCTION TO RADAR

1.1 Applications

There are many applications of radar, some of which have direct relevance to our everyday
lives. Examples of these are :

Civil Applications

Air Traffic Control

Sea Traffic Control
Speed Traps 
Intruder Alarms
Secondary Radar

Navigation (Sea, Air, Ground)

Weather Monitoring
Monitoring of the Sea State
Locating Historical Relics (Archaeology)
Distance and Rotation of Planets
Imaging planetary surfaces
Earth Resources
Mapping - 900 m error in the position of Corsica on the map !!!
etc...

Military Applications

Detection of Own or Enemy Forces

Guidance
Detection of Missiles
Espionage

1.2 Radar Operating Frequencies

The bands of operating frequencies commonly used are as shown in the following Table
(From Electronic Communication Systems 3rd Ed., George Kennedy, Mc Graw Hill, 1987).

BAND NAME FREQUENCY RANGE MAXIMUM AVAILABLE

APPROX (GHz) PEAK POWER PER TUBE
(MW)
UHF 0.3 - 1 5
L 1.0 - 1.5 30
S 1.5 - 3.9 25
C 3.9 - 8.0 15
X 8.0 - 12.5 10
Ku 12.5 - 18.0 2
K 18.0 - 26.5 0.6
Ka 26.5 - 40.0 0.25
V 40.0 - 80.0 0.12
N 80.0 - 170.0 0.01

Note that the powers given in the third column are per tube in the transmitter; several tubes
could be used together to multiply these power values.

The physical principles behind a lot of the above applications are the same. Essentially, we
are trying to extract some or all of the following basic parameters of the target: position and
velocity. We would like to extract these parameters with a given accuracy. However,
before we can discuss methods to do this, we need to make sure that we can detect the
target(s) of interest in the first place; this leads us to discuss the Radar Equation.

1.3 Radar Equation

The process of transmission and reception of radar energy is shown in the following Figure:

Figure 1

The question that we are going to ask is the following: For a given transmitted power, what
is the maximum range such that we can detect targets? The answer to this question is very
complicated; we are going to present a simplified analysis which gives useful, if
approximate results. The letter enclosed in square brackets [ ] after each equation signifies
the units of the quantity of interest.

At Radar

Let us suppose that the transmitted peak power is Pt Watts.

Between Radar and Target

Travelling out from the radar, at distance r, the Power Density, Pden is the transmitted power
divided by the surface area of a sphere of radius r :
 (1)

Up to this point, it has been assumed that the power is transmitted isotropically , that is the
same power in all directions. In practice, transmission will be by an antenna, which will
have directivity, eg:

If the gain of the antenna transmission is G, then we modify Equation (1) as follows :

(2)

At Target

If the target of interest is at range R, then the power density at the target becomes :

(3)

Now any target has a radar cross-section, σ [m 2]. This need not be the same as the physical
cross-section that the target presents to the incoming radiation. In addition, σ will depend on
the aspect that the target presents to the radar (because of the target's geometry), and also on
the incoming frequency (from the target's geometry and the material making up the target).
Thus σ is not a constant with respect to time and frequency. σ is defined for the target such
that the reflected power, Pref , is given by
(4)
in which case, from Equation (3) :

(5)

At Radar Again

The power density of the radiation, Pincden incident at the receiver of the radar is given by :

(6)

What is the total power received by the antenna? Let Pinc be the power incident on the
antenna. Define, the effective antenna area, as

(7)

A' [m2] is related to the geometric area, A, of the antenna by :

(8)
where η is the antenna aperture efficiency (normally between 0.5 and 0.8).

Combining Equations (6) and (7) :

(9)

It can be shown from antenna theory that

(10)

where λ is the operating wavelength of the radar. Substituting for A' from Equation (10) into
Equation (9):

(11)
Now there are losses due to the electronic equipment used and due to atmospheric effects.
We can incorporate them into Equation (11) above by multiplying Pinc by a loss factor L,
which is less than 1:

(12)

Now, solving Equation (12) for R :

(13)

In Equations (12) and (13), the following are typical numerical quantities for the various
parameters :

Transmitted Power, Pt : Typically 1 kW to several MW

Antenna Gain , G: ~ 1000 ~10000 +
Target Cross-Section, σ : 0.1 m2 (eg : missile)to 100 m2
Operating Wavelength, λ : 0.03 to 0.3 m
Minimum Received Power, Pinc : ~ 10-15 W +
Loss Factor, Ls typically around 0.3 maximum.

Note that power, gain and loss factors are often expressed in decibels (dB):
PdB = 10 log10 P, GdB = 10 log10G and LdB = 10 log10L. Eg : if L = 0.3, then L dB = -5.2 dB.

Worked Example 1.3.1:

Calculate the maximum range of a radar system which operates at a wavelength of 3 cm with
a peak pulse power of 500 kW, if the minimum power that can be detected is 10-13 W, the
effective antenna area is 5 m2 and the radar cross section of the target is 20 m2. Ignore losses
(L = 1).

Solution : Looking at Equation (13), we see that we have been given most of the parameters,
except for the antenna's gain G. We can work this out from Equation (10) :

Substituting for G in Equation (13) and inserting other values that have been given :
----------------------------------------------------------------------------------------------
Problem R1

An L-band radar operating at 1.25 GHz uses a peak pulse power of 3 MW and must have a
range of 100 nautical miles (185.2 km) for missiles whose radar cross-section is 1 m2. If the
minimum power that can be received for target detection is 2 x 10-13 W, and the antenna's
aperture efficiency is 0.65, what is the smallest geometric diameter that the antenna can have
assuming that it has a circular cross section ? Ignore losses. [ Answer : 3.84 m]

--------------------------------------------------------------------------------------------------------------
--

2. PULSED RADAR

2.1 The Basic Set Up

We will now describe how pulsed radar can be used to determine the range, bearing &
elevation and radial velocity of a target.

The following diagram shows a pulse train transmitted by the radar :

The pulse width is τ and the time between pulses is T. The Pulse Repetition Frequency
(PRF) is defined by :

(16)

The Mark Space Ratio (MSR) is defined by :

(17)

Hence, the transmitter is turned on for a time τ. The transmitter is then turned off and the
radar operates in receive mode for a time T-τ . After this time the receiver is turned off and
the transmitter is turned on. It is important to make sure that the transmitter and receiver are
not turned on at the same time to avoid the whole system blowing up due to receive and
transmit signals being present in the radar at the same time ; a duplexer in the radar prevents
this happening.

The pulse train is turned on for a proportion equal to τ / T (the duty cycle) of the time during
a period. Hence, if Ppk is the peak power, the mean power for transmission, Pmean is given by:

(18)

Each transmitted pulse has a width τ. The bandwidth B of each pulse, ie the range of
frequencies for which the magnitude of the Fourier transform is significant, is proportional
to 1/τ . A rough rule of thumb is to use the following approximation :

(19)

There is a limit on the minimum value of τ , ie a limit on the maximum value of B, due to
limitations of the electronics in the transmitter.

2.2 Range Determination

The range of a target is determined by measuring the delay of a pulse between transmission
and reception. Refer to Figure 1.

Suppose that a target is at a range R. Each pulse has to travel a distance 2R from the radar to
the target and back to the radar again. Hence, if c is the speed of light, the time delay of each
pulse is given by :

(20)

Hence, the range, R, can be determined from

(21)

Resolution
The resolution is the minimum distance between two targets that can be distinguished by the
radar. The calculation of the resolution is a little complicated, but we will use an empirical
argument here.

Suppose that the time delays of a pulse sent out by the radar are T 1 and T2. Hence, the
difference in range ΔR = R1 -R2 is related to the difference in the time delays ΔT = T1 -
T2 by :

(22)

Now, if the two returning pulses overlap, and are superposed at the receiver, then it will be
difficult to distinguish when one begins and the other ends. Any such uncertainty will be
reduced, if the width, τ, of each pulse is reduced. Hence, the resolution will be increased by
reducing τ. However, from Equation (19) above, we can see that this means increasing the
Bandwidth of the transmitted signal and there is a limit on this.
Thus, as a rule of thumb, we can say that two targets can be resolved if the difference in the
time delays is greater than a pulse width :
(23)
Thus, from Equation (22), we can use as a definition of the resolution :

(24)

Of course, ΔR increases in the presence of noise.

Accuracy

The accuracy of range determination refers to how accurately we can determine the range of
a single target.

This will depend on the pulse width τ ; the smaller τ, ie the larger B, then the better will be
the accuracy in range. A semi-empirical formula for the accuracy, δR, is given by :

(25)

where SNR is the signal to noise ratio. In all systems, SNR is non-infinite, so δR can never
be exactly 0. The larger the noise power, then the smaller SNR, and the larger the error δR.
For a given SNR, from Equation (25), it can be seen that the error goes as the inverse of the
bandwidth.

Range Ambiguity

Can we always determine the range of a target unambiguously when using pulsed radar ? To
see this let us look at an example :

Example 2.2.1:

Suppose that the following pulse train is transmitted :

Figure 2

Suppose that the PRF is 1 kHz, meaning that T = 1 ms. Let the pulse width, τ , be 5 μs.
Suppose the target is at a range of 140 km. Thus, from Equation (20), the time for a pulse to
travel from the radar to the target and back to the radar again is given by :
Example 2.2.2:

Now suppose another target is at a range of 160 km. The time delay, from Equation (20) is
given by :

But there is a problem. In this case, Tdel is larger than the time between pulses T (= 1 ms).
Hence, another pulse has been transmitted before the first pulse has returned from the target.
Hence, we do not know whether the pulse that we have detected is the first pulse or the
second pulse. If the return is assumed to have come from the second pulse, then the delay is
taken to be 1.067 - 1 ms = 0.067 ms. The calculated range is then

which is wrong. Hence, for a given range R the minimum time between pulses for
unambiguous determination of the range is given by

(29)

The larger the time between pulses, T, then the larger is maximum range R that can be
determined unambiguously by the radar.

One way round this problem is to send out two sets of pulses with different PRF's :

Example 2.2.3 : Target at 160 km. Suppose that we use 2 sets of pulse trains, one with time
between pulses, T1 = 1 ms, the other with time between pulses, T2 = 0.8 ms.

Now we have seen that a pulse sent out from the radar will be delayed by 1.067 ms. We have
also seen that , when T1 = 1 ms, the delay could also be 0.067 ms from the second pulse.
For the set of pulses with T2 = 0.8 ms, the delay will be either 1.067 ms (from the first pulse)
or 1.067 - 0.8 = 0.267 ms from the second pulse. In both cases, 1.067 ms is a possibility,
hence the return is deemed to have come from the first pulse and the range can be calculated
correctly as 160 km.

Minimum Range

The transmitter turns on for τ seconds when transmitting a pulse and then is turned off for T-
τ seconds; during this time the duplexer turns on the receiver to detect any returning pulses.
In particular, we cannot receive any pulses during the transmission of a pulse. Hence, we
cannot receive pulses that have been delayed by less than τ seconds. This corresponds to a
minimum detectable range given by
 (30)

In our example, τ = 5 μs, hence

(31)

Targets closer than 750 m will not be detected.

2.3 Determination of Radial Velocity

It is also possible to determine the radial velocity of a target using pulsed radar. The radial
velocity is a useful measurement to make as you can then use it to predict where the target
will be at the next update before receiving the next return from the target.

The radial velocity is determined using the Doppler Effect.

If a radar sends out a pulse with frequency f0:

(32)
then, if the target is travelling towards the radar with a radial component of velocity V, the
return pulse will be shifted in frequency :
(33)
where, fd , the Doppler frequency which is given by

(34)

Hence

(35)

This is mixed with a reference signal stored in the receive which has the same frequency as
the transmitted pulse:

(36)
Using the trig identity cos(A).cos(B) = ½ cos(A+B) + ½ cos(A-B), sr(t) may be rewritten as :

(37)
The first terms is at a frequency 2f0 + fd whilst the second term is at a frequency of fd . The
first term may be filtered off with a low-pass filter leaving us with the second term, which is

(38)

Now, the pulse train is effectively sampling this signal. If T is the time between pulses

(T = 1/PRF) , then the nth pulse will occur at a time of nT. Hence, the set of samples that are
being processed is given by {s(nT)}, where the nth sample is given by

(39)

These samples could, for example, be fed into a processor that computes the Fourier
transform of the signal which will give us the frequency f d. Having found fd , V can be found
from Equation (34 ) as :
(40)

The Fast Fourier Transform (FFT) is a convenient way to compute the Fourier transform
digitally ; the restriction here is that the number of samples should be a power of 2 e.g. : 32,
256, ..... By delaying the input signal by 90 degrees, it is possible to make the input signal
(39) complex as follows (putting A = B/2 for convenience):

It can then be shown (see Appendix 1) that, for this complex signal, it is possible to
distinguish between positive shifts in the Doppler frequency (i.e. target travelling towards the
radar) and negative shifts in this frequency (i.e. target travelling away from the radar). This
would not be the case if just sr(nT) in Equation (39) were to be processed – for further
details, see Appendix 1.

Resolution

The resolution, which is the difference in range rate that can be measured between two
targets using pulsed radar, depends on the method used to determine the Doppler frequency,
fd , from Equation (35). If FFT methods are used, it can be shown that the resolution, V, is
given by:

(41)
where n is the number of pulses fed to the FFT processing and the pulse repetition frequency
is PRF; this result is derived in Appendix 2 , where more details of FFT methods are given.
The Accuracy in the determination of V depends on the method used to calculate the radial
velocity and also on the signal to noise ratio. Using an empirical formula similar to Equation
(25), the accuracy V in the measurement of V is given by
(42)

Ambiguity

In unfavourable situations, the determination of radial velocity can be ambiguous. To see

this, note from Equation (39) that we are processing a sampled cosine wave. Now,
Shannon's Sampling Theorem states that if one is sampling a cosine with frequency f d , then
the maximum time between samples should be given by :

(43)

For further details of sampling, please see Appendix 3.

If T is larger than this value, then one is undersampling the cosine signal, which leads to
ambiguities in the determination of the frequency. It turns out (see Appendix 3) that, in this
case, one could not be able to distinguish between a signal with the following frequencies:

(44)

Substituting for fd from Equation (34) into Equation (43) for a given V:

(45)

The larger V, the larger the Doppler frequency then the smaller is the maximum time
between pulses, Tmax , that is allowed.

As for range determination, radial velocity or range rate ambiguities can be overcome by
using two sets of PRF's.

Example 2.2.4 : Let us revisit Example 2.2.3. with two pulse trains, one with time between
pulses, T1 = 1 ms, the other with time between pulses, T 2 = 0.8 ms. Suppose that the radar is
L-band and operating at 1 GHz and the target speed is 500 ms-1. It is known that the
maximum target speed is 1000 ms-1. From Equation (34), the Doppler frequency is given by
:

(46)

The maximum Doppler frequency, corresponding to the maximum possible radial velocity of
1000 ms-1 is given by
 (47)

Now from Equation (44), we see that when T1 = 1 ms, corresponding to a PRF of 1000 Hz,
the following frequencies are possibilities (taking into account the maximum value of f d that
the target can have) :

6333, 5333, 4333, 3333, 2333, 1333, 333 Hz

When T2 = 0.8 ms, corresponding to a PRF of 1250 Hz, the following are possible
frequencies:

5833, 4583, 3333, 2083, 833 Hz

Looking at these two sets of values, we see that 3333 Hz is the only frequency in common,
so this must be the Doppler frequency.

2.4 Range and Range Rate Ambiguities

The question now to be asked is whether one can simultaneously determine unambiguously
the range and range rate of a target. Equations (29) and (45) will be used here.

Example 2.4.1
Suppose that one wishes to unambiguously determine a target's range out to 150 km and one
also wishes to determine the radial velocity up to 300 ms-1 . Suppose that the radar is L-
band and is operating at 1 GHz. Can we choose a PRF such that range and radial velocity
can be determined unambiguously?

Equation (29) implies that the minimum time between pulses should be given by :

(48)

for unambiguous determination of the range. Equation (45) implies that the maximum time
between pulses should be given by :

(39)

Clearly, there is no value of T that is going to satisfy both conditions. If, for example, T was
chosen to be 2 ms, corresponding to a PRF of 500 Hz, then from Equation (48), we would
obtain unambiguous determination of range, but from Equation (49) ambiguous radial
velocity determination would result.

Radar which use a relatively high PRF, are called pulse-Doppler radar because they can
unambiguously determine radial velocity but not range. Those radar which use a relatively
low PRF, can determine large ranges unambiguously, but not radial velocity. Intermediate
PRF radar have both range and Doppler ambiguities.

As mentioned above, one can avoid range and Doppler ambiguities by using two sets of PRF
values to sort out the ambiguities; this is usually done at the intermediate PRF mentioned
above.

2.5 Noise

The radar equation (13) does not explicitly take into account noise. There are many sources
of noise. External factors are interferences from galaxies, from the atmosphere and, in
military situations, jamming. The electronics in the radar produce their own noise due to the
random thermal motion of the electrons. Define the received noise power as N watts.
Hence, from Equation (12), the received signal to noise ratio is given by :

(40)

Rewriting this in terms of the target range, R :

(51)

A typical value for SNR is 13 dB for reliable detection of signals. The noise power
accompanying the incoming signal can be approximated by:
(52)
where k is Boltzmann's constant (1.38 x 10-23 J K-1), B is the bandwidth of the signal (Hz)
and θ is the temperature in degrees Kelvin. However, the receiver's electronics increases the
amount of noise. This can be taken into account by multiplying N0 by the noise figure of the
receiver, F, so that N in Equation (51) is given by :
(53)
Note that the minimum value of F is 1. Sometimes this is expressed in dB as F dB = 10log10 F.

Hence, Equation (50) can be rewritten as :

(54)
Now let us specialise to the case of pulsed radar. We can increase the effective peak power
by integrating over several pulses. If n pulses are processed, then it can be shown that the
effective peak power is nPt. In this case, the SNR in Equation (54) is given by:

(55)

Finally, this equation can be rewritten in terms of the mean transmitted power, Pmean . It can
be worked out from Figure 2 that if Pt is the peak power and Pmean is the mean power, then

(56)

Substituting for Pt in Equation (56):

(57)

or in terms of R :

(58)

Equations (57) and (58) are two alternative equations for the radar equation when pulsed
radar is used.

It should be noted that these are not exact equations ; they give a good rule of thumb when
designing a radar system. In practice, these equations are used before more sophisticated
design techniques are used.

Equations (57) and (58) also tell us about the factors limiting range determination. For
example, in Section (2.2), it was mentioned that using a large bandwidth radar means that we
can obtain better range resolution. However Equation (58) suggests that this reduces the
maximum range that can be determined ; the reason for this is that a larger B introduces more
noise power, through Equation (53), hence there is a trade-off here when choosing B between
resolution and received signal to noise.

2.6 Determination of Elevation and Azimuth

The radiation transmitted by the radar is directional ; a typical beam pattern is shown
below :

Figure 3

The theory behind this is covered in the lectures on antennas. An approximation to the
beamwidth of the mainlobe of an antenna is :

where is the wavelength, d is the spacing between dipoles and N is the number of dipoles.

Radar can operate in two basic modes : surveillance and tracking. In surveillance mode, a
single beam or several beams are sent out and rotated about the space. This rotation is
caused either by mechanical rotation of the radar or by electronic beam steering (covered in
the lectures on antennas). The aim here is to sweep the whole sky looking for targets. In the
case where the radar rotates mechanically, the beam is steered continuously to scan the space
for targets.

There are two conflicting requirements when determining the beamwidth for transmission. If
the beam is too narrow then several directions will need to be looked at per scan in order to
cover the space of interest. The returns from each direction need to be updated every 10
seconds or so in order to follow manoeuvring targets. Steering a beam this quickly in 10
seconds, looking at all directions, puts considerable strain on the radar's resources. If the
beam is too broad, then although less beams are necessary to sweep all of space, the angular
resolution will be very poor, that is azimuthal and elevation measurements will have poor
accuracy.

To quantify the problem, refer to the following diagram:

Figure 4

Here, the beam is being rotated once in the azimuthal direction every T seconds. Suppose
that the width of the beam is radians. Hence, the dwell time DT on the target is given
by:

Now, pulses are being transmitted continuously with the time between pulses being given by
1/PRF. It is required to process n pulses for the return from each target to increase detection
probability and to determine the radial velocity. Hence, the time of transmission of n pulses
must be less than the dwell time DT of the target; otherwise we will receive less than n
pulses.

Hence

If the PRF is fixed, then n can be determined from:

Once a target is detected, then one uses a tracking radar. In this mode, the radar has locked
on to a single target and is dedicated to tracking it. If the target moves out of the main beam
then it is automatically steered towards the target again. The transmitted beam in such a
radar is usually narrow in order to (i) obtain good angle resolution (ii) obtain a high gain in
the chosen direction (iii) minimise the effects of unwanted returns appearing in the main
beam. There are various methods that have been adopted to track targets, but one popular
method is to use monopulse methods. One monopulse method that one can use is the
amplitude comparison method.

Amplitude Comparison Methods

To illustrate the working of this method, suppose we wish to determine the azimuth of a
target. Two beams are sent out by the radar which are slightly offset from each other. If we
subtract the envelopes of the two beampatterns, then we obtain the difference beam pattern
shown in Figure 5. If we add the envelopes, we obtain the sum pattern shown in Figure 5.
In this particular example, d/λ = 0.25, N = 32 and the two beams that are used point at +0.5 0
and -0.50.

Figure 5
The sum beam is used to detect the target and to determine the range and range rate using
the methods described in Section (2.2) and (2.3). As this beam is formed by summing two
individual beams, the noise becomes averaged out leading to improved signal to noise. This
beam cannot be used to determine the azimuth accurately, however, as it is too broad.

The difference beam is used to find the azimuth. If there is no return from the difference
beam, then the azimuth of the target is at 0 degrees. If the target is off the bore sight, then
there will be a non-zero return detected by the difference beam. Note that, for small angular
deviations, the voltage is proportional to the angular error ; in the above example beyond
about 3 degrees, this calibration cannot be used. The magnitude of the return will give the
angle, the sign of the return will give the direction with respect to the bore sight. Ideally, the
return from the difference beam should be 0 indicating that the beam is pointing straight at
the target. If a non-zero return is detected by the difference beam, then this information is
used to steer the antenna beam towards the target again. In addition, if one has been tracking
the target for some scans, then one could predict ahead where we expect the target to be and
steer the beam accordingly ; digital signal processing methods have been developed to do
this.

If one wishes to extend this to looking at elevation, then the following arrangement can be
used :

Figure 6

Three beams are formed as follows:

Beam 1 : A + B + C + D
Beam 2 : A + B - (C + D)
Beam 3 : (A + C) - (B + D)

Here, the target is detected and the range and range rate are found by looking at Beam 1.
The return from Beam 2 will determine the azimuth and the return from Beam 3 will
determine the elevation.

Ideally the target will be in the bore sight of the radar and the returns from Beam 2 and Beam
3 should be 0. Non-zero returns from either or both of these beams can be used to steer the
beam back towards the target.

One of the advantages of monopulse methods is that the beams are sent out simultaneously
hence one can overcome the problem when the target radar cross-section is varying rapidly in
time. In other tracking methods, one processes the returns from several scans which could
cause difficulty if the amplitude of the returns vary with time because of the rapidly varying
target cross-section.

----------------------------------------------------------------------------------------------

Problem R2
A radar is required to have an unambiguous range of 400 km. What is the maximum PRF
that may be used ? [375 Hz]
If the pulse length is 3 μs, what is the bandwidth of the signal ? [333.3 kHz]
What is the minimum range that can be determined ? [450 m]

Problem R3
An L-band surveillance radar operates at 1.25 GHz. The desired range rate resolution is
1 ms-1 and the maximum speed of the target is 1000 ms-1. The minimum PRF is used such
that the determination of range rate is unambiguous for all possible speeds of the target.
What is the minimum number of pulses that is needed to obtain this resolution ? Assume that
you are using FFT methods. [2048 pulses]

Problem R4
A particular surveillance radar has a beamwidth of 30. The antenna beam rotates every 10 s.
With the minimum value of the number of pulses and PRF determined in Problem R3, is it
possible to scan the target within the required time ? [No]

Problem R5
A guided missile tracking radar has the following parameters :

Peak Power = 1 MW
Operating frequency = 5 GHz
Signal to Noise ratio for detection = 13 dB
Noise Figure = 11 dB
Diameter of antenna dish = 3.66 m
Aperture efficiency = 0.65
Missile cross-section = 1 m2
Loss factor = - 5 dB
Beamwidth = 1 degree
Pulse width = 0.625 μs

Determine:

(i) The maximum detection range [119.3 km]

(ii) The PRF such that the maximum unambiguous range is equal to the maximum
detection range determined in (i) [1.257 kHz]
(iii) The minimum range that can be determined. [94 m]

Problem R6
The Italian Company Alenia manufactures the ATCR-44K, an L-band medium range
primary radar designed for use in modern air traffic control (ATC) systems. The parameters
are as follows :

(i) Transmitter power = 1.2 MW

(ii) Gain = 4500
(iii) Target cross-section = 2m2
(iv) Loss factor = -7 dB
(v) Minimum signal to noise ratio for detection = 10 dB
(vi) Operating frequency = 1.3 GHz
(vii) Noise figure = 3 dB
(viii) Pulse width = 3 μs

What is the maximum range for this radar ? [314 km]

Problem R7
This question requires the use of a calculator or computer (MATHCAD and MATLAB might
be particularly useful). It is meant to illustrate monopulse methods of determining angle.

Suppose that we are determining angle in one direction, say azimuth. You are given that the
beampattern from an array of N dipoles, spacing between dipoles = d, is given by:

with

where is a general azimuth and is the direction of the mainlobe. is the

wavelength.

Suppose that two beampatterns are produced, which have mainlobes which are at and
degrees. Given = 0.25 and N = 50 Determine the sum and difference patterns
when :

(i) = 0.2 degrees

(ii) = 5 degrees

In each case, comment on the effectiveness of the beampatterns produced to determine

azimuthal angle.

2.7 Track While Scan (TWS) Radar

In track while scan radar, the antenna is rotated mechanically whilst at the same time the
beam is rotated continuously in a circular plane. The range, bearing and range rate of the
returns from each scan are monitored. Software methods are used to plot out the track of
each detected target. The beamwidth is typically large in order not to miss any targets, but
the accuracy in the estimation of the position can be improved considerably using digital
signal processing techniques. The disadvantage if this method is that the rotation rate should
be slow enough so that the beam from the antenna is able to scan the whole of the space of
interest. However, the rotation rate should be large enough so that manoeuvring targets (ie
accelerating, turning) can be tracked accurately enough. We thus have conflicting
requirements on the rotation rate.

2.8 Phased Array Radar

The main beam of an antenna can be rotated electronically rather than mechanically ; the
former method does not suffer from problems with inertia of the antenna array hence the
main beam can be rotated much more quickly than in the case of track while scan.

A phased array radar, consists of a two dimensional array of dipoles which allows one to
scan elevation as well as azimuth.

The maximum array size is about 4900 dipoles. This could be a 70 X 70 array, but if, for
example, more resolution is required in the azimuth than the elevation, then a rectangular
array would be more appropriate (this is the case for Air Traffic Control applications - see
later).

Phased array radar are very flexible in their use. One could use different sub-arrays to send
out beams in different directions. Very high scan rates are possible compared with track-
while-scan. One could use the radar in both surveillance mode and tracking mode at the
same time by using different sub-arrays.. A central computer is used to ensure that efficient
use of radar resources is made. The radar can change modes (ie change the position of a
beam send out new beams) in a time duration the order of microseconds. Unlike TWS, one
can alter the scan rate for a particular beam according to the dynamics of the target ; for
example if the target manoeuvres, then one would wish to increase the scan rate in order to
track the target more accurately. If the scan rate were kept constant, as in TWS, one could
lose the target.

The main disadvantage of phased array radar and the reason why it is used mainly in
specialised military applications is its huge cost compared to the relatively cheap track-while-
scan.

3. BEACONS, TRANSPONDERS
So far, we have considered the case where a radar transmits pulses, these bounce back from
the target to the receiver and then we process these pulses to obtain position and speed
information about the target.

Let us revisit the Radar Equation again and look at an example of communication between
NASA and an Apollo spacecraft orbiting the Moon.

The following might be typical parameters for the radar used:

Peak Power = 0.5 MW

Operating Frequency = 2.5 GHz. Hence wavelength = 3x108 / 2.5x109 = 0.12 m.
Antenna physical diameter = 64 m
Antenna Aperture Efficiency = 0.65 (Parabolic)
Target Cross Section = 1 m2
Noise Figure of Receiver = 1.1 (high quality receiver)
Minimum signal to noise ratio for detection = 1 (low compared to terrestrial applications).
Receiver bandwidth = 5 kHz
Antenna noise temperature = 290 K
Ignore loss factors, loss factor, Ls = 1.

What is the maximum operating range of the radar ?

Let us use Equations (51) and (53) to calculate Rmax.

To calculate, G we use Equation (10) which gives the gain of the antenna in terms of the
effective aperture. The effective aperture is determined from the actual geometric area from
Equation (8).

Inserting numbers (where D = diameter of the antenna)

To determine G, we use Equation (10) :

Substituting for G and the other parameters in Equation (51) :

Thus the operating range of the radar is just over 27 000 km whilst the mean distance to the
Moon is about 380 000 km (!!!!!!!)

The problem can be seen by referring to Equation (51). The transmitted power received by a
target at range R goes as 1/R2. The radar's radiation is reflected back to the receiver (in this
case on earth) and the power received at the radar again varies as (1/R2 )2.

Whilst in terrestrial applications, R is of the order of tens of km, for space applications we
are talking about hundreds of thousands, if not millions of kilometres. Hence, due to the 1/R 4
factor the received power is too small for detection.

We could increase the transmitter power but this would be costly. The antenna used is
relatively large and could not be increased much further.

The solution to this problem is to place a transponder, otherwise termed beacon, on the
spacecraft as shown:

Figure 7

On receipt of a signal from the radar, the transponder turns on and transmits another signal
back to Earth. The strength of the transponder signal is independent of the signal power
received from earth, except that this latter power should be larger than the minimum power
detectable by the transponder. In this case, the Power received by the spacecraft goes as
(1/R2) ; this signal transmitted by the transponder goes as 1/R2 hence is stronger on reaching
the earth than the reflected signal from the previous example on the basis that for large R,
1/R2 > 1/R4 .

Let us analyse this situation in more detail.

Suppose that the radar and beacon have the following parameters :

RADAR : Peak transmitted power = PT , gain = GT , geometric area AT and aperture

efficiency, ηT . Noise figure FT , minimum signal to noise ratio for detection : SNRT
BEACON : Peak transmitted power = PB , gain = GB , geometric area AB and aperture
efficiency, ηB. Noise figure FB , minimum signal to noise ratio for detection : SNRB

The power density transmitted by the radar is given by :

Taking into account the gain of the antenna, then we modify the above to:

If the effective antenna area of the beacon is AB' then the power incident on the beacon is
given by :

AB' is related to the geometric area AB by

The gain of the transmitter antenna is given by :

The signal to noise ratio at the beacon is given by :

with the noise NB given by

where FB is the noise Figure of the beacon's antenna and θB is the antenna's noise
temperature.

Let the losses be lumped into a factor L1

Combining the above equations, we find for the maximum up-link range, R1:

(75)

Now if the signal to noise ratio at the beacon is greater than the threshold for detection, the
transponder will transmit a signal back to the radar. The maximum range, called the down
link range, may be derived using a similar argument to the above (exercise for student). It
may be shown that the down link range, R2 , is given by :
 (76)

where L2 represents down-link losses.

Hence the maximum range for communication between the radar and the beacon is given by
the minimum of R1 and R2 :
(77)
Now let us look at the Apollo example, only this time using a transponder mounted on the
spacecraft.

The radar's parameters are as before :

PT = 0.5 MW, = 0.12 m, D = 64 m, = 0.65, F = 1.1, SNRT = 1

The beacon's parameters are as follows :

PB = 50 W, = 0.12 m, D = 1 m, = 0.65, FdB = 13 dB, SNRB = 1

The antenna noise temperature (290 K) and bandwidth of transmitted signal (5 kHz) are
assumed to be the same for both the beacon and the radar. Ignore losses so that L 1 = L2 = 1.
The up-link range (R1) and the down-link range (R2) are now calculated from (75) and (76)
as

million km

Hence, from Equation (77), Rmax = min(R1 , R2) = 410 million km. This is much larger than
the mean distance from the Earth to the Moon, so communication should be possible.
However, it should be noted that the calculation that we have done is only very approximate.
The minimum signal to noise for detection at the beacon mounted on the spacecraft and at
the radar, could be much less than 1 if one integrates several pulses coming in as described in
Section (2.5). In addition, using longer pulses reduces the bandwidth B in the denominator
which increases the range. Increasing the size of the antenna on the transponder, for example
using a fold out arrangement would increase the range. More detailed calculations are
needed here, but our approximate calculations show that it is feasible to track the Apollo
spacecraft and indeed it is feasible with some redesigning to be able to track spacecraft to the
edges of the solar system.
-----------------------------------------------------------------------------------------------------
Problem R8

A radar is transmitting a signal to a beacon on a satellite which sends back a reply to the
transmitter on receipt of a signal from the beacon. The radar and beacon's parameters are as
follows :

Radar

Peak transmit power = 1 MW

Operating Frequency = 5.7 GHz
Aperture Efficiency = 0.65
Antenna Gain = 43 dB
Loss factor (radar to beacon) = -5 dB
Minimum Signal to Noise Ratio for Detection = 20 dB
Noise figure for transmitter = 1.8 dB
Pulse width for transmitted signal = 1 μs
Noise temperature = 290 K

Beacon

Peak transmit Power = 1000 W

Operating Frequency = 5.7 GHz
Aperture Efficiency = 0.65
Antenna Gain = 4.8 dB
Minimum Signal to Noise Ratio for Detection = 23 dB
Noise figure for beacon = 7 dB
Loss factor (beacon to radar) = -5 dB
Pulse width for transmitted signal = 1 μs
Noise temperature = 290 K

(i) What is the maximum range of operation ? [23 470 km]

(ii) If the beacon was not present on the target and the radar was receiving signals
reflected off the target, what would then be the maximum range of operation ?
Assume a target cross-section of 1 m2 and that the loss factor is - 5 dB. Compare
your answer to part (i). [130.4 km]

4. SECONDARY RADAR AND AIR TRAFFIC CONTROL

4.1 Introduction

The previous example of tracking of a spacecraft is an example of "Secondary Radar". The

"primary radar" transmits a signal. On reception of the signal on the secondary radar (in this
case mounted on the spacecraft) a transponder on the spacecraft sends back another message
to the primary radar. This has already been illustrated in Figure 7 above.

As mentioned before, one can track targets over longer distances with smaller transmitter
powers than if a single primary radar was used which replies on detection of signals reflected
off targets.

Secondary radar is also used in another familiar situation to us, that is Air Traffic Control
(ATC). Here, the primary radar at the airport sends out an interrogation signal of frequency
1.03 GHz. A transponder on the aircraft sends out signals at a frequency of 1.09 GHz back
to the Control Centre providing details of the aircraft's identity and height to be sent to a
display. After this identification, range and azimuth measurements from the primary radar
can be combined with height information provided by the transponder to track the target as it
comes into land.

As well as requiring smaller transmitter powers, the following are advantages of using
secondary radar as well as a primary radar:

(1) The signals transmitted by a transponder on the aircraft are independent of the radar
cross-section. If just a primary radar were to be used, there would be a dependence of
the return signal on the radar cross-section of the target.

(2) If just a primary radar was used, the returns would be affected by clutter (reflections
off extraneous targets) and other interference effects. This is not the case with
secondary radar.

4.2 Communication Between Aircraft and Control Centre

Let us look at the whole process of transmission and reception.

(A) MODES 3/A and C

Step 1 : Transmission of Interrogation Pulses

Interrogation pulses, which modulate a sine wave of frequency 1.03 GHz are sent out by the
primary radar at regular intervals. First, we wish to identify the target. In the so-called
Mode 3/A, the following pulses are sent out :
Figure 8

Pulses P1 and P3 are sent out by a highly directional antenna whilst pulse P2 is sent out by an
omnidirectional antenna which has no directivity ; the reason for sending out this additional
pulse will be given in the next stage of the process. The pulse widths are 0.8 , the time
between P1 and P3 is 8 .

Nowadays, phased array radar are used to send out the beams. Now, we are going to receive
height information from the target. Hence, monopulse methods need only be used to
determine the azimuth. This means that the beams sent out are only required to have a fine
resolution in the azimuthal direction and can be relatively broad in the elevation direction
thus cutting down on the number of dipoles used in the phased array. The typical
dimensions of the antenna are 8 m wide and 1.6 m high.

Step 2 : Aircraft Receives Interrogation Pulses

The transponder on the aircraft detects that the spacing between pulses P1 and P3 is 8
and hence identifies this as Mode 3/A interrogation. However, before sending out a reply,
the transponder will need to know whether pulses P1 and P3 come from the main lobe or one
of the sidelobes ; this problem is illustrated below :
In this case, the returns to the radar come from the sidelobe not the mainlobe.
If pulses P1 and P3 come from the mainlobe, then it was intended to interrogate the aircraft.
If these pulses come from a sidelobe, then it was not intended to interrogate the aircraft and
the aircraft should not send back a reply. This is where pulse P2 comes in. Remember that
this pulse was transmitted by an omnidirectional antenna. The gain in any direction is less
than the mainlobe gain of the directional antenna but is larger than the gain of the largest
sidelobe.

This is illustrated below :

In the direction of the mainlobe, the gain of the directional antenna is greater than the gain of
the omnidirectional antenna. In the direction of the sidelobes, this situation is reversed and
the gain of the omnidirectional antenna is larger.

Hence, if the amplitudes of P1 , P3 > P2 , then it is safe to assume that the mainlobe was
directed at the aircraft and the transponder sends back a reply. If P1 , P3 < P2 , then no reply
will be sent back to the Control Centre. P2 is termed the control pulse and occurs 2 μs after
pulse P1 .

Step 3 : Transponder Sends Back Reply Identifying Aircraft

Suppose that the transponder accepts the transmission from the primary radar and recognises
the mode as Mode 3/A. It now sends back a stream of pulses representing its identity code.
The pulse positions are represented as below:
F1 and F2 are framing pulses. The pulses marked A, B, C and D identify the aircraft. Note
that the A, B, C and D pulses may or may not be present. The presence of a pulse signifies a
'1' and the absence of a pulse signifies a '0'. The replies are given as binary coded octal
values, with (A4A2A1 )2 representing the most significant octal digit, (B4B2B1 )2 representing
the next most significant octal digit, etc. The pulse position marked X is not at present being
used.

Some special codes are used :

77008 : (ie all A and B pulses present, all C and D pulses absent) - MAYDAY !!!!
76008 : aircraft has lost communication with the controllers
12008 : responding aircraft is not working within the air traffic control system and should be
ignored.

Because of the seriousness of the first two of the above codes, identification codes 77xx8 and
76xx8 are not used, where x indicates any octal value.

Example:

What reply pulse code results from Mode 3/A interrogation with the transponder set to a
value of 45738 ?

Solution : The ID digit "4" is encoded by the A (most significant) group. Thus a pulse
would be present in the A4 position and absent in the A1 and A2. The next digit "5" is
encoded with the "B" group and B1 and B4 are present and B2 is absent. Likewise, C1 , C2
and C4 are present as are D1 and D2 with D4 absent.

Step 4 : Determine the Range and Azimuth of the Target

The range and azimuth of the target are determined using the techniques described in
Sections (2.2) and (2.6).

Step 5 : Find Out Height of the Target

Having locked onto the target we still need to know the height of the target. In this case,
Mode C transmission is used. In this case, the following pulse train is sent out :
The time difference between pulses P1 and P3 is 21 .

As for 3/A, P1 and P3 are from the directional antenna whilst P2 is from the omnidirectional
antenna, again as a check that the aircraft is not being interrogated through the sidelobes of
the former antenna. The spacing between pulses P1 and P3 is different than for Mode 3/A,
but the spacing between P1 and P2 is the same. The transponder on the aircraft recognises
this as a different mode and sends back a pulse train of the same form as for Mode 3/A
except that in this case, the binary coded octal value formed by the A, B, C and D pulses
represents the aircraft's height which is determined from the aircraft's altimeter.

Step 6 : Subsequent Steps

The above procedure can be repeated to track the aircraft. It should be noted that three
independent parameters are required to track a target in three dimensions. In the ATC
situation that we are discussing, these parameters are : range, azimuth and height. The height
information is provided by the aircraft. In a military situation, the height would not be so
readily available ; in this case we would have to determine the elevation of the target instead
just using the primary radar and this would require a larger phased array radar than the one
used in ATC in order to obtain a beam that is narrow in elevation as well as in azimuth ;
needless to say, such a larger radar would be more expensive.

--------------------------------------------------------------------------------------------
Problem R9

The following reply pulse code is received from Mode 3/A interrogation. Decode this and
comment on the result.
 (B) MODE S

The amount of information transfer when using Modes 3/A and C is limited. In recent years,
a new method of transferring information from the Control Centre and aircraft and back
again has been developed where each aircraft is polled Selectively ; this is referred to as
Mode S. Not all aircraft use Mode S transponders, so it is important that the Control centre
can interrogate and receive information from both the old Mode 3/A and C type transponders
and the new Mode S transponders.

Let us now look at the steps in the communication process when using Mode S.

Step 1 : Transmission of Interrogation Pulses

Interrogation pulses, which modulate a sine wave of frequency 1.03 GHz are sent out by the
primary radar at regular intervals. We cannot assume that all transponders operate using
Mode S, so we need to determine this in the interrogation pulses. The following pulses are
sent out by the main (directional) beam (P1 , P3, P4) and the control (directional) beam (P2 ).

If the transponder is operating in Mode 3/A or C, then it will reply according to the
separation between the pulses P1 and P3 ; this assumes that the control pulse P2 has an
amplitude less than P1and P3 so that the interrogation is taking place through the main beam
and not its sidelobes. If the transponder is operating in Mode S, then it will only operate
when pulse P4 is received which is 1.6 long ; this pulse is ignored by Mode 3/A and C
transponders. P4 occurs 2 after P3. This form of interrogation is called the 'all call'
mode, because at this stage it accepts replies from Mode 3/A, Mode C and Mode S
transponders.

Step 2 : Reply from Mode S Transponder

Assume that the transponder is Mode S. The next stage is that it sends back a digital signal
to the Control centre identifying the aircraft. The from of the reply is as below :
The first 8 μs represents 4 "preamble pulses" that are arranged in such a way that they cannot
mimic the replies from Modes 3/A and C ; this is so that the Control Centre knows that the
reply is from a Mode S transponder and that the other two modes are not being used.

After this, there is a stream of either 56 or 112 pulses representing a binary coded signal.
The reply used pulse position modulation (PPM). Each pulse is 0.5 μs wide and occupies a
width of twice that amount. If the pulse occupies the left hand side of the 1 μs slot, then it
represents a 1, otherwise it represents a 0. For example, the pulses a,b,c,d in the above
diagram represent 1101. The last 24 pulses identify the aircraft.

Step 3 : Selective Interrogation of the Aircraft

The Control Centre now knows (i) the identity of the aircraft and (ii) that the transponder
operates in Mode S.

Now, the Control Centre wishes to interrogate this particular aircraft and send it a message;
perhaps it wants an update on the aircraft's height or send instructions.

The special interrogation pulse and the control pulse (P5) are sent out as shown below :

In this case, P5 is from an omnidirectional antenna and represents the control pulse ; by
comparing its amplitude with those of the upper signal, the transponder on the aircraft can
determine if interrogation is taking place through the sidelobes. P 2 is in the same position as
the control pulse used when interrogating Mode 3/A and C transponders. Hence, any aircraft
in the vicinity that use Mode 3/A and C will ignore this signal because pulse P2 is of the same
height as P1 and hence these transponders will think that they are being interrogated through
the sidelobes and will not reply.

The transmission is done using a technique called Differential Phase Shift Keying (DPSK).
At a certain point in the Main Beam, there is what is known as Synchronisation Phase
Reversal, where the signal goes from a 0 bit followed by a 1. If pulse P 5 in the Control Beam
has a smaller amplitude than the corresponding pulse in the main beam, then the Mode S
transponder on the aircraft recognises this phase reversal as the beginning of the message.

Now the primary radar uses parity checking to ensure that there are no errors that occur
during transmission (ie bit reversals). This is achieved by calculating a parity word from the
first 32 (if 56 bit transmission) or 88 (if 112 bit transmission). This parity word is added on
to the last 24 bits of the transmission which represent the address of the aircraft being
interrogated.

The transponder on the aircraft calculates, using the same algorithm, from the first 32 or 88
bits the parity word used and subtracts this form the last 24 bits of the transmission. If, after
subtraction, the transponder own address is obtained, then the transmission is deemed to be
good and the message is accepted ; if not, then the message is ignored and the interrogation
from the Ground will need to be repeated.

Subsequent Steps

Several transmissions can take place between the Control Centre and the aircraft using Mode
S, with some quite complicated messages being passed between the two; this is not possible
using the older Modes 3/A and C.

Usage of Mode S

For civilian air travel, Mode S is used more than Modes 3/A and C. A recent research
project investigating the surveillance of aircraft after they had landed at Atlanta Airport,
USA, mentioned that 90 % of aircraft are equipped with Mode S transponders. Having said
this, it is still important for the time being, that aircraft fitted with older Mode 3/A and C
transponders can also be communicated with, hence the use of the all-call mode in Step 1.

In Appendix 4, some further topics on the use of secondary radar in Air Traffic Control are
described : (i) The use of Modes 3/A and C by light aircraft, hot air balloons and other minor
air traffic and (ii) the use of Mode S in aircraft collision avoidance.

Advantages and Problems with Secondary Radar

The advantages of Secondary Radar are that as we are receiving returns from a transponder
on the target rather than relying on reflection of radiation from the target, then less
transmitter power is required along with a modest power requirement for the transponder. In
addition height information can be obtained in the transponder reply so we do not need to
have fine resolution in elevation. The reply is at a different frequency from the interrogation,
so problems with returns from clutter are eliminated.

There are some problems as well. The Secondary Radar may expect to receive replies from
one aircraft but instead receive replies from a transponder mounted on another aircraft that it
did not interrogate. However, the measured ranges of the unexpected returns will, in general,
be very different from the expected range hence this problem can be detected quite easily.
This problem is termed FRUIT (False Replies Unsynchronised In Time).

Garbling occurs when a FRUIT reply is received which corresponds to the same range as
the expected reply. However, this can be detected by measuring the azimuth using the
monopulse method ; if one of the replies corresponds to a different azimuth, then this can be
eliminated even though it has a similar range to the expected reply.

5. SUMMARY OF LECTURES AND FUTURE TRENDS

The aim of these 5 lectures has been to introduce you to the physical principles of operation
of a radar, in particular concentrating on the communication aspects. Approximate, but
useful, design equations have been derived to calculate maximum ranges of operation in
terms of the radar's parameters. The principles of range, angle and range-rate determination
have been discussed. The final part of these lectures have covered the use of secondary
radar is air traffic control ; this final application is very important given the increased air
traffic that airports have to cope with.

As with most technology, developments in radar have arisen from the needs of the user.
Traditionally, a considerable amount of effort has been put in developing radar for military
use. This particular application is more demanding than for civil use. In the former case,
the target(s) of interest are uncooperative and hence unpredictable in their intended motion.
Hence, one needs to update frequently position and radial velocity measurements with a
view to deciding whether the target is becoming a threat. This requires the use of tracking
radars that can steer the beam quickly to the target. Phased array radars are particularly
useful here, although they are expensive to construct. The limiting factor is in the
computational power of the software and the speed of the A/D converter used. Some of the
signal and data processing steps can be implemented using VLSI which speeds up the
processing considerably compared with the use of software.

Civil applications, for example Air Traffic Control (ATC), have less stringent tracking
requirements. As the targets are co-operative, they are able to identify themselves to the
control centre and provide useful information such as their height ; thus cruder elevation
beampatterns are tolerable compared to the above military situation. The manoeuvres
executed by such targets are gentler than in the military case and are hence easier to follow.
The emphasis is on the communication between the Control Centre and the aircraft and the
avoidance of collisions in a crowded airspace.

Clutter can corrupt returns from targets ; these are returns from unwanted objects such as
buildings, hills, the sea, birds, other aircraft etc.

However, some of this clutter is of interest to study by radar.

For example, there is interest in studying the weather using radar. Radar can probe a larger
volume of the atmosphere than the traditionally used weather balloons. Satellite radar
images of dense cloud, signifying rain, can be obtained and one can monitor the movement
of such clouds to help predict the weather at late times.

There is interest in monitoring sea currents and sea wave motion ; this could be used to plot
windmaps and to predict hurricanes. Ground radar can be used to image buried objects (eg
pipes in walls, archaeological applications, ...).

Another interesting area is Synthetic Aperture Radar (SAR). This is a method to obtain fine
resolution images of the ground. However, the resolution of a radar is proportional to the
inverse of the size of the antenna ; for the resolution that is usually required, the antenna
dimensions would be impractically large. To overcome this, an ordinary sized antenna is
mounted on an airplane. Returns from the ground are taken every T seconds. The areas of
coverage overlap considerably. Knowing the trajectory of the aircraft one can combine the
returns in such a way that one is synthesising the image obtained with an antenna of length
n.V.T where V is the speed of the aircraft and n is the number of returns processed. One can
image down to very small resolution, imaging tanks from satellites for example, or
investigating land use (eg : rainforest reduction).
There are many more interesting applications, see the references at the end if you are
interested.
 APPENDIX 1

Doppler Shift Determination using Complex Input Signals

The Doppler shifted input signal to the radar is given by

(A.1.1)

where the high frequency term that depends on the radar’s frequency has been filtered off.
The Doppler Shift is given by

(A.1.2)

If the target is travelling towards the radar, then V > 0 hence fd > 0 in Equation (A.1.2).
However, if the target is travelling away from the radar, then V < 0 and fd < 0 in Equation
(A.1.2).

One method to determine this frequency is the Fourier transform. If the Fourier transform
of the signal sr(t) in Equation (A.1.1) is taken, then peaks at both +fd and –fd will be
observed. This can be seen by noting that the Fourier transform of cos(2f0t) is given by :

½ [(f+f0)+(f-f0)] (A.1.3)

where (f+f0) represents a delta function, or spike, centred at f = -f 0 and (f-f0) represents a
delta function centred at f = f0 . Hence, changing f0 to –f0 in Equation (A.1.3), the Fourier
transform of cos(-2f0t) is given by :

½[(f+f0)+(f-f0)] (A.1.4)

which is identical to (A.1.3). Hence, one cannot distinguish between positive and negative
frequency shifts when determining the Fourier transform of sr(t), or, put another way, we
cannot tell whether the target is coming towards or away from us. Such information is
crucial in many situations !!!!

This problem may be solved by using Hilbert transform techniques. Essentially, what
happens is that the signal, sr(t), in Equation (A.1.1) is shifted in phase by 90 degrees to form
the signal si(t) given by :

(A.1.5)

sr(t) and si(t) can together be thought of as a complex signal :

 (A.1.6)

sr(t) is referred to as the in-phase component and si(t) is referred to as the quadrature
component. What is the Fourier Transform of s(t) ? To determine this, note that

:
(i) the Fourier Transform of cos(2fdt) is ½[(f+fd)+(f-fd)]
(ii) the Fourier Transform of sin(2fdt) is ½ j[(f+fd)-(f-fd)]

where (x) is the delta function.

-----------------------------------------------------------------------------------------------------------
Problem A.1

(a) Using the above results, show that the Fourier transform of exp(j2fdt) has a single
peak at fd only.
(b) Show also that the Fourier transform of exp(-j 2fdt) has a single peak at –fd only

What you should have found from Problem A.1 is that if the Fourier transform of the
complex signal, s(t), in equation (A.1.6) is computed and the amplitude spectrum displayed,
then if fd is positive, a positive peak at fd Hz occurs in the amplitude spectrum with the peak
at –fd absent. Similarly, if fd is negative, then a negative peak at –fd Hz occurs in the
amplitude spectrum with the peak at fd Hz absent

The above may seem odd as it is often quoted that the amplitude spectrum has even
symmetry about dc and that if there is a peak at f d Hz, then there is another peak at –fd Hz.
However, this is only true if the input signal to the Fourier transform is real, which would be
the case if the signal sr(t) in equation (A.1.1) is fed in . However, if the input signal is made
complex, as in Equation (A.1.6) then the output amplitude spectrum no longer has this
symmetry.

------------------------------------------------------------------------------------------------------------
Problem A.2

An L-band radar with operating frequency 1 GHz is determining the radial velocity and
direction of travel of an aircraft. The aircraft has a radial velocity of 300 ms-1.

(a) Suppose that just the in-phase signal sr(t), equation (A.1.1) is being processed. Sketch
the amplitude spectrum if just the Doppler signal, sr(t), in Equation (A.1.1) is being
processed if the target is
(i) Travelling towards the radar
(ii) Travelling away from the radar

(b) Suppose that both the in-phase and quadrature components are being processed.
Sketch the amplitude spectrum of the complex signal s(t) = sr(t) + jsi(t) if the target is
(i) Travelling towards the radar
 (ii) Travelling away from the radar

(c) Comment on your answers to (a) and (b)

APPENDIX 2 - Discrete Fourier Transform (DFT)

The Discrete Fourier Transform (DFT) is a method to compute the Fourier transform of a
sampled signal. The main assumption is that the underlying signal is periodic, which means
that the frequency spectra are plotted out at discrete frequencies from 0 to in steps of

, where fs is the sampling frequency, which is the inverse of the time between samples,
and n is the number of samples processed.

For a pulsed radar, where one sample per pulse is taken, then and the DFT is
plotted out at discrete frequencies from 0 to PRF in steps of :

Amplitude
Spectrum

Frequency
(Hz)

The Fast Fourier Transform (FFT) is a quick method to compute the DFT. The only
restriction is that the number of samples n is a power of 2, e.g. 64, 512,…

The pth point on the frequency axis, represents a frequency of

Hz

The (p+1)th point on the frequency axis, represents a frequency of

Hz

The resolution is the ability to distinguish between two closely spaced frequencies.
For the DFT, one definition is the difference represented by adjacent frequency points (or
“bins”) in the DFT spectrum:

(A.2.1)

Now, the Doppler frequency is given by

Hence, if is the resolution between two closely spaced Doppler frequencies, then taking
differentials of the above equation:

(A.2.2)

To determine the equivalent resolution in radial velocity, we put the resolution in the
DFT spectrum, ,in Equation (A.2.1) equal to the resolution in Doppler frequency,
in Equation (A.2.2):

Hence

Solving for :

as stated in Equation (41).

 APPENDIX 3 – REVIEW OF SAMPLING

Digital signals are obtained from analogue signals by two operations: sampling and binary
coding. We shall discuss just the sampling process as this is the most critical stage in the
conversion of an analogue signal to digital form. When we sample an analogue signal, we
take information of this signal at constant intervals of time, for example:

x(t)

t
-2 -1 0 1 2

Figure A.3.1 Sampled signal

This is sampled as shown to obtain the sampled signal, xs(t). The time between adjacent
samples = . The sampling frequency is defined as
Hz or rad s-1.
Let us represent the set of samples as: {x0, x1, x2,…,xn,…}. Now sampling of any signal
means that we are ignoring information between samples and hence we are losing this
information. The larger the time between samples, or the lower the sampling frequency,
then the more of this information is lost. Let us look at the following example. Suppose
that we are given a set of samples of a signal as follows:
1.5

1
Voltage (arbitrary units)

0.5

-0.5

-1

-1.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
Time (seconds)

Figure A.3.2 Samples of a signal

The sampling frequency is 20 Hz.

Suppose that you are told that it is a cosine signal. Could you reconstruct the original
continuous signal? Now, the most obvious answer might be:

1.5

1
Voltage (ArbitraryUnits)

0.5

-0.5

-1

-1.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (seconds)

Figure A.3.3 A possible reconstruction of the original signal from Figure A.3.2

i.e. a cosine with amplitude = 1, phase = 0 and frequency = 10 Hz. However, another
possible reconstruction is as follows:
1.5

1
Voltage (Arbitrary Units)

0.5

-0.5

-1

-1.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (seconds)
Figure A.3.4 Another possible reconstruction

i.e. a cosine with a frequency of 30 Hz. Yet another reconstruction is:

 1.5

Voltage (Arbitrary Units)

0.5

-0.5

-1

-1.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (seconds)
Figure A.3.5 Yet another possible reconstruction

that is a cosine with frequency 50 Hz. In fact, considering all possible reconstructions of the
continuous signal, the following are possible frequencies: 10, 30, 50, … Hz. So there is an
ambiguity here in determining the frequency.

In general, if the underlying cosine signal has a frequency of fd and the sampling frequency is
fs then one can fit cosines with the following frequencies to the sample points:

If T is the time between samples, then and the possible frequencies become:

as stated in Equation (44) of the notes. Now suppose that we are given prior information
about the underlying continuous cosine, that its frequency has a maximum value of 10 Hz.
We can then see that the only possible reconstruction is that shown in Figure A.3.3; that is
the frequency is 10 Hz. In fact, it can be shown that in general, that if a signal has
frequency fd , then the minimum sampling frequency for us to unambiguously determine the
frequency of the underlying signal is 2fd . Or, put another way, the times between samples T
should satisfy the following relation:

The maximum value of T is thus given by

as stated in Equation (43) of notes.

This is a basic form of the Sampling Theorem (Shannon 1948).

Hence, we can reconstruct the original signal, but we need to know something about the
maximum possible value of the underlying frequency.
 APPENDIX 4 – FURTHER TOPICS IN SECONDARY RADAR

A.4.1 Use of Modes 3/A and C for Minor Air Traffic

It might be thought that Modes 3/A and C transponders will eventually be phased out leaving
all air traffic to use Mode S. Whilst it is true that all passenger airliners will eventually be
using Mode S, there is still a market for Modes 3/A and C for light aircraft, gliders,
microlites and hot-air balloons. Mode S transponders are too expensive and are too
sophisticated for such air traffic. Al that is required is that the control tower knows where
these craft are so that they can prevent them causing problems for major passenger airlines.

On take-off, the Control issues a code to the pilot and the Mode 3/A transponder on the light
aircraft etc. can give back this information to the Control along with height information from
the Mode C transponder. Primary radar can give information on the range and azimuth.
These transponders can also be used for collision avoidance of these small craft with the
larger aircraft – see section A.4.2 below. Racal are one of the companies involved with the
manufacture of lightweight transponders for light aircraft use.

Reference : New Electronics on Campus, Summer 1999, pp 10 – 11 (author Paul Hanbury)

A.4.2 Traffic Alert and Collision Avoidance System (TCAS)

So far, we have talked about communication between the Control Tower and the aircraft.
However, there is need for the same type of communication between aircraft.

To illustrate this, let us look at a problem in Air Traffic Control that sometimes occur.

Two aircraft A1 and A2 were at a similar range and azimuth with respect to the control
tower. The intention has been to ask aircraft A2 to descend first and land, then to ask aircraft
A1 to land. Unfortunately, as the labels on the display for both these aircraft were practically
on top of one another, the controller mistakenly asked aircraft A1 to descend first, which it
dutifully did as the pilot could not see A2 through the clouds. A1 was thus on a collision
course with A2. Fortunately, the pilot of A2 saw A1 approaching and took avoiding action
and a major disaster was averted.

The increase in air traffic has resulted in aircraft being fitted with a TCAS so that they can
automatically detect the proximity of another aircraft and take avoiding action. Essentially
each aircraft is fitted with a primary radar and they can also receive transmissions from
Modes S, 3/A and C.

Suppose that two aircraft are on a collision course, for example in the situation in the above
diagram.

Suppose that each of these aircraft are fitted with Primary Radar and Mode S transponders.
Look at the situation from Aircraft A1’s point of view.

The distance from A1 to A2 can be determined from the time delay of pulses from the
primary radar to go from A1 to A2 back to A1.

The bearing of A2 with respect to A1 can be determined using monopulse methods.

The height of A1 above ground can be determined from its own on-board altimeter.

The height above ground of the other aircraft A2 can be determined from its Mode S reply.

Similar measurements are taken on A2.

Using this information from a few measurements, the TCAS on A1 can predict if a collision
is likely to occur. At the most basic level the TCAS can give a warning that a collision is
likely to occur and leaves it to the pilot to take avoiding action. More sophisticated systems
actually give out advice to the pilot as to the avoiding action that should be taken.

-------------------------------------------------------------------------------------------------------------
Acknowledgement : I would like to thank Air Cdre. Norman Bonnor, of the Institute of
Engineering Surveying and Space Geodesy (IESSG), University of Nottingham for providing
me with information on secondary radar and TCAS,
--------------------------------------------------------------------------------------------------------------
-
 REFERENCES

1. S Kingsley and S Quegan, Understanding Radar Systems, Mc Graw-Hill (1992) -

Good introduction to the subject with applications discussed

2. B Edde, RADAR : Principles, Technology, Applications, PTR Prentice Hall (1993) -

Detailed description of Secondary Radar (Section 14-4)