George W. Stimson introduction to Airborne Radar (Se)

Подождите немного. Документ загружается.

But if V is constant, this advance will be the same for

both pulses. Therefore, the round-trip distance traveled by

the pulses to any one point on the ground will be the same,

making the phases of the returns the same (Fig. 3).

So, for each resolvable range interval the radar returns

received from the ground may be canceled simply by pass-

ing the digitized video outputs of the radar receiver through

a single-pulse-delay clutter canceler. As illustrated below, it

delays the return of pulse (n) by the interpulse period, T,

and subtracts it from the return of pulse (n + 1).

CHAPTER 24 Separating Ground-Moving Targets from Clutter

319

3. Returns of each pulse are received by the same segment that

transmitted the pulse. Consequently, round-trip distances trav-

eled by both pulses n and (n + 1) will be equal.

Successive returns from a moving target, however, will

differ in phase as a result of the radial component of the tar-

get’s velocity

. Consequently, they will not cancel but will pro-

duce a useful output.

Effective as this technique is, it has four limitations:

1. The PRF is tied to the aircraft’s velocity

2. Very tight constraints are placed on aircraft and

antenna motion

3. The phase and amplitude characteristics of the two

antenna segments and of the receive channels for

both segments must be precisely matched

4. Only half of the aperture is used at any one time

The fourth limitation may be partially removed by adjust-

ing the velocity and PRF so that during the interpulse period

the phase centers advance by only half the distance between

them (Fig. 4), the entire aperture may be used for transmis-

sion. But still only half the aperture may be used for recep-

tion. Returns of pulse (n) must be received by the forward

segment; returns of pulse (n + 1), by the aft segment.

Although both pulses are not transmitted from the same

point in space and returns from the same ranges are not

received at the same points in space, the result is the same

as if they were. For as indicated in the table at right, the

phase centers’ total displacement for transmission and

reception is the same for both pulses. Therefore, the round-

trip distance traveled to any one point on the ground is the

same for both pulses—just as when the pulses are alternate-

ly transmitted by the fore and aft antenna segments.

Return

(n)

Return

(n + 1)

= round trip transit time to

range, R, of target from

which return is received

Pulse

(n)

Pulse

(n + 1)

COMPLETE DPCA CYCLE

Time

V t

–

Delay = T

(1/PRF)

Returns

from

Range R

Summer

To Doppler

Filtering

Pulse

(n + 1)

Pulse

(n)

CLUTTER CANCELER

To transmit with full aperture, velocity, V, and PRF are adjusted

so the radar travels only half the distance between phase cen-

ters during interpulse period. Returns of pulse

are received

by the forward antenna segment; returns of pulse

(n + 1)

, by

the aft segme

nt. So the round trip distance traveled by both

pulses is the same.

t + T

Return (n)

t = t

(t + T) + t

Time

Pulse

(n)

V t

MODIFIED DCPA

Pulse

(n+1)

= round-trip transit

time to range, R,

of target from

which returns

are received.

Return (n + 1)

V t

Pulse

Displacement of Phase Centers

Transmit Receive Total

(n)

(n + 1)

0V t

+ V t

V t

5. Relationship between return received from a slowly moving

target on the ground and mainlobe clutter. Because of tar-

get‘s radial component of velocity, clutter having the same

doppler frequency as the target, f

, is received from a differ-

ent angle off the boresight line.

6. Placing a notch in the antenna receive pattern at angle, θ

from which the radar receives mainlobe clutter having the

same doppler frequency, f

, as target, n, prevents clutter from

interfering with the target‘s detection. Doppler filtering isolates

the target return from other clutter.

PART V The Problem of Ground Clutter

320

Notching Technique

Notching has the advantage over Classical DPCA of not

requiring that the PRF be tied to aircraft velocity and of

relaxing the constraints on aircraft and antenna motion.

Mainlobe clutter is rejected without rejecting target returns

by taking advantage of the doppler shift due to the radial

component of a target’s velocity, small as it may be. Because

of this shift, clutter having the same doppler frequency as

target n comes from a slightly different angle,

, off the

boresight line (Fig. 5). Therefore, by placing a notch in the

antenna receive pattern at

, the clutter can be rejected

without rejecting the return from target n (Fig. 6).

Return From

Target n

Clutter having

same doppler

frequency as

target n.

Boresight Line

MLC

Doppler Filter Bank

Target

Return

Angle Off Boresight

(+)(–)

Shift due to target’s radial

component of velocity.

Return From

Target n

Clutter having

same doppler

frequency as

target n.

Boresight Line

MLC

Angle Off Boresight

(+)

(–) 0

Doppler Frequency

(+)(–) 0

Shift due to target’s radial

component of velocity.

Moreover, the return from target n is isolated from clutter

received from other directions—and therefore having other

doppler frequencies—by doppler filtering.

Because a target’s angular position and radial component

of velocity generally are not known in advance, and because

returns from targets in different directions may be received

simultaneously, a separate notch must be formed for each of

the N resolvable doppler frequencies. To avoid rejecting tar-

get returns along with the clutter, notching must be per-

formed after, not before, doppler filtering.

The notches are produced with an interferometric tech-

nique similar to that used in phase-comparison monopulse

angle tracking (see panel, left). As with Classical DPCA, a

two-segment electronically steered antenna is typically

used. So that very small differences in doppler frequency

may be resolved, dwell times are increased to allow prede-

tection integrated over long periods, t

int

. So that the notch-

HOW A NOTCH IS MADE

In a Two-Segment Antenna’s Receive Pattern

At an Angle θ

Off the Boresight Line

1. Calculate the difference in distance, ∆

, traveled to the

two phase centers, A and B, by returns from a distant

point at the angle, θ

, off the boresight line.

2. Convert ∆

to phase, φ.

3. Subtract

φ from π radians (180 °). Divide by 2. Result is

the phase rotation, ∆

φ, which—when made in opposite

directions to antenna outputs A and B—will increase the

phase difference,

φ, between returns received from θ

180°.

5. Sum the phase-rotated outputs, A' and B'. The returns

received from θ

will then cancel, producing the equiva-

lent of a notch in the antenna receive pattern at angle θ

6. To produce a notch on the opposite side of the boresight

line, reverse the directions of the two phase rotations.

φ = ∆

= W sin θ

2 π

Radar return

from angle q

∆

W sin θ

∆φ =

π – φ

+ ∆

– ∆

ing can be done in the signal processor, a separate receive

and signal processing channel is generally provided for each

antenna segment.

Implementation of the notching process is illustrated in

abbreviated form in Fig. 7. For every range bin in both

Channel A and Channel B, a separate doppler filter bank is

formed. The outputs of each pair of filters, n, passing

returns of the same frequency, f

, from the same range, m,

are then rotated in opposite directions through the angle,

∆φ

. This rotation causes the returns received by the two

antenna segments from the ground at an angle θn off the

boresight line to be 180° out of phase.

The phase-rotated returns are summed, with the result

that the ground returns from

cancel, while returns from

targets at any other angle off the boresight line whose

doppler frequency is f

do not.

For radars in which monopulse sum and difference sig-

nals for angle tracking are produced ahead of the receiver

(i.e., at microwave frequencies), notching is performed sim-

ilarly with the outputs of the sum and difference channels.

In that case, though, rather than being phase rotated and

summed, the outputs of corresponding doppler filters, f

are weighted and summed to shift the null of the difference

output to the angle

off boresight.

In view of the fact that a good many targets on the

ground will have high enough radial velocities to fall in the

clutter-free portion of the doppler spectrum, notching is

generally time shared with conventional moving-target-

indication processing.

Combined Notching and Classical DPCA

Generally, notching provides very good mainlobe clutter

cancellation. But, under some conditions—such as when

frame-time requirements limit dwell times hence achievable

doppler resolution—clutter rejection performance can be

substantially improved by combining notching and

Classical DPCA. This improvement may at any time be

traded to various degrees for an easing of Classical DPCA’s

strict constraints on aircraft and antenna motion and/or for

uncoupling PRF from aircraft velocity.

Implementation differs from that just described primarily

in that for each range bin, two doppler filter banks are

formed from the outputs of each receive channel, and the

inputs to one of these banks are delayed by the interpulse

period, T (Fig. 8).

Although further improvements in clutter rejection per-

formance can be expected, it should be borne in mind that,

since the cancellation techniques rely on clutter scatterers

being stationary, cancellation ultimately will be limited by

the “internal motion” of the clutter.

CHAPTER 24 Separating Ground-Moving Targets from Clutter

321

8. Combination of classical DPCA and notching. Technique can

be used to ease constraints DPCA places on aircraft and

antenna motion or improve clutter rejection performance of

notching in applications that limit dwell time.

9. Implementation of notching technique. Video outputs of

receive channels A and B are collected in range bins. For

each range bin, m, a doppler filter bank is formed. Output of

each filter, n, is rotated in phase in Channel A by +∆φ

and in

Channel B by –∆φ

. Rotated outputs are then summed, creat-

ing the equivalent of a notch in the antenna receive pattern at

, while passing returns from targets at other angles whose

doppler frequency is f

Range

Bins

Filter

Banks

Phase

Rotation

+ ∆φ

− ∆φ

Output of

Receive

Channel

Output of

Receive

Channel

Bin m

Moving

Target

Returns

Range, m;

Doppler

Frequency, f

Bin m

Moving

Target

Returns

Filter

Banks

Phase

Rotation

+ ∆φ

1/PRF

From

Receive

Channel A

Bin m

Range

Bins

Delay

+ ∆φ

1/PRF

From

Receive

Channel B

Bin m

PART V The Problem of Ground Clutter

322

Precise Angle Measurement

While the DPCA and notching techniques enable a target

to be detected which would otherwise be hopelessly embed-

ded in mainlobe clutter, they don’t tell at what angle within

the antenna beam the target is located. For although the

direction of the interfering clutter can be determined from

the frequency of the doppler filter that passes the target

return, without knowing the target’s radial velocity, hence its

doppler frequency, it is impossible to tell directly how far

removed from that angle the target actually is.

In conventional operation of a two-segment antenna and a

two-channel receiving system, a target’s precise direction may

be obtained by comparing the phases, or amplitudes, of the

outputs the target produces from the two channels. But the

slow-moving-target detection techniques fully utilize the out-

puts of both channels for clutter rejection and target detection.

Accordingly, where precise angle measurement is required,

a three-segment antenna and three receive channels are gener-

ally provided. As illustrated in Fig. 9, the outputs of receive

channels A and B are used to provide clutter rejection and tar-

get detection for an effective phase center half way between

the phase centers of antenna segments A and B. The outputs

of channels B and C are similarly used to provide clutter

rejection and target detection for an equivalent phase center

half way between those of antenna segments B and C. The tar-

get’s precise direction is then estimated on the basis of the dis-

tance between the two effective phase centers and the differ-

ence in phase of the two output signals.

Summary

Targets whose true doppler frequencies fall in mainlobe

clutter are commonly separated from the clutter with either

Classical DPCA or notching, both of which employ a two-

segment side-looking antenna.

For DPCA, aircraft velocity and PRF are adjusted so the

radar advances the distance between the segments’ phase cen-

ters during the interpulse period. By transmitting and receiv-

ing alternate pulses with fore and aft segments, both pulses

travel the same round-trip distance to any one point on the

ground; so MLC can be eliminated by a clutter canceller.

For notching, radar returns are sorted with a doppler filter

bank, and a notch is placed in the antenna receive pattern for

each filter. Because of the doppler shift due to a target’s veloc-

ity, MLC having the same doppler frequency as the target will

come from a different direction than target return. So, the

clutter is “notched out” without rejecting target returns.

By combining DPCA with notching, greater flexibility

may be obtained than with either technique alone.

To pinpoint a detected target’s position within the anten-

na’s mainlobe, a third antenna segment must be provided.

9. Approach to precise angle measurement. Since returns

received by two antenna segments are required for clutter

elimination and target detection, a third segment must be pro-

vided to determine target‘s angle,

, within the radar‘s beam.

Channel

Angle

Measure-

ment

Clutter

Elimination

Detection

Channel

Clutter

Elimination

Detection

Angular

Position

325

The Crucial

Choice of PRF

1. PRFs used by airborne radars range all the way from a few hundred hertz to several hundred kilohertz.

ew parameters of a pulsed radar are more impor-

tant than the PRF. This is particularly true of

doppler radars. Other conditions remaining the

same, the PRF determines to what extent the

observed ranges and doppler frequencies will be ambigu-

ous. That, in turn, determines the ability of the radar not

only to measure range and closing rate directly, but to

reject ground clutter. In situations where substantial

amounts of clutter are encountered, the ability to reject

clutter crucially affects the radar’s detection capability.

In this chapter we will survey the wide range of pulse

repetition frequencies employed by airborne radars and

see in what regions significant range and doppler ambigui-

ties may occur. We will then take up the three basic cate-

gories of pulsed operation — low, medium, and high

PRF— and learn what their relative merits are.

Primary Consideration: Ambiguities

The pulse repetition frequencies used by airborne

radars vary from a few hundred hertz to several hundred

kilohertz (Fig. 1). Exactly where, within this broad spec-

trum, a radar will perform best under a given set of condi-

tions depends upon a number of considerations. The most

important of these are range and doppler ambiguities.

PART VI Air-to-Air Operation

4. Maximum opening rate is usually that of ground from which side-

lobe clutter is received behind radar. Maximum closing rate is

that of fastest approaching target.

326

Range Ambiguities. As we have seen, for range to be

unambiguous, the echoes from the most distant detectible

targets must be first-time-around echoes. In other words, all

sources of detectable return must lie in the first range zone:

their ranges must be less than the unambiguous range, R

(Fig. 2). Because targets of large radar cross section, as well

as return from the ground, may be detected at exceptionally

long ranges, range is almost always ambiguous.

However, if the PRF is sufficiently low so that the maxi-

mum required operating range falls within the first range

zone, ambiguities can be eliminated by rejecting any return

received from beyond R

. (Techniques for rejecting it are

described in Chap. 12.) Under this condition, the first

range zone is a region of unambiguous range.

The first range zone extends to a range very nearly equal

to 80 nautical miles divided by the PRF in kilohertz. In

Fig. 3, this range is plotted versus PRF. The area under the

curve encompasses every combination of PRF and true

range for which range is unambiguous. The area above the

curve encompasses every combination for which range is

invariably ambiguous.

Notice how rapidly the curve plunges as the PRF is

increased. From a range of 400 miles at a PRF of 200 hertz,

it drops to 10 miles at a PRF of 8 kilohertz and to only 4

miles at a PRF of 20 kilohertz.

Doppler Ambiguities. Like range, doppler frequency is

inherently ambiguous. Whether the ambiguities are signifi-

cant, however, depends not only upon the PRF but upon

the wavelength and the spread between the maximum

opening and closing rates likely to be encountered. The

maximum closing rate is usually the rate of the most rapidly

approaching target. The maximum opening rate may be

either that of the most rapidly opening target or that of the

ground from which sidelobe clutter is received behind the

radar (Fig. 4). In fighter applications, it is generally the lat-

ter. This rate is very nearly equal to the maximum velocity

of the aircraft carrying the radar.

2. Assuming all return from beyond first range zone is rejected,

this zone is a region of unambiguous range.

3. Area under curve encompasses every combination of range and

PRF for which a target‘s observed range will be unambiguous,

assuming all return from beyond first range zone is either

negligible or rejected.

Click for high-quality image

CHAPTER 25 The Crucial Choice of PRF

The relationship between the PRF and the doppler fre-

quency at which ambiguities arise in a clutter environment

is illustrated in Fig. 5.

It shows the doppler profile for the flight situation pre-

sented in Chap. 23 and includes both the true profile (cen-

tral-line frequencies) and the next-higher-frequency repeti-

tion of it (first upper sideband frequencies). Corresponding

points in the two profiles are, of course, separated by the

pulse repetition frequency. A high closing rate target (B)

appears in the clear region above the highest true clutter

frequency. If this target’s closing rate were progressively

increased, the target would move up the doppler frequency

scale and into the repetition of the sidelobe clutter spec-

trum. On the basis of doppler frequency, alone, the radar

would have no way of separating the target echoes from the

sidelobe clutter, even though their true doppler frequencies

are quite different.

From the standpoint of clutter rejection, therefore, the

highest unambiguous doppler frequency a target can

have—i.e., the highest frequency at which the target will

not have to compete with clutter whose true doppler fre-

quency is different from the target’s—equals the pulse repe-

tition frequency minus the maximum sidelobe clutter fre-

quency. The latter frequency, as we just noted, corresponds

to the radar’s velocity.

Maximum unambiguous doppler = PRF –

(

)

The maximum closing rate for which the doppler fre-

quency will be unambiguous in a clutter environment is

plotted versus PRF in Fig. 6. A wavelength of 3 centimeters

and a radar velocity of 1000 knots are assumed. The plot

decreases linearly from a closing rate of about 8000 knots at

a PRF of 300 kilohertz to a closing rate of 1000 knots

(radar’s ground speed) at a PRF of 70 kilohertz. (It is termi-

nated at this point since at lower PRFs the maximum posi-

tive and negative sidelobe clutter frequencies overlap.)

The area beneath the curve encompasses every combina-

tion of PRF and closing rate for which the observed doppler

frequencies are unambiguous. Conversely, the area above

the curve encompasses every combination for which the

doppler frequencies are ambiguous. For example, at a PRF

of 250 kilohertz, and a closing rate of 5000 knots the

doppler frequency is unambiguous, whereas at a PRF of

150 kilohertz and the same closing rate, the doppler fre-

quency is ambiguous.

If the radar-carrying aircraft’s ground speed is greater

than 1000 knots, the area beneath the curve will be corre-

spondingly reduced, and vice versa.

327

5. True doppler profile and next higher repetition of it. As com-

ponent of target velocity along line of sight to radar increas-

es, target moves through doppler clear region and ultimately

enters repetition of negative frequency sidelobe clutter.

6. Combinations of PRF and closing rate, R

, for which observed

doppler frequencies are unambiguous.

Click for high-quality image

PART VI Air-to-Air Operation

8. Dramatic

increase

in region of unambiguous doppler frequency

resulting from increase in wavelength to 10 centimeters.

9. When region of unambiguous range is plotted to same scale

as PRF, it becomes apparent that choice of PRF is, at best, a

compromise.

328

Since doppler frequency is inversely related to wave-

length, the shorter the wavelength, the more limited the

region of unambiguous doppler frequencies will be. To

illustrate the profound effect of wavelength on doppler

ambiguities, Fig. 7 plots the maximum closing rate at

which the doppler frequency will be unambiguous for a

one centimeter wavelength. Not only is the area under the

curve comparatively small, but even at a PRF of 300 kilo-

hertz, the maximum closing rate at which the PRF is unam-

biguous is less than 2000 knots.

Yet, at a wavelength of 10 centimeters the area under the

curve extends from 1000 knots at 21 kilohertz to about

27,000 knots (off the scale) at 300 kilohertz (Fig. 8).

7. Dramatic reduction in region of unambiguous doppler fre-

quencies resulting from decrease in wavelength λ from 3 to 1

centimeter.

Putting the Plots in Perspective. The regions of unam-

biguous range and unambiguous doppler frequency (for λ

= 3 cm and V

= 1000 knots) are shown together in Fig. 9.

Drawn to the scale of this diagram, the region of unambigu-

ous range (first range zone) is quite narrow. To the right of

it is the comparatively broad region of unambiguous

doppler frequencies. In between is a region of considerable

extent within which both range and doppler frequency are

ambiguous.

Clearly, the choice of PRF is a compromise. If the PRF is

increased beyond a relatively small value, the observed

ranges will be ambiguous. And unless the PRF is raised to a

much higher value than that, the observed doppler fre-

quencies will be ambiguous.

While both range and doppler ambiguities make clutter

rejection difficult, as we shall see their effects on a radar’s

operation are fortuitously quite different. It turns out that

by designing a radar to operate over a wide range of PRFs

and by judiciously selecting the PRF to suit the operational

requirements at the time, the difficulties can be almost

completely obviated.

Click for high-quality image

CHAPTER 25 The Crucial Choice of PRF

329

The Three Basic Categories of PRF

Because of the tremendous impact the choice of PRF has

on performance, it is customary to classify airborne radars

in terms of their PRFs. Recognizing that the regions of

unambiguous range and unambiguous doppler frequency

are very nearly mutually exclusive, three basic categories of

PRF have been established: “low,” “medium,” and “high.”

These are defined in terms not of the numerical value of

the PRF per se, but of whether the PRF is such that the

observed ranges and/or doppler frequencies are ambiguous.

While exact definitions vary, all are similar. The following is

a widely used, consistent set of definitions.

• A low PRF is one for which the maximum range the

radar is designed to handle lies in the first range zone.

In the absence of return from beyond this zone, range

is unambiguous.

• A high PRF is one for which the observed doppler fre-

quencies of all significant targets are unambiguous.

• A medium PRF is one for which neither of these con-

ditions is satisfied. Both range and doppler frequency

are ambiguous.

Which category a particular PRF falls in depends to a

considerable extent upon the operating conditions. A PRF

of 4 kilohertz—first range zone extending to 20 nautical

miles—would be “low” if the maximum target range were

less than 20 miles (Fig. 10). Yet the same PRF, 4 kilohertz,

would be “medium” if the maximum range were greater

than 20 miles and the spread between maximum positive

and negative doppler frequencies exceeded 4 kilohertz.

Similarly, a PRF of 20 kilohertz might be “medium” for a

3-centimeter radar (X-band), yet “high” for a 10-centimeter

radar (S-band) if, say, the radar’s velocity were 200 knots

and the velocity of the fastest target, 1000 knots—maxi-

mum closing rate 1200 knots (Fig. 11).

10. A PRF of 4 kilohertz would be LOW if maximum range of

interest was less than 20 miles; yet MEDIUM if maximum

range of interest was greater than 20 miles.

11. A PRF of 20 kilohertz might be medium at a wavelength of 3 cen-

timeters, yet high at a wavelength of 10 centimeters.

CATEGORIES OF PRF

PRF RANGE DOPPLER

HIGH Ambiguous Unambiguous

MED Ambiguous Ambiguous

LOW Unambiguous Ambiguous

Click for high-quality image

PART VI Air-to-Air Operation

12. In practice, not all of the PRFs within each category are used. Reason will be made clear in subsequent chapters.

330

In practice, not all of the possible PRFs within each cate-

gory are used for any one radar band (Fig. 12, above). At X-

band, for example, PRFs in the low category typically run

from 250 to 4000 hertz; PRFs in the medium category are

on the order of 10 to 20 kilohertz; PRFs in the high catego-

ry may range anywhere from 100 to 300 kilohertz.

One should not get the idea that the classifications are

technicalities of little practical importance. To the contrary,

in the everyday world of radar development and applica-

tion, they have proved to be immensely useful—especially

regarding fighter and airborne early warning radars. As we

shall see in subsequent chapters, whereas changing the PRF

within any one category does not alter the radar’s design in

any fundamental way, changing the PRF from one category

to another radically affects both the radar’s signal processing

requirements and its performance.

In fighter radars, the three categories of PRF complement

one another nicely from the standpoint of performance.

Low PRF Operation. Because range is unambiguous at

low PRFs, this mode of operation has two important advan-

tages. First, range may be measured directly by simple pre-

cise pulse delay ranging. Second, as will be explained in the

next chapter, virtually all sidelobe return can be rejected

through range resolution.

However, unless the mainlobe clutter is separated in

range from the targets the radar encounters, it can be reject-

ed only on the basis of differences in doppler frequency. And

because of the overlapping of successive repetitions of the

doppler spectrum at low PRFs, the clutter cannot be rejected

without also rejecting the returns from a considerable por-

tion of the spectrum in which targets may appear. If the

wavelength is long enough, the ground speed low enough

∝V

/λ), and the antenna large enough (θ

3dB

∝λ/d) the

mainlobe clutter spectrum will be sufficiently

narrow that

the possible loss of target return is quite tolerable.

But for the conditions under which most fighter radars

must operate—short wavelength, small antenna, and

potentially high ground speed—if the PRF is low enough to

extend the first range zone out to reasonably long ranges—

30 or 40 miles—the unambiguous doppler spectrum is col-

1. Except for return from point

targets of exceptionally large

radar cross section.

LOW PRFs

ADVANTAGES LIMITATIONS

1. Good for air-to-air look- 1. Poor for air-to-air look-

up and ground down—much target

mapping. return may be rejected

along with mainlobe

clutter.

2. Good for precise range 2. Ground moving targets

measurement and fine can be a problem.

range resolution.

3. Simple pulse delay 3. Doppler ambiguities

ranging possible. generally too severe

to be resolved.

4. Normal sidelobe return

can be rejected

through range

resolution.

Click for high-quality image