George W. Stimson introduction to Airborne Radar (Se)

Подождите немного. Документ загружается.

the desired range resolution (Fig. 30). Surprisingly, if this is

done in the absence of mainlobe clutter, a low PRF radar

can actually obtain greater search detection ranges than a

high PRF radar employing the same average power even

against nose hemisphere targets.

What makes the difference are the losses incurred by the

high PRF radar due to eclipsing—return being received

while the radar is transmitting and the receiver is blanked

out.

True, even at low PRFs, a considerable amount of return

may be lost as a result of eclipsing. But eclipsing is much

less of a problem at low PRFs than at medium and high

PRFs.

For as long as a returned pulse is not received at exactly

the same time as the transmitter is transmitting, some of the

pulse will get through to the receiver, and with low PRFs

only the return from zero range is so synchronized. As the

range increases the portion of the return getting through

increases. For targets at ranges greater than one pulse

length, none of the return is lost.

This is so, of course, only up to the point where the trail-

ing edge of the echo from a target at the maximum range of

interest is received as the leading edge of the next pulse is

being transmitted. In other words, the interpulse period

must be at least one pulse width longer than the round-trip

transit time for the most distant target of interest (Fig. 31).

CHAPTER 26 Low PRF Operation

351

31. To avoid eclipsing of long-range targets, the interpulse period

must be at least one pulsewidth longer than the transit time for

the most distant target of interest.

30. Duty factor can be increased by transmitting very long pulses

and using large amounts of pulse compression to obtain the

desired range resolution.

Provided this requirement is met, duty factors of up to 20

percent may be used by a radar operating at low PRFs with-

out incurring a significant eclipsing loss.

In contrast, at medium and high PRFs, because of range

ambiguities, the severity of eclipsing is independent of the

range from which the return is received. The eclipsing loss

increases directly with the duty factor.

Moving Targets on the Ground. In air-to-ground opera-

tions detecting ground moving targets (GMTs) may be a pri-

mary objective, but in air-to-air operations rejecting GMTs

Click for high-quality image

PART VI Air-to-Air Operation

352

may be essential. When operating over territory where

there are hundreds of moving vehicles on the ground—

cars, trucks, trains, and so forth (Fig. 32, above)—a radar

may detect a great many more GMTs than airborne targets.

The GMTs may so clutter up the display that the operator

cannot discern his own targets among them—even though

the targets have been solidly detected and are clearly dis-

played.

If the separation between mainlobe clutter lines is suffi-

ciently wide, the number of filters at the ends of the

doppler filter bank whose outputs are not used can be

increased enough to exclude GMTs without rejecting an

unacceptable amount of possible target return.

If the lines are too closely spaced for this, the GMTs may

be identified by observing the effect of PRF switching on

their apparent doppler frequencies. Because of the greater

ground speeds of airborne targets, at such low PRFs an air-

craft’s apparent doppler frequency will generally be its true

doppler frequency minus some multiple of the PRF. Con-

sequently, the apparent frequencies of these targets will

usually change when the PRF is switched. On the other

hand, the observed doppler frequencies of ground moving

targets, whose speeds are much lower, will generally be true

frequencies and so will not change (Fig. 33). By disregard-

ing those threshold crossings which occur in the same

doppler filter after the PRF has been switched, GMTs may

be prevented from appearing on the display.

In air-to-ground applications where GMTs rather than

airborne moving targets are of interest, the same procedure

is commonly used in reverse: airborne moving targets are

prevented from appearing on the display by discarding

those threshold crossings which do not occur in the same

doppler filter after the PRF has been switched.

32. When searching for aircraft over areas where hundreds of vehicles may be moving on the ground, means must be provided to eliminate

these targets from the radar display.

33. Ground moving targets may usually be distinguished from air-

borne targets because their apparent doppler frequencies do

not change if the PRF is changed slightly.

Click for high-quality image

Summary

At low PRF, mainlobe clutter may largely be eliminated

by offsetting the doppler spectrum so the central line is at

dc and passing the return through a clutter canceller and

bank of doppler filters. To keep the clutter in the canceller’s

rejection notches, the “offset” must be varied with radar

speed and antenna look angle.

Sidelobe clutter as well as mainlobe clutter residue and

background noise are then minimized through a combina-

tion of range gating and doppler filtering. Maximum detec-

tion range is usually limited only by receiver noise.

The principal limitation of low PRFs is doppler blind

zones—regions in the doppler spectrum for which a target’s

“observed” doppler frequency is the same as that of the

mainlobe clutter. The zones are the same width as the

mainlobe clutter lines and are spaced at intervals equal to

the PRF. While not a serious problem where large antennas

and low radar speeds are practical, in fighter radars blind

zones can be acceptably reduced only by employing such a

high PRF that R

, hence the maximum operating range, is

severely reduced or by limiting the maximum look angle.

The possibility of a target remaining in a blind zone for

an entire time-on-target may be minimized by switching

among widely separated PRFs. Alternatively, a limited span

of doppler frequencies may be kept clear by jittering or

sweeping the PRF. Or, the doppler frequency of a given tar-

get may be kept clear by adaptively selecting the PRF.

Multiple-time-around target echoes may be singled out

by jittering the PRF and be blocked from reaching the dis-

play. Both PRF switching and PRF jittering, however,

reduce detection sensitivity.

By transmitting very long pulses and employing pulse

compression to provide adequate range resolution, duty

factors of up to 20 percent may be achieved.

In fighter applications, severe doppler ambiguities make

discrimination between airborne and ground moving tar-

gets difficult. The problem can be alleviated at the cost of

wider blind zones by discarding the outputs of a larger

number of filters at the ends of the doppler filter bank or by

noting whether a target appears in the same doppler filter

after the PRF has been changed slightly.

Because of the blind zone problem, low PRFs are gener-

ally used only where mainlobe clutter can be avoided or

where large antennas and low radar speeds are practical.

CHAPTER 26 Low PRF Operation

353

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 4. Higher peak powers or

can be rejected larger amounts of

through range pulse compression

resolution. generally required.

355

Medium PRF

Operation

1. A medium PRF is one for which both range and doppler fre-

quency are ambiguous.

medium PRF is, by definition, one for which

both range and doppler frequency are ambigu-

ous (Fig.1). In practice, only the lower reaches

of the relatively wide band of PRFs satisfying

this definition are actually used. In general, the optimum

value increases with the radar’s radio frequency. For the X-

band, medium PRFs typically range from about 8 to 16

kilohertz—slightly higher than the top of the low PRF

range, which falls somewhere between 2 and 4 kilohertz.

Medium PRF operation was conceived as a means of get-

ting around some of the limitations of low and high PRFs in

fighter applications. The primary reason for operating

above the low PRF region is to improve the radar’s ability to

contend with mainlobe clutter and GMTs. And the primary

reason for operating below the high-PRF region is to

improve the radar’s ability to contend with sidelobe clutter

in tail hemisphere (low-closing-rate) approaches.

In this chapter, we will take a closer look at medium PRF

operation. We will see what must be done to separate tar-

gets from clutter and how the signal processing is per-

formed. We will then take up the problems of rejecting

ground moving targets, eliminating blind zones, minimiz-

ing sidelobe clutter, and rejecting sidelobe return from

those targets on the ground which have exceptionally large

radar cross sections.

Differentiating Between Targets and Clutter

As in the preceding chapter, to get a clear picture of the

problem of rejecting ground clutter, let us look at the range

and doppler profiles for a representative flight situation. We

will assume that the radar has a PRF of 10 kilohertz and

that the maximum range of interest is 24 nautical miles.

PART VI Air-to-Air Operation

356

Range Profile. This profile with the true range profile

above it, and the flight situation from which the profiles

were derived above that, is illustrated in Fig. 2.

2. Range profile of a representative flight situation. PRF is such

that the maximum range of interest is broken into three range

zones.

3. Doppler profile of the representative flight situation. Mainlobe clutter lines are much more widely spaced than for low PRF operation; other

conditions remaining the same.

The width of the range zones is about 8 miles (80 ÷ 10 =

8). The maximum range of interest, therefore, is in effect

broken into three segments, 8 miles long.

As seen by the radar, however, these are indistinguish-

ably superimposed. Ground clutter completely blankets the

observed range interval. Mainlobe clutter extends from one

end of it to the other. Strong sidelobe clutter received from

short ranges covers a substantial portion of it. None of the

targets are discernible.

Except in the case of very large targets in comparatively

light clutter—such as are encountered in modes of opera-

tion provided for detecting and tracking ships—no amount

of range discrimination alone is going to enable the radar to

isolate the target echoes from the clutter. To reject both

mainlobe and sidelobe clutter, we must rely heavily on

doppler frequency discrimination.

Doppler Profile. As in the case of low PRFs, this profile

consists of a series of mainlobe clutter lines separated by

the pulse repetition frequency, f

(Fig. 3). Between any two

Click for high-quality image

consecutive lines (Fig. 4) appears most, but not all, of the

sidelobe clutter and the return from most but not necessarily

all of the targets. The rest of the sidelobe clutter and target

return is indistinguishably intermixed with the mainlobe

clutter.

Rejecting Mainlobe Clutter. While the doppler profiles

for low and medium PRF operation are similar, there is one

important difference: other conditions remaining the same,

at medium PRFs the mainlobe clutter lines are spread far-

ther apart. Since the width of the line is independent of the

PRF, there is considerably more “clear” room between them

in which to detect targets. Even if the mainlobe clutter is

reasonably broad, it can be rejected on the basis of its

doppler frequency without at the same time, on an average,

rejecting the return from an inordinately large fraction of

the radar’s targets.

Rejecting Sidelobe Clutter. Because of the more severe

range ambiguities, this is not as simple as at low PRFs. To

illustrate, in Fig. 5 the range profile as seen by the radar is

repeated with the mainlobe clutter removed. You will notice

two things in this plot. First, the sidelobe clutter has a saw-

tooth shape. Second, only the short-range target (A) can be

discerned above the clutter. Targets B and C are still

obscured.

The sawtooth shape is due to the strong sidelobe return

from the first range zone being superimposed over the

weaker return from subsequent range zones (Fig. 6).

As for the obscured targets, Target B, in the second range

zone, must compete not only with sidelobe clutter from its

own range but with the far stronger clutter from the corre-

sponding range in the first range zone. Targets C and D, in

the third range zone, must compete not only with sidelobe

clutter from their own range but with the much stronger

return from the corresponding ranges in the first and sec-

ond zones.

The clutter can, of course, be reduced substantially. It

comes not only from different ranges but from different

angles. Since returns from different angles have different

doppler frequencies, we can differentiate between the target

echoes and a great deal of the competing sidelobe clutter if

we sort the return by both range and doppler frequency.

Sorting by range may, of course, be done by range gating

(sampling), just as in low PRF operation. The range gates

will isolate the returns received from relatively narrow

strips of ground at constant range. Because of range ambi-

guities, though, the return passed by each gate will come

from not just one strip, but several. And, as already noted,

one or more of these strips may lie at relatively short range.

CHAPTER 27 Medium PRF Operation

357

5. Range profile as seen by the signal processor, with mainlobe

clutter removed. Only the short-range target can be discerned

above the clutter.

6. Sawtooth shape is due to strong sidelobe return from the first

range zone being superimposed over the weaker return from

the second and third zones.

4. Portion of the doppler profile processed by the radar. The

doppler spectrum is normally shifted to place the central line

of mainlobe clutter at zero frequency (dc).

PART VI Air-to-Air Operation

358

Still, the reduction in clutter obtained through range gating

will be substantial (Fig. 7).

7. Each range gate passes the return from a series of circular

strips (only three of which are shown here) separated from

one another by R

8. Each doppler filter receives only that portion of the total side-

lobe return passed by a single range gate. The filter passes

only that fraction of this return which comes from strips of

ground whose angles relative to the radar‘s velocity are such

that the return falls in the filter‘s passband.

Sorting by doppler frequency may be accomplished by

applying the output of each range gate to a bank of doppler

filters. They will isolate the returns received from strips of

ground lying between lines of constant angle relative to the

radar’s velocity (Fig. 8).

Because of doppler ambiguities, though, any one filter

will pass the return from not just one strip, but several.

Nevertheless, the amount of clutter with which a target’s

echoes must compete will be only a fraction of that passed

by the range gate.

Signal Processing

As illustrated in Fig. 9 (above) for medium PRFs, signal

processing is quite similar to that for low PRFs. There are,

however, three main differences. First, because range ambi-

guities preclude the use of sensitivity time control, addition-

al automatic gain control is needed to avoid saturation of the

A/D converter. Second, to further attenuate the sidelobe clut-

ter (which because of range ambiguities piles up more

deeply at medium PRFs), the passbands of the doppler filters

may be made considerably narrower. Third, additional pro-

cessing is required to resolve range and doppler ambiguities.

As at low PRFs, the first step in processing the IF output

of the radar receiver is to shift the doppler spectrum so as

to place the central mainlobe clutter line at dc. Again, the

shift must be dynamically controlled to account for changes

in radar velocity and antenna look angle. The I and Q out-

puts of the synchronous detector that performs this shift are

likewise sampled at intervals on the order of the transmit-

ted pulse width

and digitized.

However, to reduce the dynamic range required of the

A/D converter, automatic gain control is provided ahead of

the converter. For this, the converter’s output is monitored

and a continuously updated profile of the output over the

course of the interpulse period is stored. On the basis of

this profile, a gain control signal is produced and applied to

the amplifiers ahead of the A/D converter.

By reducing the gain when the mainlobe clutter and

strong close-in sidelobe clutter are being received, the con-

trol signal keeps the converter from being saturated, yet

maintains the input to the converter well above the local

noise level when weaker return is coming through. Since

CHAPTER 27 Medium PRF Operation

359

9. How signal processing for medium PRF operation may be handled. Clutter canceller is optionally included to reduce dynamic range required

of doppler filters. Postdetection integration (PDI) may be provided if filter integration time is less than time-on-target.

1. Compressed pulse width,

when pulse compression is

used.

PART VI Air-to-Air Operation

360

the control signal is derived after the return is digitized, this

is called digital automatic gain control (DAGC).

To reduce the dynamic range required in the subsequent

processing, once the output of the A/D converter has been

sorted into range bins, an optional next step is to get rid of

the bulk of the mainlobe clutter in each bin. This may be

accomplished with a clutter canceller (Fig. 10).

The return in each bin is next applied to a bank of

doppler filters. At the end of every integration period, the

magnitude of each filter’s output is detected. If the scan of

the radar antenna and the bandwidth of the doppler filters

are such that the filter integration time is less than the time-

on-target, magnitude detection may be followed by postde-

tection integration.

In either event, the integrated return passed by each

doppler filter during every time-on-target is applied to a

separate threshold detector. The threshold of this detector

is adaptively set to keep the probability of clutter producing

false alarms acceptably low. The setting may be based on

the average level of the clutter for (a) several range incre-

ments on either side of that in question, (b) several doppler

frequencies on either side of that in question, (c) several

integration periods before and after that in question, or (d)

some combination of these. In general, the optimum aver-

aging scheme for a clutter background is different from that

for a noise background.

When a target is detected, we can tell its apparent range

by observing which bin (or adjacent bins) it was detected

in. Similarly, we can tell its apparent doppler frequency,

hence range rate, by observing which doppler filter (or

adjacent filters) the detection occurred in.

The observed range and doppler frequency will, of

course, be ambiguous. Range ambiguities are resolved by

PRF switching, as outlined in Chap. 12. Doppler ambigui-

ties may be resolved by the methods described in Chap. 21.

Rejecting Ground Moving Targets (GMTs)

GMTs are not nearly the problem they are at low PRFs.

For, at medium PRFs, the mainlobe clutter lines are spread

sufficiently far apart that GMTs appear only near the ends

of the region between lines. Targets with positive closing

rates appear at the lower end; targets with negative closing

rates, at the upper end (f

– f

). GMTs, therefore, can be

eliminated without losing an unreasonable fraction of the

target return simply by discarding any return in the fre-

quency bands where GMTs may appear.

The width of these bands depends upon the wavelength

and the velocities of the targets. Most surface vehicles travel

at less than 65 miles per hour. At X-band (30 hertz per mile

per hour of closing rate), the maximum doppler shift of the

10. Mainlobe clutter is reduced by passing digitized video-

frequency output of receiver through simple clutter canceller

having characteristic such as this.

Click for high-quality image

GMTs relative to the center frequencies of the mainlobe clut-

ter lines would be about 2 kilohertz (65 x 30 = 1950 Hz).

So, for an X-band radar, those GMTs having a compo-

nent of velocity toward the radar (positive doppler shift)

can be eliminated by discarding all return whose frequency

is less than 2 kilohertz above the center of each mainlobe

clutter line (Fig. 11) And those GMTs having a component

of velocity away from the radar (negative doppler shift) can

be eliminated by discarding all return whose frequency is less

than 2 kilohertz below the center of each line. At the same

time, the mainlobe clutter residue will also be eliminated.

When this approach is taken to the problem of GMTs,

the anticipated maximum doppler frequency of the GMTs

relative to the doppler frequency of the mainlobe clutter

usually puts the lower limit on the selection of PRF.

Suppose that to provide a reasonable amount of room in

which to look for airborne targets, we establish a design cri-

terion that at least 50 percent of the doppler spectrum be

clear, i.e., not covered by blind zones. If to eliminate GMTs

we discard all return whose frequency is within 2 kilohertz

of the center of each clutter line (Fig. 12), we must make

the filter bank’s passband at least 4 kilohertz wide. To

accomplish this, the PRF must be at least 2 + 4 + 2 = 8 kHz.

Since the doppler shift is inversely proportional to wave-

length (f

= 2R

⋅

/λ), the shorter the wavelength, the higher

the minimum PRF will be, and vice versa. Take a wave-

length of 1 centimeter, for example. At this wavelength, the

maximum relative doppler shift for a 65 mile per hour

vehicle is 6 kilohertz as opposed to 2 kilohertz.

Consequently, if we apply the above design criterion to a 1-

centimeter radar, the minimum PRF is 6 + 12 + 6 = 24 kHz.

Eliminating Blind Zones

Blind zones, too, are still a problem at medium PRFs. In

fact, because of range ambiguities, the radar must contend

not only with blind zones in the doppler spectrum but with

blind zones in the range interval being searched, as well.

Doppler Blind Zones. Because mainlobe clutter covers a

much smaller portion of the doppler frequency spectrum,

doppler blind zones are far less severe at medium PRFs

than at low PRFs and so can be eliminated by switching

among fewer PRFs. However, additional PRFs are required

to resolve range ambiguities and eliminate ghosts.

Typically, the radar is cycled through a fixed number of

fairly widely spaced PRFs (Fig. 13). If a target is in the clear

on any three of these and its echoes exceed the detection

threshold on all three, the target will be deemed to have

been detected. The range ambiguities will then be resolved

and “de-ghosted.” The optimum number of PRFs varies

CHAPTER 27 Medium PRF Operation

361

11. Most ground moving targets (GMTs), as well as residue of

mainlobe clutter (MLC) passed by clutter canceller, can be

rejected by discarding the return between 0 and 2 kilohertz

and between (f

– 2 kHz) and f

12. If 4 kilohertz of doppler spectrum is discarded to eliminate

GMTs, the PRF must be at least 8 kilohertz to meet criterion

that 50 percent of doppler spectrum be clear.

13. To eliminate blind zones and resolve range ambiguities, a

radar may cycle through a number of widely spaced PRFs.

Click for high-quality image