George W. Stimson introduction to Airborne Radar (Se)

Подождите немного. Документ загружается.

range gates. In fact, if the average transmitted power is held

constant (by increasing the peak power), the maximum

detection range can be significantly increased by backing

off from a 50 percent duty factor and then range gating.

The reduction in duty factor reduces the eclipsing loss, and

the range gates reduce the competing noise or clutter.

Adding range gates, however, is costly. Since range gating

must precede doppler filtering, the entire doppler filter

bank and all subsequent signal processing must be dupli-

cated for every range gate that is provided (Fig. 12).

Increasing the number of range gates from one to just two,

for instance, entails forming twice as many doppler filters,

detecting the magnitudes of twice as many filter outputs,

setting twice as many detection thresholds, and so on.

Moreover, where the value of the PRF is very high (as in

most fighter applications), a great many more doppler fil-

ters are required to provide the same doppler resolution for

high PRF operation as for medium. The cost of range gat-

ing, therefore, can be much higher in a high PRF radar than

in a medium PRF radar.

Mechanization

Just how the signal processing functions outlined in the

preceding paragraphs are actually performed varies widely

from one radar to another. The mechanization, for instance,

may be either analog or digital. Or, to reduce the dynamic

range required of the analog-to-digital converter, the initial

filtering may be analog and only the final doppler filtering,

digital. Also, in some cases the processor may be designed

to look for targets in both sidelobe clutter and doppler clear

regions; in others, only in the doppler clear region.

Analog Mechanization. In analog processors, the IF out-

put of the receiver is applied at the outset to a bandpass fil-

ter which passes only the central band of doppler frequen-

cies.

The pulsed return is thereby converted to a continu-

ous wave signal (Fig. 13), and all subsequent processing is

handled as in a CW radar.

CHAPTER 28 High PRF Operation

373

12. For every range gate that is provided, all filtering and subse-

quent signal processing must be duplicated.

2. The central band is selected

because it contains the most

power. Loss of signal power in

the other bands is no problem

since the noise and clutter

they contain is rejected too.

13. For analog processing, the IF output of the receiver is fed to a

filter which passes only the central spectral band, thereby con-

verting the return to a CW signal.

SOME OPTIONS FOR SIGNAL PROCESSING

Filtering Regions Processed

Initial Remainder Sidelobe Clutter Doppler Clear

Analog Analog X X

Analog Digital X X

Analog Digital X

Digital Digital X

Click for high-quality image

15. To avoid desensitization by strong signals, return is segregat-

ed into subbands for the application of automatic gain control,

AGC.

16.

After the altitude return, transmitter spillover, and mainlobe clut-

ter are filtered out, the radar return is applied to a bank of

doppler filters.

PART VI Air-to-Air Operations

374

The output of the central band filter (Fig. 14, above) is

passed through a

filter which removes the clutter having zero

doppler frequency

—transmitter spillover and altitude

return. Along with them, the echoes of any zero-closing-

rate targets will unavoidably be rejected.

The entire doppler spectrum is then shifted in frequency

so as to center the mainlobe clutter in the rejection notch of

a second rejection filter. As with low and medium PRF

operation, this shift must be dynamically controlled to

match the changes in clutter frequency due to changes in

radar velocity and antenna look angle. Once the mainlobe

clutter has been removed, the same frequency offset may be

applied in reverse to center the doppler spectrum again at a

fixed frequency.

Up to this point, processing is normally done at very low

signal levels. To keep the stronger return in the target

regions from saturating subsequent stages, its relative

amplitude is now reduced through automatic gain control

(AGC). For this, the doppler spectrum is commonly broken

into contiguous subbands (Fig. 15). One or more subbands

may span the sidelobe clutter region, and one or more may

span the doppler clear region. By applying the AGC sepa-

rately in each subband, strong return (or jamming) in one

band is prevented from desensitizing the other bands.

Signal levels equalized, the subbands are, in effect,

recombined and applied to a single, long bank of doppler

filters (Fig. 16).

At the very end of every filter integration time, t

int

, the

amplitude of the signal that has built up in each filter is

detected. If t

int

is less than the time-on-target for the radar

antenna, the detector outputs for the complete time-on-

target are added up in postdetection integrators. Finally, the

integrated output of each doppler filter is supplied to a sep-

arate threshold detector.

14. First steps in signal processing: (1) reject altitude return and transmitter spillover; (2) offset spectrum to track mainlobe clutter;

(3) reject mainlobe clutter; (4) remove frequency offset.

Digital Mechanization. In digital signal processors, as

with medium PRF operation, the IF output of the receiver

is applied to an I/Q detector (Fig. 17). Its outputs (video)

are then sampled at intervals equal to the (compressed)

pulse width by an analog-to-digital converter. To prevent

saturation, digital automatic gain control (DAGC) is applied

to amplifiers ahead of the converter.

In contrast to the output of an analog central band filter,

the samples taken by the A/D converter represent the

power in all of the doppler bands passed by the receiver’s IF

amplifier. However, since the power includes both signal

and noise plus clutter, the signal-to-noise ratio is essentially

the same as when central band processing is employed.

Following analog-to-digital conversion, all of the same

steps may be performed as in analog processing. The only

difference is that they are performed by digital rather than

analog filters.

Ranging

Because of the difficulty of pulse-delay ranging at high

PRFs, FM ranging is generally employed. The accuracy of

FM ranging is proportional to the ratio of the frequency res-

olution of the doppler filter bank to the rate of change of

the transmitter frequency, f

(Fig. 18). The finer the frequen-

cy resolution and the greater f

, the more accurately the

ranging time can be measured. Frequency resolution rough-

ly equals the 3-dB bandwidth of the doppler filters, so

Range accuracy ≈

3 dB

To illustrate, let’s say that the 3-dB bandwidth of the doppler

filters is 100 hertz and the rate of change of the transmitter

frequency f

= 3MHz per second. The accuracy with which

the ranging time t

can be measured then is 100 ÷ (3 x 10

)

= 33 µs. At 12.4 microseconds per nautical mile of range,

this corresponds to an accuracy of about 2.7 miles.

One might suppose that virtually any degree of range

accuracy could be obtained simply by narrowing the filters

and/or increasing f

. But there are practical limits on both.

Filter bandwidth is limited by the integration time (BW

3 dB

≅

1/t

int

). Furthermore, with three slope modulation, the

maximum integration time is less than 1/3 of the time-on-

target (Fig. 19). The minimum possible filter bandwidth,

therefore, is roughly

3/t

hertz.

Not so apparent, f

is limited by the spreading of the clut-

ter spectrum. The spreading is due to the clutter being

received from a wide span of ranges. To keep spreading with-

in acceptable bounds, the maximum shift in the frequency of

the radar return due to range must be no more than a small

fraction of the maximum doppler shift of the clutter.

CHAPTER 28 High PRF Operation

375

17. For digital processing, output of the I/Q detector is sampled

at intervals equal to the (compressed) pulse width and digi-

tized. Digital AGC prevents saturation of the A /D converter.

18. Slope of the modulation curve (f

⋅

) and bandwidth of the filters

(BW

3 dB

) determine the minimum resolvable difference in

ranging time (∆t

19. Filter bandwidth is inversely proportional to the integration

time. With three-slope ranging, the integration time is less

than 1/3 of the time on target.

Click for high-quality image

21. For maximum detection range, a velocity-search mode may

be provided. Once the target is detected, the operator may

switch to the range-while-search mode.

PART VI Air-to-Air Operations

376

The reason is illustrated (for the descending slope of the

modulation cycle) in Fig. 20. It shows a radar detecting two

long range targets. One has a high doppler frequency. The

other, a doppler frequency only slightly higher than the

highest clutter frequency. The antenna’s mainlobe illumi-

nates the ground for a very great distance.

Beneath the situation diagram are three frequency pro-

files. The first, is a plot of the doppler frequencies of the

targets and the ground return.

The second profile shows what happens if the frequency

shift corresponding to the targets’ range is comparable to

the doppler shift of the higher closing rate target. As you

can see, the mainlobe clutter not only shifts into the nor-

mally clutter-free region, but spreads to the point where it

blankets both targets.

In the third profile, the frequency shift corresponding to

range is a small fraction of the doppler shift. Although the

clutter still spreads, it does not spread enough to interfere

with target detection.

It turns out that for a typical fighter application the con-

straints on minimum filter bandwidth and maximum rate

of change of transmitter frequency are such that the range

accuracy is on the order of a few miles. This is substantially

poorer than can be obtained with pulse-delay ranging.

While range information is always highly desirable, it is

not essential when searching for targets at extremely long

ranges. What is most important then is detecting targets and

knowing their direction. Determining whether a target in a

given direction is 100 or 150 miles away can come later.

Because a target must be detected on all three slopes of

the modulation cycle to be detected at all and the integra-

tion time per slope is only one-third of what it would be

without FM ranging, the price one pays for range measure-

ment is a reduction in detection range.

Accordingly, for situations where extremely long detec-

tion ranges are desired, a special mode may be provided in

which range is simply not measured. In this mode, called

velocity search or pulse-doppler search, targets are displayed

in range rate versus azimuth (Fig. 21). Once a target is

detected, the operator can switch to the more standard

range-while-search mode, in which range is measured by

FM ranging and the targets are presented on a range versus

azimuth display.

Problem of Eclipsing

When operating at very high duty factors, a considerable

amount of target return is lost as a result of eclipsing—i.e.,

echoes being received in part or in whole when the radar is

transmitting and the receiver is blanked.

20. How clutter spreading limits the rate (f

⋅

) at which the transmit-

ter frequency may be changed for FM ranging.

a) Two targets approaching from long range. One has much higher closing

rate than the other.

b) Doppler profile, without FM ranging—no clutter spreading.

c) Doppler profile, with FM ranging and

large

value

f. Mainlobe clutter

spread shifts out in frequency and spreads over doppler clear region,

obscuring targets.

d) Doppler profile, with FM ranging and

small

value

f. Clutter spread only

moderately, leaving targets in the clear.

Click for high-quality image

Eclipsing, however, is not always as severe a problem as

it might at first seem. A target is totally eclipsed only when

its range is such that the period during which its echoes are

received exactly coincides with the period during which the

receiver is blanked. Otherwise, at least a portion of the

return gets through (Fig. 23). As the degree of coincidence

is reduced, so is the eclipsing loss.

Even so, eclipsing reduces the signal-to-noise ratio suffi-

ciently to leave periodic holes of appreciable size in the

radar’s range coverage. Fortunately, when searching for tar-

gets approaching from very long ranges, one is concerned

mainly with the cumulative probability of detection—the

probability that a target will be seen at least once before it

has approached to within a given range. Moreover, once a

target has been detected, it generally need not be detected

continuously. A rapidly approaching target will not remain

at an eclipsed range very long. And as the range decreases

and the signal strength increases, the gaps in range cover-

age tend to fill in (Fig. 24).

In applications where closing rates may be comparatively

low and/or more nearly continuous detection is required,

the length of time any one range remains eclipsed may be

reduced by switching the PRF among different values,

much as it is switched to eliminate blind zones in medium

CHAPTER 28 High PRF Operation

377

22. YF-12A of the early 1960s. First airplane capable of sustained Mach-3 flight; first interceptor to be equipped with a high PRF pulse-doppler

radar.

23. As long as the received echoes and periods of receiver

“blanking” do not coincide exactly, a portion of the return will

get through.

24. Reduction in signal-to-noise ratio due to eclipsing for an

approaching target. As range decreases, gaps in range cover-

age grow narrower.

Click for high-quality image

PART VI Air-to-Air Operations

378

PRF operation. In single-target tracking, by periodically

changing the PRF at appropriate times, a target can be kept

largely in the clear. At high duty factors, though, the holes

in range coverage are not easily eliminated, particularly at

short ranges. Furthermore, PRF switching introduces losses

which reduce the maximum detection range in the doppler

clear region.

Eclipsing may also be reduced by lowering the duty fac-

tor. This, of course, will reduce the average transmitted

power. But that reduction can be compensated for by using

multiple range gates. Suppose, for example, that the duty

factor is reduced from 50 percent to 20 percent and four

range gates are provided (Fig. 25). If the peak transmitted

power remained the same, the average power hence the

total received energy would be decreased by a factor of 0.2

÷ 0.5 = 0.4. But as can be seen from the figure, with four

range gates the noise energy with which the signal would

have to compete at any one time would be reduced by the

same factor, so these two effects would cancel. For a contin-

uously closing target, then, the signal-to-noise ratio would

increase in direct proportion to the increase in the fraction

of the time the receiver is not blanked. In this case, the

increase would be on the order of 0.5 ÷ 0.2 = 2.5.

Thus, by reducing the duty factor somewhat and provid-

ing multiple range gates, not only may the detection range

be increased, but the holes in range coverage due to eclips-

ing may be correspondingly narrowed. As noted earlier,

though, providing multiple range gates substantially

increases the cost of implementation.

Improving Tail Aspect Performance

Several approaches may be taken to improving perfor-

mance against low-closing-rate targets in severe clutter.

Since the root of the problem is sidelobe clutter, a logical

first step is to minimize the antenna sidelobes.

For a given sidelobe level, the amount of sidelobe return

with which a low-closing-rate target must compete may be

further reduced by narrowing the passbands of the doppler

filters (Fig. 26). This, of course, entails adding more filters

and, as noted earlier, there are practical limits on how nar-

row the passbands can be made.

At the expense of still greater complexity and a lower

duty factor, the competing return may be still further

reduced by narrowing the pulses and employing more

range gates.

Even then, because of the transmitter spillover

and altitude return, the radar will be blind to zero closing

rate targets—those being pursued at constant range.

A particularly attractive solution to the problem is to

employ high PRFs when long detection range against nose

hemisphere targets is essential and to interleave high and

25. Eclipsing loss may be reduced by reducing the duty factor

and employing multiple range gates.

26. Signal-to-clutter ratio, hence performance against low closing

rate targets, may be improved by reducing the duty factor and

providing multiple range gates or by narrowing the passband

of the doppler filters.

3. For a given duty factor, hence

degree of eclipsing, the num-

ber of range gates may be

increased still further, without

loss of signal, through pulse

compression.

Click for high-quality image

medium PRFs when long detection ranges are required

against both nose and tail hemisphere targets.

An effective way of accomplishing this is illustrated in

Fig. 27. High and medium PRF modes are employed on

alternate bars of the antenna scan. The bars assigned to the

high PRF mode in one frame are assigned to the medium

PRF mode in the next frame, and vice versa. Since adjacent

bars overlap, virtually complete solid-angle coverage is

achieved in both modes. Rapidly approaching targets which

are beyond reach of the medium PRF mode are detected in

the high PRF mode. Low-closing-rate targets, as well as any

shorter range targets which may be eclipsed in the high PRF

mode, are detected in the medium PRF mode.

When PRF interleaving is used, the complexity of the sig-

nal processor for the high PRF mode can be substantially

reduced by processing only the return falling in the doppler

clear region (Fig. 28). This return, of course, must first be

CHAPTER 28 High PRF Operation

379

28. When medium and high PRFs are interleaved, signal processing for the high PRF mode is simplified by processing only the return in the

clutter-free region.

27. Interleaving of high and medium PRFs on alternate bars of the

search scan to achieve maximum detection range in both

nose hemisphere aspects and tail chases.

Click for high-quality image

PART VI Air-to-Air Operations

380

isolated from the clutter—mainlobe, sidelobe, and altitude

return. But that can readily be accomplished by passing the

receiver output through one or more broad bandpass filters.

After performing automatic gain control, their outputs are

supplied to a suitably long bank of doppler filters.

Summary

To maximize detection range, high PRF waveforms for

fighter applications usually have duty factors approaching

50 percent. Even higher duty factors may be used when

illuminating targets for long range semiactive missiles.

To provide an adequate doppler clear region, the PRF

must at least equal the maximum sidelobe clutter frequency

plus the doppler frequency of the highest closing rate tar-

ILLUMINATING TARGETS FOR SEMIACTIVE MISSILE GUIDANCE

If the pulsed transmission of a high PRF radar is used to

illuminate targets for semiactive missiles to home on, both the

PRF and the duty factor may be made somewhat higher than

they would be for normal searching and tracking.

Increased PRF. As explained in detail in Chapter 15, because of

the high velocity of the missile relative to the launch aircraft, a

target’s doppler frequency is generally much higher as seen by

the missile than as seen by the radar in the launch aircraft. To

ensure that the velocity data obtained by the missile is

unambiguous, in computing the minimum acceptable PRF one

must add to the maximum target closing rate half the velocity of

the missile relative to the radar.

Increased Duty Factor. Since for a given peak transmitted

power, detection range increases with duty factor, in the case of

a semiactive missile which must be launched at long ranges it is

desirable to make the duty factor as high as possible. Although

the maximum useful duty factor which a radar may employ for

detecting and tracking targets is limited by the eclipsing loss to

somewhat less than 50 percent, when the radar’s pulsed

transmission is used to illuminate a target for a missile, this

limitation does not necessarily hold. Because the missile is

remote from the radar, blanking is not needed to keep transmitter

noise from leaking directly into the missile receiver when the

radar is transmitting. Consequently, if the missile’s seeker must

be capable of long detection range, the duty factor may be made

considerably higher than 50 percent. The maximum acceptable

duty factor is, of course, limited by eclipsing losses in the radar.

What about the directly received signal from the radar?

Because of the doppler shift, the frequency of the radar’s

transmitted signal is sufficiently different from the frequency of

the target echoes that the seeker in the missile can separate the

two. However, care must be taken in the design of the radar to

minimize the radiation of transmitter noise, some of which

invariably has the same frequency as a target’s echoes.

(

∆V

)

Click for high-quality image

get. For semiactive missile guidance an allowance must be

added for the velocity of the missile relative to the radar.

In analog signal processing, the receiver output is usually

converted at the outset to a CW signal by a central band fil-

ter. Next, successive filters reject the combined spillover

and altitude return and the mainlobe clutter. The remaining

return may then be divided into subbands for the applica-

tion of AGC. The return is then applied to a bank of

doppler filters. If a target’s closing rate is greater than the

radar’s ground speed, the target echoes compete only with

the noise passed by the same doppler filter that passes the

echoes. But if the closing rate is less than this, they must

compete with sidelobe clutter passed by the filter, much of

which may come from comparatively close range.

In many high duty factor fighter radars, the only range

gating is that provided by receiver blanking. At the cost of

increased complexity, noise and sidelobe clutter may be

reduced by using multiple range gates.

Range must generally be measured with FM ranging.

Because the rate at which the transmitter frequency may be

changed is limited by the spreading of the clutter spectrum,

range accuracy is poor. Since range measurement reduces

detection range, where maximum detection range is

desired, a special velocity search mode may be provided in

which range is not measured.

When operating at high duty factors, eclipsing losses are

significant. They may be minimized by switching PRFs,

and/or providing multiple range gates. PRF switching,

though, reduces detection range in the doppler clear

region, and employing multiple range gates increases the

cost of implementation.

Performance against low closing rate targets may be

improved by providing a low sidelobe antenna, greater

doppler resolution, and multiple range gates. One attractive

approach is to interleave high and medium PRF operation

on alternate bars of the antenna scan.

CHAPTER 28 High PRF Operation

381

HIGH PRFs

ADVANTAGES LIMITATIONS

1. Good nose-aspect 1. Detection range

capability—high-clos- against low-closing-

ing-rate targets appear rate targets may be

in clutter-free region of degraded by sidelobe

spectrum. clutter.

2. High average power 2. Precludes use of sim-

can be provided by in- ple, accurate pulse de-

creasing PRF. (Only lay ranging.

moderate amounts of

pulse compression, if

any, are needed to

maximize average

power.)

3. Mainlobe clutter can be 3. Zero-closing rate tar-

rejected without also gets may be rejected

rejecting taget echoes. with altitude return

and transmitter

spillover.

383

Automatic Tracking

n the preceding chapters, we became acquainted with

various approaches to target detection. In this chapter,

we’ll take a closer look at the techniques for tracking

the targets that are detected: the single-target track

(STT) and track-while-scan (TWS) modes introduced in

Chap. 2.

Single-Target Tracking

The goal of single-target tracking is to continuously and

accurately provide current data on a given target’s position,

velocity, and acceleration—all of which may be continuous-

ly changing. Toward that end, separate semi-independent

tracking loops are typically established for range, range rate

(or doppler frequency), and angle.

Functions Included in a Tracking Loop. Each tracking

loop includes four basic functions: measurement, filtering,

control, and response (Fig. 1).

Measurement is the determination of the difference

between the actual value of the parameter (e.g., the target’s

range) and the radar’s current knowledge of the parameter:

in short, the tracking error.

Filtering is the processing of successive measurements to

minimize the random variations (noise) due to target scin-

tillation, thermal agitation, and other corrupting interfer-

ence. Needless to say, tracking accuracy depends critically

on how effectively filtering is done. A tracking filter may be

thought of as a low-pass filter (Fig. 2) whose key parame-

ters—cut-off frequency, gain, etc.—are constantly adjusted

in light of the signal-to-noise ratio, the target’s potential

maneuvers, and the radar-bearing aircraft’s actual maneu-

Measure Filter Control

Respond

Actual

Value of

Parameter

Radar’s Knowlege of

Parameter

Tracking

Error

1. Basic functions performed by a single-target tracking loop.

Frequency

Gain

2. A tracking filter may be thought of as a low-pass filter whose

gain and cut-off frequency are adjusted to eliminate as much

noise as possible without introducing excessive lag.