RU2037841C1 - Method of optimal detection of pulse signals with nonmodulated carrier frequency - Google Patents

Abstract

FIELD: radio engineering. SUBSTANCE: signal is received when providing excess transmission band ΔF of the receiver in comparison with width of basic spectrum of pulse signal at duration τ_i,, where ΔF is higher or equal to 10/τ_i,. Signal received is delayed for time τ_к=τ_i to satisfy the condition f_sτ_к= d1/2 ,, where d is integer, f_s is signal carrier frequency. Nondelayed and delayed signals are summed and filtered subsequently in quasioptimal filter. Transmission band of the filter should satisfy ΔF_к relation, where ΔF_к is much lower than 1/τ_i,, and central frequency satisfies f_р=f_s relation. Envelope is separated and compared with threshold level. If threshold level is beaten, the signal is detected. EFFECT: improved noise immunity. 3 dwg

Description

 The invention relates to radio engineering and can be used in radar and in some other radio communication systems to detect signals under the influence of interference, especially fluctuation.

 The purpose of the invention is the improvement of noise immunity.

 Figure 1 shows the amplitude-frequency characteristic of the optimal filter (right branch); figure 2 the right branch of the spectral density of noise; figure 3 is a structural electrical diagram of a device that implements the method.

A device that implements the method includes an intermediate frequency amplifier (IFA) of receiver 1, a delay device for the interval τ _to τ _and 2, an amplifier compensating for losses in the delay device 3, adder 4, an amplifier loaded on a quasi-optimal filter 5, amplifier 6, detector 7 , a decisive (threshold) device 8, a driver shaper for “zeroing” the pulses 9, a device for “zeroing” the output circuits of the amplifiers (5 and 6) 10.

 The proposed method is as follows.

 It is known that methods for the optimal detection of pulsed signals are based on the operations of linear filtering consistent with the spectrum of the signal or one or another correlation processing. As a result of these operations, the highest signal-to-noise ratio (SNR) is provided, and hence the best signal detection quality according to one or another probabilistic criterion.

, (1) где Е энергия входного сигнала;
N_o const спектральная плотность (СП) шумовой помехи (односторонняя). При случайной амплитуде входного сигнала величины Е и q² носят усредненный характер. Выражение (1) по сути соответствует ОСШ
ρ

(2) в полосе пропускания приемника ΔF

; P_m 2Р пиковая мощность сигнала. В случае пассивной помехи отношение типа ρ на выходе резко повышается за счет обработки сигнала и помехи методами "сжатия" и СДЦ, но, естественно, зависит от уровня мощности этой помехи σ_п ² на входе (подавление помехи ограничено некоторой величиной К_п).According to well-known theoretical principles, this optimal result in the linear receiver system is completely determined by the highest SNR or the energy parameter of the signal plus fluctuation interference (noise) that undergoes optimal processing,
q ² =

, (1) where E is the energy of the input signal;
N _o const spectral density (SP) of noise interference (one-way). With a random amplitude of the input signal, the values of E and q ² are averaged. Expression (1) essentially corresponds to SNR
ρ

(2) in the receiver passband ΔF

; P _m 2P peak signal power. In the case of passive interference, the ratio of the ρ type at the output sharply increases due to signal processing and the interference using the “compression” and SDC methods, but, of course, depends on the power level of this interference σ _p ² at the input (the suppression of the noise is limited to a certain value of K _p ).

In modern conditions of increasing levels of interference, it is increasingly difficult to provide the required value of the parameter q ² . In fact, in a number of applications, technical capabilities are already limited due to the impossibility of creating transmitters and antenna-feeder paths to a large level of energy and, moreover, peak power, overall weight limitations, etc. Therefore, increasing noise immunity (PZ) due to more efficient processing of signals when they are detected is relevant and appropriate. And not only to overcome these limitations, but also in the economic sense (reduction in cost, size, power consumption, etc.).

The purpose of the invention is achieved by introducing additional signal processing operations when it is detected, the main part of which is the so-called extension antikogerentnoe the received pulse signal _and a duration τ and the corresponding optimal (quasioptimality) filtering the 1st pulse in the plurality of extended ν-multiple of the number of times the input signal. This ensures relatively higher requirements for the stability of the carrier frequency of the signal, which is ultimately offset by increased immunity to narrowband and harmonic interference.

(3)
Этот избыток полосы (по сравнению с оценочной шириной спектра входного сигнала) необходим для последующего четкого формирования импульсного сигнала увеличенной длительности. На результирующую полосу линейной части приемника этот "излишек" не влияет, так как в оконечной части КФ (или оптимальный фильтр ОФ) обеспечивается полоса
ΔF_к< ΔF_ос=

2ΔF_c= 2

< ΔF, (4) где ΔF_ос основной спектр ВЧ импульсного сигнала с прямоугольной огибающей (соответствует главному лепестку спектра);
ΔF_c оценочный спектр импульсного сигнала.In the receiving path, including the IF amplifier, provide an excess bandwidth ΔF, eliminating significant distortion of the received signal. Enough to have
ΔF ≳

(3)
This excess band (compared with the estimated width of the spectrum of the input signal) is necessary for the subsequent clear formation of a pulse signal of increased duration. This “surplus” does not affect the resulting strip of the linear part of the receiver, since a strip is provided in the end part of the CF (or optimal filter OF)
.DELTA.F _to <ΔF _a =

2ΔF _c = 2

<ΔF, (4) where ΔF _{os the} main spectrum of the HF pulse signal with a rectangular envelope (corresponds to the main lobe of the spectrum);
ΔF _{c is the} estimated spectrum of the pulse signal.

Consider the twofold expansion of the signal (ν 2) with the formation of two successive pulses of the same duration. In this case, the initial phase of the second pulse should differ from the final phase of the first by ϑ π. This is precisely the "anti-coherence" of the 2nd impulse with respect to the 1st impulse. Provided that the time shift between them _to τ _and τ equal and integer d
f _c ˙ τ _to d (5) where f _{c is the} carrier frequency of the signal, there is no phase jump (ϑ 0) at the junction between pulses, since an integer number of waves of frequency f _о fit within τ _and τ _k . Therefore, to fulfill the task of forming a 2-fold extended anticoherent signal when conditions (5) are fulfilled, the following operations are necessary:
1) the delay of the initial pulse of duration τ _and and the value of τ _to = τ _and ;
2) the addition of the initial pulse with a delayed inversion using an inverter in the delayed pulse circuit and the adder.

 Those. delay and subtraction.

. (6)
Поскольку последний вариант несколько проще, следует его считать основным.The same result of the formation of an extended anticoherent signal is achieved without the use of an inversion operation at a carrier frequency of the signal satisfying the condition
f _c · τ _k = d ±

. (6)
Since the latter option is somewhat simpler, it should be considered the main one.

 To determine the SNR, knowledge of the SP noise (interference) after the noted operations is necessary. According to the shape of the joint venture, on the basis of the existing theory of optimal linear filtering, you can choose the frequency response of the optical frequency or CF. However, in this proposal, the approach to such a choice has received significant development, fundamentally different from the traditional one.

4N_ocos

(7)
Здесь n_ζ(ω) СП входного случайного стационарного процесса (помехи) ζ(t), равная в данном случае N_o cosnt. Результат (7), естественно, подтверждается экспериментально. Квадрат амплитудного спектра сигнала S(i ω), после тех же операций на выходе сумматоре также изменится по закону 4cos

, что нетрудно показать. В общем это объясняется тем, что устройство задержки и сумматор, составляющие гребенчатый фильтр, в одинаковой степени преобразуют спектры сигнала и помехи. Поэтому и считается, что известное максимальное ОСШ ρ_т при линейной оптимальной (согласованной) фильтрации не ухудшается никаким другим фильтром по сравнению с ОФ (СФ)
ρ_т=

(8)
Здесь S(iω) спектр исходного сигнала, в этом случае величина ρ_тсоответствует традиционной оптимальной фильтрации;
n

'(ω) двухсторонняя СП на входе.During the operation of delay by τ _k and addition (see above), a noise SP (noise) is formed at the output of the adder, equal to
n _η (ω) 4n _ζ (ω) cos

4N _o cos

(7)
Here, n _ζ (ω) of the input random stationary process (noise) ζ (t) of the input process, which is equal in this case to N _o cosnt. The result (7), of course, is confirmed experimentally. The square of the amplitude spectrum of the signal S (i ω), after the same operations at the output of the adder, will also change according to the law 4cos

that is not difficult to show. In general, this is due to the fact that the delay device and the adder making up the comb filter transform the signal and interference spectra to the same extent. Therefore, it is believed that the known maximum SNR ρ _t with linear optimal (matched) filtering is not degraded by any other filter compared to the OF (SF)
ρ _t =

(8)
Here S (iω) is the spectrum of the initial signal, in this case the value ρ _t corresponds to the traditional optimal filtering;
n

'(ω) two-way joint venture at the inlet.

K(iω)

результат не меняется. При этом, естественно, речь идет об ОСШ на выходе сумматора для вновь образованного сложного сигнала, состоящего в данном случае из двух идентичных по форме импульсов, но с противоположными фазами (ϑ π) на стыке при τ_и τ_к.Indeed, when multiplied under the integral of the numerator and denominator by

K (iω)

the result does not change. In this case, of course, we are talking about the SNR at the output of the adder for the newly formed complex signal, consisting in this case of two identical in shape pulses, but with opposite phases (ϑ π) at the junction at τ _and τ _к .

The question arises, why, for optimal detection, it is necessary to take into account precisely the combination of these two separately existing pulses at the output of the adder? After all, the filter in the signal isolation circuit (CF, OF) does not “know” that the 2nd pulse will follow the 1st pulse. By the moment of the end of the 1st pulse (τ _to τ _and ), the output will have the maximum voltage corresponding to the matching of the filter with the parameters of the 1st pulse. This, by the way, clearly follows from the direct physically clear method for determining the filter output voltage of the Duhamel integral method, even if we take into account the entire sequence of both pulses. The spectral method, as a mathematical technique, just does not reveal it, since it takes into account the entire complex signal as a whole, not in parts, and besides, as a rule, it is much more complicated. However, if it is applied correctly, it is not limited to general considerations about the signal spectrum and filter band, i.e. adjusted to determine the time function U _o (t), the picture of the passage of a complex signal to the filter output should be the same as with the Duhamel integral technique. The main thing is that with the adopted spectral approach, the opportunity to determine the optimal SNR for the times of existence of each of the elementary pulses of a complex signal is missed, which corresponds to the principle of superposition.

Having determined the peak voltage (U _mk ) due to the 1st pulse, it is quite possible to judge the possibilities and features of subsequent detection. In this case, the action of the 2nd pulse (antiphase) is hardly conducive to obtaining a larger peak within the limits of the 2nd pulse (τ _k .2 τ _k ) than within the limits of the 1st pulse, since damped due to the antiphase of the 2nd pulse the free vibrations after the 1st are damped to a certain degree (for a filter such as a resonant LC circuit, this precisely follows if the signal frequency is equal to the resonant frequency of the filter). In general, we will assume that the peak of the filter output signal is determined by the 1st pulse and does not depend on the subsequent anticoherent pulse, which corresponds to calculations and experiment.

 As for the noise SP (noise), then it (7), determined by the introduced operations, acts at any moment during which the process ζ (t) takes into account delays. Those. (7) Certainly true for stationary interference.

·

:4

cos²ωτ_и/2

dω

·

:2N_osin²ω′τ_и/2

dω′ (9)

·

dx

C

-

∞ Здесь ω ω_к ω'; так как f_c τ_к=d

, то cos² ω' τ_и/2 sin² ω'τ_и/2; dω′ dx·

; τ_и τ_к.Based on the foregoing, we determine the SNR ρ and frequency response of the OF applied to the 1st pulse at the output of the adder and noise (7) according to the general expression (8)
π

·

:4

cos ² ωτ _and / 2

dω

·

: 2N _o sin ² ω′τ _and / 2

dω ′ (9)

·

dx

C

-

∞ Here, ω ω _to ω '; since f _c τ _k = d

, then cos ² ω 'τ _and / 2 sin ² ω'τ _and / 2; dω ′ dx

; τ _and τ _to .

 The result (9) is unexpected and only theoretical. The main limitation is the unrealizability of OFs with infinite gain at individual frequencies, and the noise of the adder circuits and subsequent circuits can, in principle, reduce the effect. The physical meaning of the result (9) will become more clear when applying CF.

, (10) где S*(i ω) комплексно-сопряженный спектр сигнала;
t_o момент возникновения максимального напряжения на выходе ОФ (в данном случае t_o τ_и).The frequency response of the OF is determined according to the expression
K _of (iω) C

, (10) where S * (i ω) is the complex conjugate signal spectrum;
t _{o the} moment of occurrence of the maximum voltage at the output of the OF (in this case t _o τ _and ).

K_оф(iω)

C

·

:2N_osin²ω′τ_и/2

C

C

(10a)
Вид правой ветви этой АЧХ при С₁ 1 показан на фиг.1 в пределах Х 0.π. На частотах f 0, ±

усиление теоретически должно быть бесконечным, что нереально. Для КФ практическое значение имеет полоса в области f' 0.The frequency response module of the equal to

K _of (iω)

C

·

: 2N _o sin ² ω′τ _and / 2

C

C

(10a)
The view of the right branch of this frequency response at C ₁ 1 is shown in FIG. 1 within X 0.π. At frequencies f 0, ±

gain theoretically should be infinite, which is unrealistic. For CF, the band in the region f '0 is of practical importance.

(f') 4N_o sin² π f' τ_к (11) Пунктиром показана правая граница КФ, полоса которого равна
ΔF_к= βΔF_к1; ΔF_к1=

(12)
ОСШ определяется как
ρ_п=

U_m·

S_i

:2

n_η(f′)df′ (13)
Здесь максимальная амплитуда сигнала на выходе КФ зависит от величин ΔF_к, τ_и, так как
Z

πΔF_к·τ_и= πβΔF_к1τ πβ (14) (величина Si

табулирована). В (13) n_η (f') определяется (11) при плотности N_o как мощности на 1 Гц полосы. Из физических соображений и смысла вида АЧХ ОФ наибольшие значения ОTШ будут в области ΔF_к_→ 0. Для малых аргументов z имеем
S_i

≈

Поэтому
ρ_п=

U_m·

·

:

2·4N

sin²πf′τ_иdf

(U_mβ)²:

in²xdx (U_mβ)²:

(15)

·

q

Здесь U_m ² τ_и 2Е. В пределе

m

=

m

q²

∞,(16) что соответствует ρ согласно (9) для ОФ. Повышение

1 достигается при весьма малых значениях β, т.е. рабочих полос КФ ΔF_к=

, например, при β= 0,01, ρ_п q² 30,4.Consider the traditional CF with a rectangular shape of the frequency response. Its central tuning frequency coincides with the signal frequency f _c and the central frequency of the noise band f '0, satisfying (6). The right branch of the noise SP (7) is shown in figure 2, where f 'ff _c . Moreover, the dependence of the joint venture on f 'takes the form
n

(f ') 4N _o sin ² π f' τ _to (11) The dotted line shows the right boundary of the CF, the strip of which is equal to
ΔF _k = βΔF _k1 ; ΔF _k1 =

(12)
SNR is defined as
ρ _p =

U _m

S _i

: 2

n _η (f ′) df ′ (13)
Here, the maximum amplitude of the signal at the output of the CF depends on the values ΔF _k , τ _and , since
Z

πΔF _k · τ _and = πβΔF _k1 τ πβ (14) (the value of Si

tabulated). In (13), n _η (f ') is determined by (11) at a density of N _o as a power per 1 Hz band. From physical considerations and meaning of the form of the frequency response of the OF, the highest values of the OTN will be in the region ΔF _to _ → 0. For small arguments z we have
S _i

≈

therefore
ρ _p =

U _m

·

:

2 · 4N

sin ² πf′τ _and df

(U _m β) ² :

in ² xdx (U _m β) ² :

(fifteen)

·

q

Here U _m ² τ _and 2Е. In the limit

m

=

m

q ²

∞, (16) which corresponds to ρ according to (9) for the OF. Increase

1 is achieved at very small values of β, i.e. working strips KF ΔF _to =

, for example, with β = 0.01, ρ _p q ² 30.4.

(17)
При этом U_mк 1,062 U_m и
ρ_o= (1,062U_m)²:

N_o·

0,82

0,82q² (18) что на 0,85 дБ ниже оптимального значения ОСШ ρ_т.Known CF at n (f ') N _o has a band
ΔF _opt =

(17)
Moreover, U _mk 1,062 U _m and
ρ _o = (1,062U _m ) ² :

N _o

0.82

0.82q ² (18) which is 0.85 dB lower than the optimal SNR ρ _t .

Thus, assuming a correspondingly small value of β, ρ _p > ρ _{t is} achieved, especially ρ _p > ρ _o .

It should be noted that for small values of ΔF _to ˙ τ _and = β, the voltage at the output of the CF is practically proportional to the amplitude of the input voltage U _m only for limited types of CF, for example, for the considered ideal, one- and two-stage type of resonant LC circuit, single-stage on coupled LC -contours (with two-humps with a relative level of the depression of 0.7 and with critical connection), which follows from the calculated curves for establishing the voltage at the output of the CF. In the case of real types of CF with multi-stage and small β, the output signal (and noise) levels nonlinearly depend on the input levels, and the question of decreasing the SNR growth requires special studies. In general, when applying real types of CF, it is necessary to calculate the SNR value, taking into account their frequency response.

, где α затухание контура; Δω расстройка частоты сигнала ω_сотносительно резонансной частоты контура ω_р. ЧС многоконтурных фильтров, как известно, в полной мере проявляется при β

0,5.Since relatively narrow-band filters are used as KFs, the question naturally arises of the high frequency selectivity (ES). In this case, the mode of small values of β ΔF _to ˙ τ _{and is used} . In this mode, it is somewhat known that the emergency situation is preserved when applying filters such as a resonant LC circuit, since the amplitude of the established voltage at the output is proportional to the factor 1 /

where α is the attenuation of the circuit; Δω is the frequency detuning of the signal ω _c relative to the resonant frequency of the circuit ω _p . ES of multi-loop filters, as is known, is fully manifested when β

0.5.

ω_p и близости АЧХ к основной области АЧХ ОФ на частоте f'=0 при f_c, соответствующей условию (6) (фиг.1). Это обусловит высокое ОСШ (ρ_к > ρ_т).Since the frequency response of a simple quartz filter is practically equivalent to the frequency response of a single LC circuit (in certain applications, QFs such as resonant LC circuits can be used), using this QF, we can expect a relatively large gain at a frequency ω _c

ω _p and the proximity of the frequency response to the main region of the frequency response of the OF at a frequency f '= 0 at f _c corresponding to condition (6) (Fig. 1). This will cause a high SNR (ρ _to > ρ _t ).

Obtaining the growth of SNR (ρ _p > ρ _t ) is also possible when forming a multi-pair sequence of adjacent anti-coherent signals (ν> 2) and choosing the signal frequency f _c according to (6). However, the study of this possibility showed its relatively lower efficiency than when ν = 2 (by the value of ρ _p for the same value of ΔF _k ). Therefore, the case ν> 2 is not considered.

From the output of the amplifier of the receiver 1, the bandwidth ΔF of which (as well as the subsequent circuits 2, 3, and 4) corresponds to condition (3), a working pulse signal of duration τ _and at an intermediate frequency f _{k is} supplied to delay device 2 and adder 4. Delay device, practically without distorting the pulse signal, it delays it for a time τ _to τ _and . The value of τ _to and the frequency of the signal f _c must satisfy condition (6). In this case, the delayed RF pulse arriving at adder 4 will differ in the initial phase from the phase of the end of the undelayed pulse by ϑ = π.

5·10^-5. Изменение частоты Δf_с 50 Гц и частоты центра "ямы" СП шума на 5·10

25 Гц весьма незначительны по отношению к полосе КФ ΔF_к=10⁴ Гц, что практически не повлияет на эффективность достижимую величину ОСШ. Величин 5 ˙ 10^-5 легко достижима при кварцевой стабилизации частоты f_c (на всех этапах) и при применении в устройстве задержки ультразвуковой ЛЗ на кварце. Требование к стабильности может быть снижено, так как реальная избирательность КФ при β<< 1 остается на уровне полосы порядка 1/τ_и. Термостабилизация не требуется при установочной точности и температурном коэффициенте 10^-6.The stability of the delay and frequency f _c determines the accuracy of combining the frequency of the signal with the center of the noise “well” of the SP noise (7). So, if τ _and τ _to 10 ^-6 s, then the width between the ridges of this joint venture is Δf _g = ΔF _k1 1 / τ _to 10 ⁶ Hz. When β = ΔF _to τ _and 0.01, i.e. at ΔF _k β ΔF _k1 10 ⁴ Hz the stability and accuracy of the delay τ _k and frequency f _c 5 ˙ 10 ^-5 corresponds

5 · 10 ^-5 . The change in the frequency Δf _from 50 Hz and the frequency of the center of the "hole" SP noise by 5 · 10

25 Hz are very insignificant with respect to the CF band ΔF _k = 10 ⁴ Hz, which practically does not affect the achievable SNR value. Values of 5 ˙ 10 ^{-5 are} easily achievable with quartz stabilization of the frequency f _c (at all stages) and with the use of ultrasonic laser radiation on quartz in the delay device. The stability requirement can be reduced, since the real selectivity of CF at β << 1 remains at the level of a band of the order of 1 / τ _and . Thermal stabilization is not required with installation accuracy and a temperature coefficient of 10 ^-6 .

(ω) (7)
С выхода сумматора 4 пара антикогерентных импульсов поступает на усилитель 5, нагруженный на КФ. Полоса пропускания КФ ΔF_к в общем выбирается из требуемого отношения β ΔF_к:

ΔF_к·τ_и для обеспечения нужного значения ОСШ с учетом конкретного вида АЧХ КФ. Тип КФ должен удовлетворять требованию пропорциональности амплитуды выходного напряжения при малых β U_mвых U_m ˙ A ˙ ΔF_к ˙ τ_и, где А коэффициент пропорциональности. Если в качестве КФ используется простой кварцевый резонатор (работающий на основной гармонике), то, как известно, номинал частоты f_c ≅ 9.10 МГц. Применение кварца гарантирует высокую стабильность его резонансной частоты f_p f_c, что важно ввиду малости полосы ΔF_к << f _c. Кроме того, такое состояние требует очень большого значения Q f_p ΔF_к, что как раз и может быть удовлетворено в КФ на базе кварцевого резонатора. Так, при f_с 10 МГц и ΔF_к 10⁴ Гц Q 1000.Amplifier 3 (non-inverting) is used in the delayed pulse summation circuit to compensate for losses in the ULZ. It should be noted that the gain control of this amplifier (which does not invert the phase of the signal) is intended to equalize the intensities ζ (t) and ζ (t- τ _к ) in order to correctly form (with "zero") SP n

(ω) (7)
From the output of the adder 4, a pair of anti-coherent pulses is fed to amplifier 5 loaded on CF. The passband KF ΔF _to generally selected from the desired ratio β ΔF _to :

ΔF _to · τ _and to provide the desired SNR value taking into account the specific type of frequency response of the CF. The type of CF should satisfy the requirement of proportionality of the amplitude of the output voltage for small β U _mout U _m ˙ A ˙ ΔF _to ˙ τ _and , where A is the proportionality coefficient. If a simple quartz resonator (operating at the fundamental harmonic) is used as a CF, then, as is known, the frequency rating f _c ≅ 9.10 MHz. The use of quartz guarantees high stability of its resonant frequency f _p f _c , which is important due to the smallness of the band ΔF _to << f _c . In addition, such a state requires a very large value of Q f _p ΔF _k , which just can be satisfied in CF based on a quartz resonator. So, at f _with 10 MHz and ΔF _to 10 ⁴ Hz Q 1000.

. Полоса цепей нагрузки детектора 7, естественно, должна соответствовать не ΔF_к, а ΔF_у. При этом следует иметь в виду, что из-за малого эквивалентного коэффициента затухания контура кварцевого КФ α π ˙ ΔF_к "хвост" свободных колебаний после прохождения через него пары антикогерентных импульсов сигнала будет относительно затянут по закону e

для узкополосных цепей. Т.е. сказанное выше о широкой полосе (ΔF_у) усилителя и детектора ( ≳

F_у) обусловлено необходимостью не ослабить амплитуду нарастающего по фронту 1-го сигнала на выходе КФ 5 сформированной пары импульса на выходе сумматора 4.From the output of CF 5, the signal (with relatively suppressed noise, see, e.g., (15), for a real frequency response of quartz CF, the result can be even higher if the gain at frequency f _s is very high) is fed to an amplifier matched for operation with the detector 7. The passband of the amplifier is much wider than the band ΔF _to KF 5 in order to realize its load on a simple LC circuit to eliminate the influence on the resulting passband determined by ΔF _k and transmit the signal with increasing signal amplitude of duration τ _and . That is, ΔF _y ≈

. The load band of the detector 7, of course, should correspond not to ΔF _k , but ΔF _y . It should be borne in mind that, due to the small equivalent attenuation coefficient of the quartz CF contour, α π ˙ ΔF _{k the} “tail” of free oscillations after passing through it a pair of anticoherent signal pulses will be relatively extended according to the law

for narrowband circuits. Those. what was said above about the wide band (ΔF _y ) of the amplifier and detector (≳

F _y ) due to the need not to weaken the amplitude of the rising along the front of the 1st signal at the output of CF 5 of the generated pulse pair at the output of the adder 4.

 From detector 7, the signal arrives at a decisive (threshold) device 8, tuned, for example, to the desired probability of false alarm according to the Neumann-Pearson criterion.

From the output of the resolving device 8, the standard pulse generated by it of the required amplitude and duration goes to the channel for processing the received information (target for specific communication systems) and to the driver shaper for “zeroing” the pulses 9. The latter, in turn, generates control pulses of the desired polarity, amplitude and duration ( order τ _u) which, through the device "zeroing" KF output circuit 5 and the amplifier 6 after the actuation decision unit 8 quenched free oscillation circuit stages 5 and 6 for the duration control pulses. After this consideration, the implementation scheme of the proposed method is again ready for processing the next working input signal of duration τ _and . In general, this time will be of the order of 2 τ _and , which almost does not reduce the real throughput (SP) of the communication system equipment. Thus, the presence in this method during its implementation of narrow-band circuits does not affect the PS.

Note that instead of the usual envelope detector 7, it is possible to use a correlator (with a known initial phase of the signal) or a quadrature correlator with an unknown phase. In this case, the narrow-band CF is used in the output video chains of the correlator. CF can be built on L- and C-elements. However, it is more difficult to perform multi-channel (in frequencies f _s ) hardware systems, the dimensions are increasing.

 It follows from the essence of the method of optimal signal detection against fluctuation noise that the interference can also be a fluctuation noise of the passive type, “non-white” (in terms of spectral uniformity) and not necessarily with a normal distribution. Under certain restrictions on the speed and duration of changes in intensity (power), this can be an unsteady interference.

Claims (1)

METHOD FOR OPTIMUM DETECTION OF PULSE SIGNALS WITH AN UNMODULATED CARRIER FREQUENCY, at which the signal is received, it is optimally or quasi-optimally filtered, the envelope is extracted, compared with the threshold level and, when the threshold level is exceeded, the signal is distinguished, which, in order to increase the noise, provides noise the excess passband ΔF of the receiver compared to the width of the main spectrum of the pulse signal of duration τ _and

the received signal is delayed for a time τ _k = τ _and when the condition

where d is an integer;
f _{with the} carrier frequency of the signal,
summarize the delayed and delayed signals, followed by filtering in a quasi-optimal filter, the passband ΔF _to which satisfies the condition
ΔF _to ≪ 1 / τ _and ,
and the center frequency f _p f _s .

Images