RU2767183C1 - Method for automatic detection of narrowband signals - Google Patents

Abstract

FIELD: radio engineering.

SUBSTANCE: invention relates to radio engineering and can be used for automatic detection of narrow-band signals in conditions of a priori uncertainty about their time parameters. Claimed technical result is achieved due to successive use of reference signals, duration of which gradually increases, to calculate based on them cross-correlation functions with processed input implementation, containing useful signal, calculating average values of components of cross-correlation functions and fixing that value, starting from which average values of components of cross-correlation functions cease to change by more than five percent. Duration of the reference signal corresponding to this value is determined as time parameters of the detected useful signal, i.e. useful signal is considered to have duration equal to duration of reference signal.

EFFECT: detection of useful signal and determination of its duration.

1 cl, 4 dwg

Description

The invention relates to radio engineering, namely to methods for automatic detection of narrow-band signals under conditions of a priori uncertainty about their time parameters.

Known "Method for automatic detection of signals" (RF Patent No. 2480901 H04L 27/22 (2006.01), publ. 27.04.2013, Bull. No. 12)

In a known method, an analog signal z(t) is received, it is digitized, for which the operations of sampling, quantization and coding are sequentially performed, a spectral representation Fj is formed, where j=1, 2, ... is the serial number of the spectral component of the digitized signal zi, where i =1, 2, ... - the serial number of the time reference, then calculate the parameters of the spectral representation Sj, the values of which calculate the threshold value of the noise level G, compare the parameters of the spectral representation Sj with the calculated threshold noise level G and, based on the results of the comparison, decide on the fact of detection signal, characterized in that when forming the spectral representation Fj, the digitized signal zi is preliminarily divided into N equal fragments, over which the Fourier transform {F1j, F2j, ..., FNj} is independently performed, and the maximum values are chosen as the parameters of the spectral representation Sj components of the Fourier transform of each of the N frags ments, and the decision on the fact of signal detection is made if the parameters of the spectral representation of at least one of the fragments exceed the threshold value of the noise level G. And the threshold value of the noise level G is calculated separately for each of the N fragments and is equal to the triple value of the averaged sum of the spectral components of the fragment.

The disadvantage of this method is the limited scope, since its implementation does not provide a determination of the duration of the detected useful signal.

Known "Method for automatic detection of narrow-band signals" (RF Patent No. 2382495, H04B 1/10 (2006.01), publ. 20.02.2010 Bull. No. 5).

In a known method, an analog signal z(t) is received, it is digitized, for which the operations of sampling, quantization and encoding are sequentially performed, then the parameters of the digitized signal zi are calculated, the obtained parameters are compared with a predetermined threshold value, and based on the results of the comparison, a decision is made about the fact of detection signal, characterized in that to calculate the parameters of the digitized signal zi, its spectral representation Fj is formed, by performing a Fourier transform on it, the threshold noise level U is calculated, by calculating the doubled value of the sample mean component of the spectral representation Fj, the levels of each of the spectral components from the sequence are compared spectral representation Fj with the calculated threshold noise level U, and the first F1j and second F2j sequences are formed, respectively, from the spectral components Fj that exceeded the threshold noise level U and did not exceed it, then separately summarize the components included in the first ∑F1 and second ∑F2 sequences, after which the ratio R is calculated as the ratio of the found sums R=∑F1/∑F2 and compared with a predetermined threshold value Rthr, and the decision on the fact of signal detection is made provided that R>Rthr, while the threshold value Rthr is selected in the interval Rthr=0.13-0.15.

The disadvantage of this method is the limited scope, since its implementation does not provide a determination of the duration of the detected useful signal.

Closest to the claimed method is the "Method for automatic detection of narrow-band signals" (RF Patent No. 2419968, H04B 1/10 (2006.01). Published: 05/27/2011. Bull. No. 15].

In the prototype method, an analog signal z(t) is received, digitized, for which the operations of sampling, quantization and coding are sequentially performed, then the parameters of the digitized signal zi are calculated, the obtained parameters are compared with the pre-calculated threshold noise level U and, based on the results of the comparison, a decision is made about the fact of signal detection, to calculate the parameters of the digitized signal z, it is divided into two equal sequences: z1 and z2, corresponding to the first and second halves of the digitized signal z. The cross-correlation function K between the sequences z1 and z2 is calculated and its spectral representation Fj is formed, where j=1, 2, … are the numbers of the spectral components of the cross-correlation function, by performing a Fourier transform on it. The threshold value of the noise level U is calculated by multiplying the average value of the components of the spectral representation Fj by the coefficient Q. The level of each of the spectral components Fj is compared with the pre-calculated threshold noise level U, and if at least one of the jth components of the condition Fj >U fix the fact of detection of a narrowband signal. The Q value is chosen in the interval Q=3.5-4.5.

The disadvantage of this method is the limited scope, since its implementation does not provide a determination of the duration of the detected useful signal.

The objective of the invention is to create a method that makes it possible to determine the duration of the detected useful signal, which is in the processed input implementation (sample).

The technical result of the claimed invention is the detection of a useful signal and the determination of its duration.

The technical result is achieved by the fact that in the known method of automatic detection of narrow-band signals, which consists in receiving an analog signal, digitizing it, for which the operations of sampling, quantization and coding are sequentially performed, the cross-correlation function is calculated, calculations are made, comparison is made, according to the results comparisons make a decision about the fact of signal detection, preliminarily form reference signals of different duration, the first reference signal is formed with the shortest duration, and the duration of each subsequent generated reference signal is gradually increased, after which, for each generated reference signal, the cross-correlation function is calculated with the received digitized analog signal , calculate the average value of the components of each of the calculated cross-correlation functions, compare the calculated average values of the components with each other, and the decision based on the comparison results about the fact signal detection is taken in the event that the value of each subsequent of the successively calculated average values of the components is not less than the value of the previous one, then among the successively calculated average values of the components, the value is fixed, starting from which it ceases to change by more than five percent, and the time parameters of the reference the signal that was used in the calculation of the cross-correlation function, the average value of the component values of which were fixed, is considered as time parameters, that is, the duration of the detected useful signal.

Due to the new set of essential features in the claimed method, a useful signal is detected and its duration is determined, which is considered as time parameters.

The claimed effect is achieved through the consistent use of reference signals, the duration of which gradually increases, to calculate on their basis the cross-correlation functions with the processed input implementation containing the useful signal, calculate the average values of the components of the cross-correlation functions and fix the value, starting from which the average values of the components of the functions cross-correlations stop changing by more than five percent. And the duration of the reference signal corresponding to this value is determined as the time parameters of the detected useful signal, i.e. consider that the useful signal has a duration equal to the duration of the reference signal.

The claimed method is illustrated by drawings, where:

fig. 1 - temporal implementation of the cross-correlation function Kzu(τ) of the digitized analog signal z(t), containing the useful signal s(t), and the reference signal u(t), provided that the duration u(t) is equal to the duration s(t) , i.e. u(t)=s(t); temporary realization s(t); temporary realization u(t);

fig. 2 - temporal implementation of the cross-correlation function K1zu(τ) of the digitized analog signal z(t), containing the useful signal s(t), and the reference signal u1(t), provided that the duration u1(t) is equal to half the duration s(t ), i.e. u1(t)≠s(t); temporary realization s(t); temporary realization u1(t);

fig. 3, which shows: the temporal implementation of the cross-correlation function K2zu(τ) of the digitized analog signal z(t), containing the useful signal s(t), and the reference signal u2(t), provided that the duration u2(t) is one and a half times longer than the duration s(t), i.e. u2(t)≠s(t); temporary realization s(t); temporary realization u2(t);

fig. 4, which shows: the temporal implementation of the cross-correlation function K3zu(τ) of the digitized analog signal z(t), containing the useful signal s(t), and the reference signal u3(t), provided that the duration u3(t) is only three quarters of the duration s(t), i.e. u3(t)≠s(t); temporary realization s(t); temporary realization u3(t).

The correlation method for detecting signals has been successfully used in various areas of radio engineering. For discrete signals X and Y , their cross-correlation function can be represented by the following expression, see [Sato Yukio. Signal processing. First meeting. - M.: Dodeka XXI Publishing House, 2002. - 176 p.]:

where N is the number of time samples; j is the current shift record of the correlated signal Y .

Further in FIG. 1-fig. Figure 4 shows the dependence of the cross-correlation functions on the argument τ, where τ is the correlation interval (in formula (1), the cross-correlation function is presented as a function of the current shift parameter j).

Detection of the useful signal s(t) is carried out according to the results of calculating the cross-correlation function Kzu(τ) of the input processed implementation z(t) with the reference signal u(t).

In this case, two situations are possible: the input processing implementation z(t) contains a useful signal s(t); the input processing realization z(t) does not contain the useful signal s(t).

In the case when the input implementation z(t) contains a useful signal s(t), the fact of its detection can be fixed by a fragment of the contour of the cross-correlation function Kzu(τ), which has a pronounced rise and fall in amplitude. If the parameters of the reference signal u(t) coincide with the parameters of the useful signal s(t), i.e. u(t)=s(t), then the cross-correlation function Kzu(τ) has one maximum, the time value of which corresponds to the beginning of the detected useful signal in the processed input implementation.

As an example, in FIG. 1 shows: the cross-correlation function Kzu(τ) obtained under the condition that u(t)=s(t); reference signal u(t); detectable useful signal s(t). In this case, the time parameters of the signals u(t) and s(t), i.e. their durations are the same. The maximum of the function Kzu(τ), indicated in Fig. 1 as Kmax, coincides with the beginning of the arrival of the useful signal s(t).

If the time parameters of the reference signal and the useful signal do not match, i.e. u(t)≠s(t), then two situations are possible:

the duration of the test signal is less than the duration of the useful signal;

the duration of the test signal is greater than the duration of the useful signal.

In the first case, the cross-correlation function Kzu(τ) will have the shape of a truncated triangle. And in the second case, the form of the cross-correlation function will coincide with the form of the cross-correlation function obtained under the condition that u(t)=s(t).

As an example, in FIG. 2 shows: the cross-correlation function K1zu(τ), obtained under the condition that the duration of the useful signal is twice the duration of the reference signal, i.e. u1(t)≠s(t); reference signal u1(t), the duration of which is two times less than the duration of the useful signal; detectable useful signal s(t).

As an example, in FIG. Figure 3 shows the cross-correlation function K2zu(τ) obtained under the condition that the duration of the useful signal is one and a half times less than the duration of the reference signal, i.e. u2(t)≠s(t); reference signal u2(t), the duration of which is one and a half times the duration of the useful signal; detectable useful signal s(t).

Analysis of the results shown in FIG. 1, fig. 2, fig. 3 shows that if the duration of the reference signal is less than the duration of the useful detected signal, then the function of their cross-correlation does not have a pronounced maximum. And if the duration is u(t)>s(t), or u(t)=s(t), then the function of their mutual correlation has a pronounced maximum.

The identified feature is the basis of the proposed technical solution aimed at choosing the temporal parameters of the reference signal in such a way that the cross-correlation function formed on its basis with the processed sample containing the useful signal has a maximum value and thereby contributes to determining the duration of the detected useful signal.

Indeed, if we successively increase the duration of the generated reference signal and at the same time monitor changes in the contour of the generated cross-correlation function with the sample containing the useful signal, then we can identify the moment from which the contour of the generated cross-correlation function ceases to change.

In this case, the duration of the generated reference signal, corresponding to the moment from which the contour of the generated cross-correlation function ceases to change, will correspond to the duration of the useful signal in the processed sample.

The implementation of the claimed method is as follows.

1. Reference signals of different duration are preliminarily formed, the first reference signal being formed with the shortest duration, and the duration of each subsequent generated reference signal is gradually increased.

As an example, in FIG. 1-fig. 4 shows generated signals u(t), u1(t), u2(t), u3(t) with different duration, which is defined as the time parameters of the reference signal. The step of changing the duration can be determined by the value of the temporary discrete report (in Fig. 1-Fig. 4 the x-axes are presented in the values of temporary discrete readings). Gradual increase means that for each successive reference signal, the duration increases by one temporal discrete value.

The procedures for generating reference signals are known, and, for example, are discussed in, see [Method for recognizing radio signals, RF Patent No. 2423735, Published: 10.07.2011 Bull. No. 19].

2. An analog signal is received, digitized, for which the operations of sampling, quantization and coding are sequentially performed.

The input implementation is taken in the form of an analog signal z(t), which may contain either only noise x(t), i.e. z(t)=х(t), or useful signal s(t) and noise x(t), i.e. z(t)=s(t)+x(t).

Reception can be carried out, for example, from the intermediate or low frequency path of the radio receiver.

The procedures for receiving an analog signal are known, and, for example, are discussed in, see [Method for recognizing radio signals, RF Patent No. 2356064, Published: 20.05.2009. Bull. No. 14].

The procedures for sampling, quantization and coding of an analog signal are known, and, for example, are discussed in, see [Method of recognizing radio signals, RF Patent No. 2261476, Published: 27.09.2005. Bull. No. 27].

3. For each generated reference signal, the cross-correlation function with the received digitized analog signal is calculated.

As an example, in FIG. 1-fig. Figure 4 shows the calculated values of the cross-correlation functions, denoted as Kzu(τ), K1zu(τ), K2zu(τ), K3zu(τ).

Procedures for calculating the cross-correlation function of signals are known, for example, see [Method for automatic detection of narrowband signals. RF patent No. 2419968, H04B 1/10 (2006.01). Published: 05/27/2011. Bull. No. 15]. In those shown in FIG. examples, the calculation of cross-correlation functions was carried out in accordance with formula (1).

4. Calculate the average value of the components of each of the calculated cross-correlation functions.

Procedures for calculating the average values of the signal components are known, for example, see [Method for detecting signals without a carrier. RF patent No. 2484581, H04B 1/10 (2006.01). Published: 06/10/2013. Bull. No. 16].

5. Compare the calculated average values of the components with each other.

Comparison procedures are known, for example, see [Method of jamming communication channels. RF patent No. 2450458, H04K 3/00 (2006.01), Published: 05/10/2012. Bull. No. 13].

6. The decision based on the results of the comparison about the fact of detecting a signal is made if the value of each subsequent of the successively calculated average values of the components is not less than the value of the previous one,

For the example shown in FIG. 1-fig. 4, the arrangement of reference signals in ascending order of their duration is as follows. The reference signal has the shortest duration u1(t). Then u3(t), u(t), u2(t) follow in ascending order. Accordingly, the average values of their cross-correlation functions have the following values: Sr1=0.08; Sr3=0.12; Sr=0.16; Sr2=0.16.

The specified average values of the cross-correlation functions correspond to the condition that the value of each of the subsequent values is not less than the value of the previous one, which corresponds to the signal detection condition.

In the course of the experiment to test the proposed technical solution, modeling was carried out for various situations, including the absence of a useful signal in the processed implementation. In all cases, in the absence of a useful signal in the processed sample, the condition that the value of each of the subsequent values is not less than the value of the previous one was violated, i.e. indicated the fact that the useful signal was not detected. And in situations where a useful signal was present in the processed sample, the condition that the value of each of the subsequent values is not less than the value of the previous one was observed.

The simulation was carried out in the range of signal-to-noise ratio (SNR) in the processed sample from 8 dB to 20 dB, with a change step of 1 dB (Fig. 1-Fig. 4 shows the results at SNR=20 dB).

For each change in the SNR value, 200 experiments were conducted in each of the situations (i.e., in the presence and absence of a useful signal in the processed implementation), which meets the requirements for calculating statistical estimates, see [Mathematical Encyclopedic Dictionary. M.: Sov. Encyclopedia, 1988. 847 p.; G. Korn, T. Korn. Handbook of mathematics. Per. from English. - M.: Nauka, 1977, p. 638-643].

7. Among the successively calculated average values of the components, fix the value from which it ceases to change by more than five percent.

For the example under consideration, this is the value of Sr=0.16, i.e. corresponds to the reference signal u(t), see FIG. one.

8. The time parameters of the reference signal, which was used in the calculation of the cross-correlation function, the average value of the component values of which were fixed, are considered as time parameters, i.e. the duration of the detected useful signal.

That is, they decide that the duration of the useful signal in the processed sample coincides with the duration of the reference signal.

As an example, in FIG. 1 shows the situation in which the duration of the reference signal u(t) coincides with the duration of the useful signal s(t).

Thus, thanks to a new set of essential features in the claimed method, a useful signal is detected and its duration is determined.

Also, the claimed method allows you to select the time parameters of the reference signal in such a way that the cross-correlation function formed on its basis with the processed sample containing the useful signal has a clearly defined maximum value and makes it possible to determine the duration of the detected useful signal.

Claims (1)

A method for automatic detection of narrowband signals, which consists in receiving an analog signal, digitizing it, for which the operations of sampling, quantization and coding are sequentially performed, the cross-correlation function is calculated, calculations are made, a comparison is made, a decision is made on the fact of signal detection based on the results of the comparison, characterized in that reference signals of different duration are preliminarily formed, the first reference signal being formed with the shortest duration, and the duration of each subsequent generated reference signal is gradually increased, after which, for each generated reference signal, the cross-correlation function with the received digitized analog signal is calculated, the average value is calculated components of each of the calculated cross-correlation functions, the calculated average values of the components are compared with each other, and the decision based on the results of the comparison about the fact of signal detection is taken at that m case, if the value of each subsequent of the successively calculated average values of the component is not less than the value of the previous one, then among the successively calculated average values of the components, the value is fixed, starting from which it ceases to change by more than five percent, and the time parameters of the reference signal that was used when calculating the cross-correlation function, the average value of the values of the components of which were fixed, is considered as time parameters, that is, the duration of the detected useful signal.

Images