WO2021040561A1

WO2021040561A1 - Method for signal extraction with frequency shift keying using square components and compensation of combination components

Info

Publication number: WO2021040561A1
Application number: PCT/RU2019/000964
Authority: WO
Inventors: Vladimir Aleksandrovich BELOGUROV; Vladimir Alekseevich ZOLOTAREV
Original assignee: Joint-Stock Company "Concern "Sozvezdie"
Priority date: 2019-08-28
Filing date: 2019-12-17
Publication date: 2021-03-04
Also published as: RU2709182C1

Abstract

This invention relates to radio engineering and may find application in communications. The technical result is an increase of the interference immunity of communications. This is achieved by generating a frequency-shift modulated signal (FSK), setting the signal frequency values so that the difference value of any pair of frequencies does not exceed the frequency value for which the difference in the amplitude-frequency characteristic (AFC) of the low-pass filter (LPF) and the bandpass filter (BF) becomes less than a certain value. After multiplying the signal and interference by reference signals (sine and cosine components) in each of the parallel bars (devices), the resulting signal is branched into two identical components. The first component is filtered by a low-pass filter, while the second component is filtered by a band-pass filter. The filter bands are matched with the signal band. Signals that have passed the low-pass filter and a band-pass filter are subtracted from one another. The received signals are converted to digital form. From the values of the quadratures, the spectral density (SP) for each subcarrier is determined proportional to the amplitude of the signals. Find the spectral density (SP) with the maximum value and the threshold value by multiplying this value of the spectral density of the spectral density (SP) by a certain coefficient. Based on the of spectral density values comparing results with a threshold, a conclusion is made about the presence of a signal with an appropriate frequency. The effectiveness of the proposed method in terms of signal / noise power ratio exceeds the efficiency of the prototype method by almost 6.7 times. The signal bandwidth ratio of the proposed method and of the prototype method is 0,1. 2 fig.

Description

METHOD FOR SIGNAL EXTRACTION WITH FREQUENCY SHIFT KEYING USING SQUARE COMPONENTS AND COMPENSATION OF COMBINATION COMPONENTS Method for signal extraction with frequency shift keying using square components and compensation of combination components

Field of the invention

This invention relates to radio engineering and may find application in communications.

Background of the invention

Known methods implementing by devices for suppressing broadband interference, are described in patents RU 2115234, H04B 1/10 , RU 2143783, H04B 1/10, RU 2190297 H04B 1/10. The disadvantage of these methods is a low degree of interference suppression.

Known method of signals extraction in presence of interference is described in the patent RU N° 2675386 H04B 1/10. The disadvantage of this method is its low efficiency when using a multiffequency signal.

The amplitude and angle modulations are described in the tutorial “Basis of the radio systems theory. Tutorial. // V.I. Borisov, V.M. Zinchuk, A.E. Limarev, N.P. Mukhin. Edited by V.I. Borisov. Voronezh Scientific Research Institute of Communications, 2004. ”, pp. 165 - 168, 170 - 174, respectively, the disadvantage of which is their low efficiency under the influence of noise.

Known methods for digital signal processing: with amplitude-pulse modulation (ASK), quadrature amplitude modulation (QAM), frequency shift keying (PSK), described in the book "Prokis John," Digital Communication ". Translation from English / Edited by D. D. Klovsky. - Moscow.: Radio and communication. 2000, p.: 148 - 152, respectively, the disadvantage of which is low efficiency under influence of noise. The closest analogue in technical essence to the proposed one is the method, which comprises using modulation with orthogonal frequency shift (FSK) (multiplexing with orthogonal frequency division of channels (OFDM)) and signal extraction using the optimal maximum likelihood of the detector, described in the book "Prokis John," Digital Communication Translation from English / Edited by / Ed. D. D. Klovsky. - Moscow.: Radio and communication. 2000, p. 141, 208, 219-221, 593-596, taken as a prototype.

The prototype method is as follows.

Using frequency shift keying method, M orthogonal signals of equal energy are formed. These signals vary in frequency

where: m=l,2,...,M;

0<t<T; e - signal energy;

T - signal change period corresponding to the minimum value of the signal spectrum frequency; f_c- signal frequency;

Df- frequency shift between signals.

The equivalent low-frequency signal can be represented as

These signal waveforms are characterized by equal energy and crosscorrelation coefficient, the real component of which is equal to

Since the case | m-k | =1 corresponds to the adjacent frequency intervals, then Df=1/(2T) represents the minimum value of the frequency separation between adjacent signals for M orthogonal signals.

An additive mixture of signal and noise is received at the input of the receiver

where: U_s - signal generated using frequency shift keying;

U_p - noise.

After multiplication by the corresponding reference signals S_oi in multiplier and integrator units, integrators at the outputs of the integrators produce the result of signal and noise transformation, i.e. multiplication by a reference signal and integration (correlation metrics):

where

are signal and noise transformation coefficients, respectively, depending on orthogonal functions system type used.

In the selection device, the maximum signal is selected that corresponds to the largest correlation metric.

The disadvantage of the prototype method is not sufficiently high efficiency under the influence of noise, which is explained by a wide signal spectrum and a high transformation noise level.

Disclosure of the invention

An object of the proposed method is to increase the signal extraction efficiency under noise influence by reducing the signal spectrum width and transformation noise level.

To solve the problem in a method for signal extraction with frequency shift keying using quadrature components and compensation of combination components, a signal is formed, consisting of several harmonic signals using a modulation with a frequency shift (FSK), after multiplication by the respective reference signals (on the sine or cosine component) in multiplier units the result of signal and noise transformation is formed, according to the invention , frequency shifts between adjacent signals (subcarriers) are set in advance so that the difference value of any pair of frequencies does not exceed the frequency value for which the difference in the amplitude-frequency characteristic (AFC) of the low-pass filter (LPF) and of the band-pass filter becomes less than a certain predetermined value, the number of frequencies N_f used is set in advance, the additive sum of signal and noise components is branched into n, where n= N_f, after multiplying additive sum of signal and noise by the cosine (sine) or sine (cosine) components in the corresponding multiplier units, the processing of multiplication results is carried out equally in corresponding bars (devices) - each of the received signals is branched into two identical components, the first component is filtered by the low-pass filter, which is matched with signal band, at the same time the second component is filtered by a band-pass filter, whose passband is chosen so that the upper band-pass filter frequency corresponds to the upper signal frequency, the lower bandpass filter frequency is set as close as possible to zero, the low-pass filter and the band-pass filter are selected with phase-frequency characteristics identical to the maximum and so that the amplitude-frequency characteristic (AFC) of the band-pass filter in the frequency domain close to zero has the highest possible slope, in the frequency domain, starting from the value for where the difference in the amplitude-frequency characteristic (AFC) values of the low-pass filter and the band-pass filters becomes less than a certain predetermined value, it is ensured that their amplitude-frequency characteristic (AFC) is identical to the maximum extent, signals that have passed the low-pass filter and a band-pass filter are subtracted from one another, subtraction result is converted into digital form in corresponding analog-to-digital converters (ADC), from these values corresponding to the sine and cosine components of the same frequency, the spectral density (SP) for each subcarrier is determined by extracting the square root from the sum of their squares and these values proportional to the amplitude of the signals are stored, from the obtained values spectral density with a maximum value is found, the threshold value is determined by multiplying the found maximum value of the spectral density by a coefficient, whose value is set in advance, the obtained spectral density (SP) values are compared with a threshold; according to comparison results, a conclusion is made about presence or absence of a signal with an appropriate frequency. An embodiment of the invention

The signals are formed as the sum of n harmonic signals (subcarriers) with different frequencies using frequency shift keying (FSK). The values of adjacent frequencies differ by a certain value Df_ij. Here, i, j are number of adjacent frequencies, j= i+1.

The frequency shifts values are set so that the difference value of any pair of frequencies does not exceed the frequency value (Fp) (see Fig. 1), for which the difference in the amplitude-frequency characteristic of the low-pass filter and the bandpass filter becomes less than some predetermined value. The reference frequencies are formed with the same values as the harmonic signals.

The number of harmonic signals n used in the signal forming, the frequency shifts values between the signals are determined at the design stage experimentally or by mathematical modeling as values that provide a maximum degree of noise immunity at the given level of data exchange rate.

After the additive sum of signal and noise is multiplied by the sine (cosine) and cosine (sine) components, components of all reference frequencies in multiplier units form results of signal and noise transformation, which processing is carried out identically. Each of the received signals is branched into two identical components, the first component is filtered by a low-pass filter, whose band is matched with the signal band, at the same time, the second component is filtered by a band-pass filter, which passband is selected so that the upper frequency of the band-pass filter corresponds to the upper frequency of the signal, the lower frequency of the band- pass filter is set equal to some predetermined value, which is set as close to zero as possible.

The choice of the low-pass filter and the band-pass filter is carried out with phase-frequency characteristics identical to the maximum extent and so that the amplitude-frequency characteristic of the band-pass filter in the frequency domain close to zero (0 - Fp, see figure 1) has the maximum possible slope, in the frequency domain, starting from the value for where the difference in the amplitude-frequency characteristic (AFC) values of the low-pass filter and the band-pass filters becomes less than a certain predetermined value (Fp - Fc, see figure 1) it is ensured that their amplitude-frequency characteristic (AFC) is identical to the maximum extent.

Signals that have passed the low-pass filter and a band-pass filter are subtracted from one another in each parallel bar (device) corresponding to a quadrature component (multiplier unit).

The received signals are converted to digital form in the corresponding analog-to-digital converters (ADC). From these values corresponding to the sine and cosine components of the same frequency, the spectral density (SP) for each subcarrier is determined by extracting the square root from the sum of their squares and these values proportional to the amplitude of the signals are stored (see, for example, Functional monitoring and diagnostics of electrical systems and devices by digital readings of instantaneous values of current and voltage. / edited by E.I. Goldstein - Tomsk: published by “Pechatnaja manufactura”, 2003, p. 92-94).

From the obtained values spectral density (SP) with a maximum value is found. The threshold value is determined by multiplying the found maximum value of the spectral density (SP) by a coefficient, by a coefficient, whose value is set in advance.

The value of this coefficient is determined at the design stage experimentally or by mathematical modeling as a value that provides the maximum probability of the correct signal detection, provided that the level of false alarm, i.e. making decisions on the presence of a signal in its absence does not exceed a predetermined level.

The obtained spectral density values (SP) are compared with a threshold; according to comparison results, a conclusion is made about presence or absence of a signal with an appropriate frequency.

The simulation of a multi-frequency signal detection using frequency shift keying using quadrature components and compensation of combination components in the presence of interference such as additive white Gaussian noise (AWGN) is done.

The interference in the simulation is presented in the form of additive white Gaussian noise, i.e. sets of harmonic oscillations with random amplitude (U_Pi) and phases values, which are distributed according to normal (amplitude) and

uniform (phase) laws (see for example tutorial “Basis of the radio systems theory. Tutorial. // V.I. Borisov, V.M. Zinchuk, A.E. Limarev, N.P. Mukhin. Edited by V.I. Borisov. Voronezh Scientific Research Institute of Communications, 2004. ”, p. 51)

where: are frequency, phase and amplitude of the i-th noise

component, respectively;

Nsp is number of harmonic interference components used to represent it.

The frequencies of noise components were modeled as random variables whose values are distributed uniformly in the signal band.

Noise samples are independent random variables.

The efficiency evaluation results of the proposed method were obtained by mathematical modeling (simulation) on a computer using the MATLAB system.

In the simulation the following initial data we used:

- number of implementations- 1000;

- number of noise components - 1000;

- number of harmonic signals - 8;

- harmonics frequency values (in arbitrary units): 10,0; 10,1; 10,2; 10,3;

10,4; 10,5; 10,6; 10,7;

- harmonic signal amplitude - 1 ;

- number of samples per period - 2;

- number of periods - 5;

- noise amplitude - 26.0;

- threshold value for signal amplitude - 5.7; - sampling frequency - 1 ;

-the compensation coefficient of the combination components in the frequency domain where the amplitude-frequency characteristic of the band-pass filter is close to the amplitude-frequency characteristic of the low-pass filter - 0.95;

-the compensation coefficient of the combination components in the frequency domain close to zero, where the slope of the amplitude-frequency characteristic of the bandpass filter is maximum, is calculated provided that in this case the amplitude-frequency characteristic of the bandpass filter has a linear dependence.

The decision process simulation results on the signal presence for the proposed method were obtained as follows: for signal / noise power ratio (SNR) of 6.7 and the probability of false alarm 10^-3, the probability of making the right decision about the signal is at least 0.999 for each frequency component of the signal.

In the simulation of the prototype method for OFDM signals following values for harmonic frequencies (in arbitrary units) are set: 1; 2; 3; 4; 5; 6; 7; 8.

According to the simulation results it was found that for the prototype method, probability of a false alarm equal to 10^-3, probability of the right decision making about the presence of a signal equal to 0.999, is provided when the signal / noise power ratio is equal to 1.

Thus, the effectiveness of the proposed method in terms of the signal / noise power ratio exceeds the efficiency of the prototype method by almost 6.7 times. In this case, the signal bandwidth ratio for the examined method (0.7) to the signal band used for the prototype method (7) is 0.1. That is, the receiver sensitivity in which the proposed method is implemented is 10 times higher than the receiver sensitivity in which the prototype method is implemented.

Thus, the efficiency of the proposed method in terms of signal / noise power ratio exceeds the efficiency of the prototype method by almost 67 times.

From the fact that for the implementation of the proposed method it is enough to use two samples per period, and for the prototype method no less than 10 - 15 samples per period and the fact that the value of the upper frequency of the signal spectrum with OFDM significantly exceeds the value of the upper frequency of the signal spectrum used in the proposed method, it follows that the rate of exchange of information when using the proposed method is significantly higher than information exchange rate, which can be achieved using the prototype method.

Brief description of the drawings

A graphical representation of the amplitude-frequency characteristics of a low-pass filter and a band-pass filter is shown in FIG. 1.

The block diagram of the device that implements the proposed method is shown in FIG. 2, where indicated:

1.1 - 1.n - multiplier units from the first to the n-th;

2.1 - 2. n - low-pass filter from the first to the n-th;

3.1 - 3. n - subtract units from first to n-th;

4.1 - 4.n - analog-to-digital converters (ADCs) from first to n-th;

5.1 - 5. n -band-pass filters from first to n-th;

6 - computing unit (CU) (BY).

The device contains n parallel bars (devices), each of which consists of corresponding series-connected multiplier units 1.1-l.n, low-pass filter 2.1-2.n, subtract units 3.1 - 3.n, ADC 4.1 - 4.n, and also band-pass filters 5.1 - 5.n, the inputs of which are connected to the outputs of the corresponding multiplier units 1.1-l.n, the outputs of the band-pass filters 5.1 - 5. n are connected to the second inputs of the corresponding subtract units 3.1 - 3.n. The outputs of the analog-to-digital converter

4.1 - 4.n are connected to the inputs of the computing unit 6 from the first to the n- th, respectively, the output of which is the output of the device. The inputs of the multiplier units 1.1-l.n are combined and are the input of the device. The second inputs of the multiplier units 1.1-l.n are the voltage inputs of the corresponding reference signals. Economic usefulness

The device operates as follows.

The signals are formed as the sum of n harmonic signals (subcarriers) with different frequencies using frequency shift keying. The values of adjacent frequencies differ by a certain value Df_ij.

Here, i, j are number of adjacent frequencies, j= i+1.

The frequency shifts values are set so that the difference value of any pair of frequencies does not exceed the frequency value (Fp) (see Fig. 1), for which the difference in the amplitude-frequency characteristic of the low-pass filter and the bandpass filter becomes less than some predetermined value.

The reference frequencies are formed with the same values as the harmonic signals.

The number of harmonic signals n used in signal forming, the frequency shifts values between signals are determined at the design stage experimentally or by mathematical modeling as values that provide a maximum degree of noise immunity at the given level of data exchange rate.

For processing one subcarrier, two bars (devices) are used. That is, if k subcarriers are used, then the number of bars (devices) is equal to

N=2k.

The adopted additive mixture of signal and interference is fed to the first inputs of the multiplier units 1.1-l.n, to the second inputs of which corresponding reference signals are fed, for example,

The result of multiplying the signal and interference by reference signals branch into two identical components. The first component is filtered by the low- pass filter 2.1 - 2.n, the band of each of which is matched with the signal band. At the same time, the second component is filtered by bandpass filters 5.1 - 5.n, the passband of each of them is selected so that the upper frequency of the bandpass filters 5.1 - 5.n corresponds to the upper frequency of the signal, the lower frequency of the bandpass filter is set as close as possible to zero.

The selection of the low-pass filter 2.1 - 2.n and the band-pass filters 5.1 - 5.n is carried out with phase-frequency characteristics identical to the maximum and so that the amplitude-frequency characteristic of the band-pass filters in the frequency domain close to zero has the maximum possible slope, in the frequency domain, starting from the value for where the difference in the amplitude- frequency characteristic of the low-pass filter 2.1 - 2.n and the band-pass filters 5.1

- 5.n becomes less than a certain predetermined value ((Fp), they ensure that their amplitude-frequency characteristic is identical to the maximum extent (an illustrative example is shown in Fig. 1).

Signals that pass through the low-pass filter 2.1 - 2.n and bandpass filters 5.1

- 5.n, are subtracted from one another, respectively. That is, the signal of the first band-pass filter 5.1 is subtracted from the signal of the first low-pass filter 2.1, the signal of the second band-pass filter 5.2 is subtracted from the signal of the second low-pass filter 2.2, etc.

The received signals are converted to digital form in the corresponding ADC 4.1 - 4.n. These signals are digitally fed to CU 6.

In the computing device 6 from these values corresponding to the sine and cosine components of the same frequency, the spectral density (SP) for each subcarrier by extracting the square root of the sum of their squares is determined and these values proportional to the amplitude of the signals are stored (see, for example, Functional monitoring and diagnostics of electrical systems and devices by digital samples readings of instantaneous values of current and voltage. / edited by E.I. Goldstein - Tomsk: published by “Pechatnaja manufactura”, 2003, p. 92- 94).

The obtained values of sums are compared with a threshold; according to comparison results, a conclusion is made about the presence or absence of a signal with a corresponding frequency.

The simulation results of the multi-frequency signal detection using frequency shift keying and compensation of combination components in the presence of interference such as additive white Gaussian noise (ABGS) are given above.

Multiplier units 1.1 - l.n can be performed, for example, in the form of a mixer (See for example tutorial “Basis of the radio systems theory. Tutorial. // V.I. Borisov, V.M. Zinchuk, A.E. Limarev, N.P. Mukhin. Edited by / Ed. V.I. Borisov. Voronezh Scientific Research Institute of Communications, 2004. ”, pp. 186-189).

ADC 4.1 - 4.n can be performed, for example, on the AD7495BR chip from firm Analog Devices.

Computing device 6 can be performed, for example, in the form of a single microprocessor device with appropriate software, for example, a processor series TMS320VC5416 from firm Texas Instruments, or in the form of a programmable logic integrated circuit (FPGA) with appropriate software, for example, FPGA XCV400 from firm Xilinx.

Thus, the claimed method may be implemented in the device described.

Sources of information

RU 2425457 Cl, 27.07.2011. Device of quadrature reception of frequency- manipulated signals; RU 2548660 Cl, 20.04.2015. Broadband signal receiving device;

RU 2247474 Cl, 27.02.2005. Device for quadrature reception of frequency-keyed signals;

RU 2262802 Cl, 20.10.2005. Device for transmitting and receiving broadband signals, modulated by phase and frequency;

US 5374903 A, 20.12.1994. Generation of wideband linear frequency modulation signals;

US 4462107 A, 24.07.1984. Radio receiver for frequency shift keyed signals.

Claims

Method for signal extraction with frequency shift keying using square components and compensation of combination components, which consists in generating a signal consisting of several harmonic signals using frequency shift keying (FSK), after multiplying by the corresponding reference signals, by sine or cosine component, in multiplier units the result of signal and noise transformation is formed, wherein, frequency shifts between adjacent signals (subcarriers) are set in advance so that the difference value of any pair of frequencies does not exceed the frequency value, for which the difference in the amplitude-frequency characteristic (AFC) of the low-pass filter (LPF) and of the band-pass filter becomes less than a certain predetermined value, the number of N_f frequencies used is set in advance, the additive sum of signal and noise is brached into n components, wherein n = N_f, after multiplying the additive sum of the signal and noise interference by the cosine (sine) or sine (cosine) components in the corresponding multiplier units, multiplication results processing is carried out identically in corresponding bars - each of the received signals is branched into two identical components, the first component is filtered by a low-pass filter, the band of which is matched with the signal band, at the same time, the second component is filtered by a band-pass filter, which passband is selected so that the upper frequency of the band-pass filter corresponds to the upper frequency of the signal, the lower frequency of the band-pass filter is set equal to some predetermined value, which is set as close to zero as possible, the low-pass filter and the band-pass filter are selected with phase-frequency characteristics identical to the maximum and so that the amplitude-frequency characteristic (AFC) of the band-pass filter in the frequency domain close to zero has the highest possible slope, in the frequency domain, starting from the value for where the difference in the amplitude-frequency characteristic (AFC) values of the low-pass filter and the band-pass filters becomes less than a certain predetermined value, it is ensured that their amplitude-frequency characteristic (AFC) is identical to the maximum extent, signals that have passed the low-pass filter and a band-pass filter are subtracted from one another, subtraction result is converted into digital form in corresponding analog-to-digital converters (ADC), from these values corresponding to the sine and cosine components of the same frequency, the spectral density (SP) for each subcarrier is determined by extracting the square root from the sum of their squares and these values proportional to the amplitude of the signals are stored, from the obtained values spectral density with a maximum value is found, the threshold value is determined by multiplying the found maximum value of the spectral density by a coefficient, by a coefficient, whose value is set in advance, the obtained values of the spectral density are compared with a threshold, according to comparison results, a conclusion is made about the presence or absence of a signal with a corresponding frequency.