RU2569554C1 - Protection method against harmonic interference at autocorrelated method for information reception using noise-like signals - Google Patents

Abstract

FIELD: radio engineering, communication.SUBSTANCE: method for protection against harmonic interference at an autocorrelated method of information reception using noise-like signals involves calculation of complex envelopes of the first and second periods of the received signal, calculation by means of Fourier discrete conversion of spectrum functions of these complex envelopes, multiplication of the spectrum function of the first signal cycle by a complex-conjugate spectrum function of the second signal cycle, calculation by means of reverse Fourier discrete conversion of mutual correlation function between these complex envelopes, selection of a maximum component of the mutual correlation function and its comparison with the threshold; with that, squares of envelopes of spectrum functions of the first and second signal cycles are calculated; dispersions of squares of envelopes of spectrum functions of the first and second signal cycles is performed; normalisation of squares of envelopes of spectrum functions of the first and second signal cycles for their corresponding dispersions is performed; in normalised spectrum functions of the first and second signal cycles there performed is searching of maximum components and their positions are determined; values of selected maximum components are compared to the value of the set threshold that is determined according to the allowable probability value of false identification of harmonic interference; in case they exceed the set threshold in spectrum functions of complex envelopes of the first and second periods, the elements located in positions of selected maximum components and their localities are zeroed; with that, localities of positions of selected maximum components are determined by the level of harmonic interference.EFFECT: improving stability of a communication system to harmonic interference at an autocorrelated method of information transmission under conditions of the varying level of received signal.

Description

The invention relates to the field of radio engineering and can be used to build communication systems and devices for synchronizing receivers.

A known method for the autocorrelation reception of noise-like signals, which consists in multiplying the received signal with a reference signal and integrating the resulting product, characterized in that the reference signal is formed by delaying the received signal by a time not exceeding the clock period, the received and reference signals are phase shifted by 90 °, multiply among themselves, integrate the resulting product, the integrated stresses are squared, summarize them, the square root is extracted from the total voltage, about restricts the amplitude of the low-frequency voltage resulting from the top, forming a short negative pulses, used for generating the modulating their function in the forward or reverse code (see. specification of a patent RF №2309550, IPC H04L 27/22, publication 27.10.2007). The disadvantages of this method is its instability to the conditions of multipath propagation of the signal and the effects of harmonic interference.

A known method of transmitting information using noise-like signals, including modulating the carrier wave with a noise-like signal on the transmitting side, transmitting the modulated signal through the communication line, finding the autocorrelation function Y (τ) of the signal on the receiving side and deciding on the value of the transmitted symbol by comparative analysis of the values of Y ( τ), calculated for different τ, wherein as the modulating signal is a periodic noise-like PN sequence, each symbol a _i alphabets ita encode periodic spread spectrum signal with its different from other repetition period T _i, and on the reception side are the values of the autocorrelation function Y (τ) of the input signal when τ delay, characterized in that on the receiving side in addition are values of the autocorrelation function of the received signal at delays t = 2T _i , 3T _i , ... nT _i , then the corresponding values of the autocorrelation function of the input signal are summed

and assign to the received symbol that value a _j for which the signal processing result S (T _i ) turned out to be maximum (see the description of the invention to RF patent No. 2435323, IPC H04L 27/00, publication November 27, 2011). The main disadvantage of this method is the low speed of information transfer and instability to harmonic interference.

Known methods [Lange F. Correlation electronics. L .: Sudpromgiz, 1963; Petrovich I.T., Razmakhnin M.K. Communication systems with noise-like signals. M .: Soviet radio, 1989; Okunev Yu.B., Yakovlev L.A. Broadband communication systems with composite signals. M .: Communication, 1968] transmitting discrete information based on autocorrelation reception, which consist of transmitting two time-shifted delay τ copies of the same noise-like signal s (t). The information parameter in this case can be, for example, a certain quantity whose value determines the delay value τ. As a signal s (t), for example, a signal is selected that is phase-manipulated according to the law of a specially selected code sequence, or a frequency signal, the phases of the harmonic components of which also take values corresponding to any code sequence. As a code sequence, for example, one of the sequences of maximum length or a Gold sequence is selected. At the receiving side, the autocorrelation function of the received signal is calculated. The time position of the maximum of this function is directly related to the delay τ, on which, in turn, the information parameter specified on the transmitting side depends. The main disadvantage of this method of autocorrelation communication is its instability to harmonic interference.

There is a method of searching for a broadband signal when exposed to narrow-band interference, which consists in dividing the samples of the input signal with interference into time-overlapping segments, the duration of which is determined by the response time to the effect of pulsed noise or to changing narrow-band interference, multiply these segments by a window with specified spectral properties, perform fast Fourier transform of the multiplication results, perform the rejection of narrow-band interference, on a time interval equal to the width of the width a band signal, accumulate samples of each frequency, multiply the conversion results by the corresponding samples of the complex conjugate copy of the signal, perform the inverse Fourier transform of the multiplication results, calculate the cross-correlation function modules and search for the results (see the description of the invention to RF patent No. 2331981, IPC Н04В 1 / 10, publication 08/20/2008). The main disadvantage of this method is the instability to changes in the level of the input signal.

The closest method, which is selected as a prototype, is the method [Bobrovsky I.V., Zakharov Yu.V. Time-frequency synchronization in sonar systems with OFDM. Scientific and Technical Collection Hydroacoustics. Vol. 18 (2). OJSC Concern Okeanpribor. - SPb .: Nauka, 2013. - P. 57-65.] Synchronization of receivers using the autocorrelation technique, consisting in the transmission of two copies of the same noise-like signal s (t) with a duration of T. That is, the delay value τ in this case is equal to the duration of the signal T. On the receiving side, the cross-correlation function between the received signal and the reference one is calculated, which uses the same received signal, but delayed by time T. The position in time of the correlation maximum gives information in this case, not about the transmitted data, but about the Doppler distortion of the received signal, which allows you to configure the receiver synchronization circuit. The main disadvantage of this method is its instability to harmonic interference.

The objective of the invention is to ensure the stability of the communication system to harmonic interference with the autocorrelation method of transmitting information in a changing level of the received signal.

The essence of the claimed invention is as follows.

A method for combating harmonic interference with the autocorrelation method of receiving information using noise-like signals, including calculating the complex envelopes of the first and second periods of the received signal, computing the spectral functions of these complex envelopes using the discrete Fourier transform, multiplying the spectral function of the first signal period by the complex conjugate spectral function second period of the signal, calculation using the inverse discrete Fourier transform cross-correlation function between these complex envelopes, selecting the maximum component of the cross-correlation function and comparing it with a threshold, characterized in that the squares of the envelopes of the spectral functions of the first and second signal periods are calculated, the variances of the squares of the envelopes of the spectral functions of the first and second signal periods are calculated, and the squares are normalized envelopes of the spectral functions of the first and second periods of the signal to the variances corresponding to them, in the normalized spectral functions of the first and second the signal iodines search for the maximum components and determine their position, compare the values of the selected maximum components with the value of the set threshold, which is determined in accordance with the allowable value of the probability of false identification of harmonic interference, if they exceed the set threshold in the spectral functions of the complex envelopes of the first and second periods, the elements located at the positions of the selected maximum components and their environs are zeroed, and the neighborhoods of the positions of the selected Maximum Feed component harmonic interference level is determined.

The inventive method is as follows.

, где m - целое число. Требуемый размер окна дискретного преобразования Фурье, необходимый для обработки одного периода принимаемого сигнала, выбирают по формуле:In the received signal s (t), samples of the complex envelope are extracted, for which, after preliminary amplification and analog bandpass filtering, s (t) is sampled with a frequency f _s and subjected to quadrature demodulation. For this purpose, in the corresponding quadrature demodulators, the samples of the received signal s (i / f _s ) are multiplied with the samples of the quadrature local oscillator exp (j2πf ₀ i / f _s ), where f ₀ is the carrier frequency and filtered using low-pass filters with a cutoff frequency F _c = ΔF _c / 2, where ΔF _c is the signal bandwidth. Further, in order to reduce computational costs, the samples of the complex envelope are subjected to the decimation procedure, as a result of which the sampling frequency of this complex envelope is reduced by a factor of m. Then the new value of the sampling rate is given by $$f_s^{\text{'}\text{'}}=f_s/m$$

where m is an integer. The required window size of the discrete Fourier transform needed to process one period of the received signal is selected by the formula:

NFFT = 2 ⁿ¹ ,

- целая часть числа, большая или равная х. Тогда обрабатываемая длительность одного периода сигнала составляет:Where $$n\mathrm{one}=]{\log }_2\left(f_s^{\text{'}\text{'}}\cdot T_c\right)[,]x[$$

is the integer part of a number greater than or equal to x. Then the processed duration of one signal period is:

, $${\dot{U}}_2\left(f\right)$$

, где $$f=\overline{\mathrm{0,}NFFT-1}$$

.After the formation of two periods of the complex envelope containing NFFT samples, the spectral functions of the complex envelopes of these two periods are calculated using the discrete Fourier transform $${\dot{U}}_{\mathrm{one}}\left(f\right)$$

, $${\dot{U}}_2\left(f\right)$$

where $$f=\overline{0NFFT-\mathrm{one}}$$

.

It is assumed that harmonic interference is described by the expression u _n (t) = U _n sin (2πf _n t),

where U _n , f _n - respectively, the amplitude and frequency of the interference.

Then, at the output of the quadrature demodulator, the complex envelope of this interference:

, $$\dot{U}\left(t\right)=U_n\cdot \exp \left(j2\pi f_n^{\text{'}\text{'}}t\right)$$

,

.Where $$f_n^{\text{'}\text{'}}=f_n-f_0$$

.

выражается через частоты настройки фильтров дискретного преобразования Фурье следующим образом:Then the frequency $$f_n^{\text{'}\text{'}}$$

expressed through the tuning frequencies of the discrete Fourier transform filters as follows:

, $$f_n^{\text{'}\text{'}}=\left(s+a\right)\Delta f$$

,

от частоты настройки фильтра дискретного преобразования Фурье, 0<а<1. Тогда спектральная функция комплексной огибающей выражается как:where s is the integer coefficient of the discrete Fourier transform, and is the coefficient determining the frequency deviation $$f_n^{\text{'}\text{'}}$$

on the frequency of tuning the filter of the discrete Fourier transform, 0 <a <1. Then the spectral function of the complex envelope is expressed as:

If a = 0, then:

, а огибающая циклической взаимно-корреляционной функции на выходе автокорреляционного приемника принимает постоянное значение.This case corresponds to the situation when the frequency of harmonic interference coincides with the frequency of tuning the filter of the discrete Fourier transform. In this case, the envelope of the spectral function contains one nonzero component at the frequency $$f=s/T_c^{\text{'}\text{'}}$$

, and the envelope of the cyclic cross-correlation function at the output of the autocorrelation receiver takes a constant value.

.If a = 0.5, then $$S_n\left(2\pi f\right)|{}_{f=p/T_c^{\text{'}\text{'}}}=U_n{T}_c^{\text{'}\text{'}}\frac{\sin \left[\pi \left(p-s-0.5\right)\right]}{\pi \left(p-s-0.5\right)}$$

.

и $$f=\left(s+1\right)/T_c^{\text{'}}$$

огибающая спектральной функции S_n(2πf) будет содержать два одинаковых максимума с уровнями $$2U_n{T}_c^{\text{'}}/\pi$$

, а уровни этой функции в дискретных точках с номерами р>s+1 составляют $$2U_n{T}_c^{\text{'}}/\pi \left(2t-1\right)$$

, где t=p-s, а при p<s - соответственно $$2U_n{T}_c^{\text{'}}/\pi \left(2t+1\right)$$

, где t=|p-s|.In this case, the harmonic interference frequency falls between the tuning frequencies of two adjacent filters of the discrete Fourier transform and the power of this harmonic interference spreads over all filters of the discrete Fourier transform, since the spectral function S _n (2πf) is the sum of all the coefficients of the discrete Fourier transform of the signal, each of which is included in it with a weight equal to the value of the frequency response of this filter at a given frequency. Moreover, at frequencies $$f=s/T_c^{\text{'}\text{'}}$$

and $$f=\left(s+\mathrm{one}\right)/T_c^{\text{'}\text{'}}$$

the envelope of the spectral function S _n (2πf) will contain two identical maxima with levels $$2U_n{T}_c^{\text{'}\text{'}}/\pi$$

, and the levels of this function at discrete points with numbers p> s + 1 are $$2U_n{T}_c^{\text{'}\text{'}}/\pi \left(2t-\mathrm{one}\right)$$

, where t = ps, and for p <s, respectively $$2U_n{T}_c^{\text{'}\text{'}}/\pi \left(2t+\mathrm{one}\right)$$

where t = | ps |.

The passage of the complex envelope of narrow-band noise acting at the input of the autocorrelation receiver through a device that calculates the square of the envelope of the spectral function S _n (2πf) is described based on the following assumptions.

Denote: η is a random process at the output of this device. Then

. Будем полагать, что входной процесс является стационарным гауссовским шумом. Тогда случайные процессы ξ и ζ также будут иметь гауссовские законы распределения плотности вероятности, а случайный процесс η будет являться стационарным процессом в широком смысле с дисперсией $${\sigma }_{\eta }^2$$

, распределенным по экспоненциальному закону. Тогда если произвести нормировку значений этого процесса на величину $${\sigma }_{\eta }^2$$

, то полученный процесс $$\overline{\eta }$$

будет иметь дисперсию $${\sigma }_{\overline{\eta }}^2=1$$

в широком динамическом диапазоне изменения уровня входного сигнала. При этом вероятность превышения случайным процессом $$\overline{\eta }$$

некоторого порогового значения TRF составитwhere ξ _i and ζ _i are the samples of the real and imaginary parts of the output DFT effect, respectively, $$i=\overline{0NFFT-\mathrm{one}}$$

. We assume that the input process is stationary Gaussian noise. Then the random processes ξ and ζ will also have Gaussian laws of the probability density distribution, and the random process η will be a stationary process in the broad sense with dispersion $${\sigma }_{\eta }^2$$

distributed exponentially. Then if we normalize the values of this process by $${\sigma }_{\eta }^2$$

then the resulting process $$\overline{\eta }$$

will have a variance $${\sigma }_{\overline{\eta }}^2=\mathrm{one}$$

in a wide dynamic range of the input signal level. The probability of exceeding by a random process $$\overline{\eta }$$

some threshold value TRF will be

. $$P\left(\overline{\eta }\ge TRF\right)=0.5\cdot \exp \left(-TRF/2\right)$$

.

.From the presented it follows that by the level of the squared ejection of the normalized envelope of the spectral function, it is possible to identify the harmonic noise acting at the input of the autocorrelation receiver, with the probability of false identification, determined by the expression $$P\left(\overline{\eta }\ge TRF\right)=0.5\cdot \exp \left(-TRF/2\right)$$

.

, $${\left|{\dot{U}}_2\left(f\right)\right|}^2$$

и определяют их дисперсии $${\sigma }_{\eta 1}^2$$

и $${\sigma }_{\eta 2}^2$$

. Осуществляют нормировку последовательностей $${\left|{\dot{U}}_1\left(f\right)\right|}^2$$

, $${\left|{\dot{U}}_2\left(f\right)\right|}^2$$

на соответствующие им дисперсии $${\sigma }_{\eta 1}^2$$

и $${\sigma }_{\eta 2}^2$$

. В результате чего формируют последовательности $${\overline{U}}_1\left(f\right)$$

и $${\overline{U}}_2\left(f\right)$$

.Based on this, to combat harmonic interference, the squares of the envelopes of the spectral functions of the first and second signal periods are calculated $${\left|{\dot{U}}_{\mathrm{one}}\left(f\right)\right|}^2$$

, $${\left|{\dot{U}}_2\left(f\right)\right|}^2$$

and determine their variance $${\sigma }_{\eta \mathrm{one}}^2$$

and $${\sigma }_{\eta 2}^2$$

. Perform normalization of sequences $${\left|{\dot{U}}_{\mathrm{one}}\left(f\right)\right|}^2$$

, $${\left|{\dot{U}}_2\left(f\right)\right|}^2$$

to the variances corresponding to them $${\sigma }_{\eta \mathrm{one}}^2$$

and $${\sigma }_{\eta 2}^2$$

. As a result, sequences are formed $${\overline{U}}_{\mathrm{one}}\left(f\right)$$

and $${\overline{U}}_2\left(f\right)$$

.

и $${\overline{U}}_2\left(f\right)$$

отыскивают максимальные компоненты $${\overline{U}}_{m1}$$

и $${\overline{U}}_{m2}$$

, осуществляют сравнение значений максимальных компонент $${\overline{U}}_{m1}$$

и $${\overline{U}}_{m2}$$

с величиной установленного порога TRF, который определяют в соответствии с допустимой величиной вероятности ложной идентификации гармонической помехи, в случае превышения этими компонентами установленного порога в соответствующих последовательностях $${\overline{U}}_1\left(f\right)$$

и $${\overline{U}}_2\left(f\right)$$

определяются позиции N_m1 и N_m2 этих компонент, из последовательностей $${\dot{U}}_1\left(f\right)$$

, $${\dot{U}}_2\left(f\right)$$

формируют последовательности $${\dot{U}}_1^{\text{'}}\left(f\right)$$

, $${\dot{U}}_2^{\text{'}}\left(f\right)$$

, элементы которых на позициях от N_m1-ΔN до N_m1+ΔN обнуляют, где ΔN определяют уровнем гармонической помехи, действующей на входе автокорреляционного приемника, умножают спектральную функцию первого периода сигнала $${\dot{U}}_1^{\text{'}}\left(f\right)$$

на комплексно-сопряженную спектральную функцию второго периода сигнала $${\dot{U}}_2^{\text{'}}\left(f\right)*$$

, с помощью обратного дискретного преобразования Фурье вычисляют взаимно-корреляционную функцию между этими комплексными огибающими и выбирают максимальную компоненту взаимно-корреляционной функции.In sequences $${\overline{U}}_{\mathrm{one}}\left(f\right)$$

and $${\overline{U}}_2\left(f\right)$$

look for maximum components $${\overline{U}}_{m\mathrm{one}}$$

and $${\overline{U}}_{m2}$$

compare the values of the maximum components $${\overline{U}}_{m\mathrm{one}}$$

and $${\overline{U}}_{m2}$$

with the value of the set threshold TRF, which is determined in accordance with the allowable value of the probability of false identification of harmonic interference, if these components exceed the set threshold in the corresponding sequences $${\overline{U}}_{\mathrm{one}}\left(f\right)$$

and $${\overline{U}}_2\left(f\right)$$

the positions N _m1 and N _{m2 of} these components are determined from the sequences $${\dot{U}}_{\mathrm{one}}\left(f\right)$$

, $${\dot{U}}_2\left(f\right)$$

form sequences $${\dot{U}}_{\mathrm{one}}^{\text{'}\text{'}}\left(f\right)$$

, $${\dot{U}}_2^{\text{'}\text{'}}\left(f\right)$$

the elements of which at positions from N _m1 -ΔN to N _m1 + ΔN are zeroed, where ΔN is determined by the level of harmonic interference acting at the input of the autocorrelation receiver, the spectral function of the first signal period is multiplied $${\dot{U}}_{\mathrm{one}}^{\text{'}\text{'}}\left(f\right)$$

on the complex conjugate spectral function of the second signal period $${\dot{U}}_2^{\text{'}\text{'}}\left(f\right)*$$

, using the inverse discrete Fourier transform, the cross-correlation function between these complex envelopes is calculated and the maximum component of the cross-correlation function is selected.

The claimed invention allows to ensure the stability of the communication system to harmonic interference with the autocorrelation method of transmitting information under conditions of a changing level of the received signal.

Claims (1)

A method for combating harmonic interference with the autocorrelation method of receiving information using noise-like signals, including calculating the complex envelopes of the first and second periods of the received signal, computing the spectral functions of these complex envelopes using the discrete Fourier transform, multiplying the spectral function of the first signal period by the complex conjugate spectral function second period of the signal, calculation using the inverse discrete Fourier transform cross-correlation function between these complex envelopes, selecting the maximum component of the cross-correlation function and comparing it with a threshold, characterized in that the squares of the envelopes of the spectral functions of the first and second signal periods are calculated, the variances of the squares of the envelopes of the spectral functions of the first and second signal periods are calculated, and the squares are normalized envelopes of the spectral functions of the first and second periods of the signal to the variances corresponding to them, in the normalized spectral functions of the first and second the signal iodines search for the maximum components and determine their position, compare the values of the selected maximum components with the value of the set threshold, which is determined in accordance with the allowable value of the probability of false identification of harmonic interference, if they exceed the set threshold in the spectral functions of the complex envelopes of the first and second periods, the elements located at the positions of the selected maximum components and their environs, zero, and the neighborhood of the positions of the selected Maximum Feed component harmonic interference level is determined.