RU177832U1 - NARROWBAND SUPPRESSION DEVICE - Google Patents

Abstract

The device relates to the field of electronics and can be used in radio receivers operating in the conditions of narrow-band interference. The technical problem, which is aimed by the invention, is to suppress narrowband interference, regardless of the probability density distribution of the amplitude of the interference, and to eliminate the suppression of the useful signal at the output of the device when exposed to low power interference or complete absence of interference. The solution to this problem is achieved by the fact that the device contains an amplitude detector, the input of which contains a mixture of a useful signal and noise, a unit for obtaining equally probable reports, Q nonlinear elements, 2Q multipliers, Q matched filters, a weight calculation unit and an adder, while the output of the amplitude detector connected to the input of the equally probable reports receiving unit, the output of which is connected to the input of each non-linear element, the output of each non-linear element is connected to the first input of the corresponding a multiplier, the second input of which is connected to the input of the amplitude detector, and the output is connected to the input of the corresponding matched filter, the output of which is connected to the corresponding input of the weighting coefficient calculation unit, and also to the first input of the corresponding multiplier, the second input of which is connected to the corresponding output of the weighting coefficient calculation , and the output is connected to the corresponding input of the common adder, at the output of which we obtain a useful signal with significantly suppressed interference. Thanks to the unit for obtaining equally probable reports, as well as the presentation of the set of amplitude characteristics of nonlinear elements in the form of a Legendre polynomial of order Q, the device is insensitive to the type of interference distribution and its power, which ultimately gives a stabilization of the detection threshold and a stable probability of false alarm. Thus, it is required to calculate once the threshold of correct detection based on the desired value of false alarm, and it will be constant for any changes in the probability density of the envelope of the interference.

Description

A narrowband interference suppression device belongs to the field of radio electronics and can be used in radio receivers operating in conditions of narrowband interference.

Known devices for suppressing narrowband interference presented in [1-3].

The main disadvantage of the considered devices is that they do not allow to achieve maximum interference suppression efficiency, since they are not adaptive with respect to the density distribution of the interference amplitude.

The closest analogue (prototype) in terms of technical nature and principle of operation is the narrowband interference suppression device described in [2, p. 300, Fig. 5.12a]. This device contains a series-connected amplitude detector, a nonlinear element, a first multiplier, the second input of which is connected to the output of the band stop, the input of which is connected to the input of the amplitude detector. The output of the first multiplier is the output of a nonlinear interference suppression device, where we obtain a useful signal with partially suppressed interference.

The technical result of the prototype is to detect signals against the background of narrow-band interference during non-linear processing of the envelope of the input mixture of the useful signal and interference.

The disadvantages of the prototype include:

- the device is designed to work in conditions of a small signal / noise ratio, with a large signal / noise ratio, the suppression efficiency is greatly reduced, i.e. the device implements an interference suppression algorithm that does not take into account the absence of interference at the input of the device. In the event of such a situation, a useful signal will be suppressed instead of interference;

- it is necessary to have a priori information about the interference distribution density, which is most often unknown, or changes over time, i.e. the device is not adaptive.

Thus, the technical problem to be solved by the claimed device is to create a device for suppressing narrowband interference, which is adaptive to the density distribution of the amplitude of the interference, and the effectiveness of which does not depend on the signal-to-noise ratio.

The technical result is achieved by the fact that the device for suppressing narrowband interference contains (see Fig. 1,) an amplitude detector, a unit for obtaining equally probable reports, Q non-linear elements, 2 * Q multipliers, Q matched filters, a weight calculation unit and an adder, while the output the amplitude detector 1 is connected to the input of the unit for obtaining equally probable reports 2, the output of which is connected to the input of each nonlinear element 3, 4, 5, the output of each nonlinear element 3, 4, 5 is connected to the first input of the corresponding variable resident 6, 7, 8, the second input of which is connected to the input of the amplitude detector, and the output is connected to the input of the corresponding matched filter 9, 10, 11, the output of which is connected to the corresponding input of the block for calculating the weight coefficients 15, as well as with the first input of the corresponding multiplier 12 , 13, 14, the second input of which is connected to the corresponding output of the weighting coefficient calculation unit 15, and the output is connected to the corresponding input of the common adder 16.

The parameter Q can take any positive integer values. In FIG. 1 is a structural diagram of a narrowband interference suppression device for the case of Q = 3.

The proposed device selects the envelope of the input mixture of the interference and the useful signal, converts the envelope reports to an equally probable distribution law, divides the resulting process into Q parallel inertialess nonlinear linearly independent channels, exposes the filtering in each channel that is consistent with the envelope of the useful signal, and also summarizes the agreed responses of each channel with a certain weight, and the vector of weighting coefficients of summation is an eigenvector corresponding to the maximum constant value mutual covariance matrix concerted responses.

The proposed device does not use the restriction of a weak useful signal at the input compared to the interference and does not require assumptions regarding the probabilistic characteristics of the interference. The only limitation is the presence of consistent filtering. Due to the well-known fact - its invariance with respect to the amplitude, phase and time of arrival of the useful signal at the input - follows a similar invariance of the proposed device.

Thus, the technical result of the operation of the device is the absence of the need for a priori information about the density of the interference distribution, as well as the absence of a restriction on the useful input signal that is weak compared with the interference.

The amplitude detector 1 can be performed on semiconductor diodes, as indicated in [4, p. 123, Fig. 7.1], or using operational amplifiers [5, p. 109, Fig. 5.10].

The remaining units of the device can be performed on a microprocessor or signal processor that has a multi-channel ADC, for example, MultiClet R1 (see the instruction manual) or TMS320F28335 (see the instruction manual) using a separate multi-channel ADC, for example, AD7779 (see manual).

The unit for obtaining equally probable reports 2 can be implemented, for example, using the ranking method, which means replacing each report received at its input with its own rank, i.e. the number of the place in the variation row. Moreover, since the Legendre polynomial of order Q, the argument of which lies in the range of values from -1 to 1, must be used as a system of amplitude characteristics of nonlinear elements, the values of the reports received at the input of nonlinear elements, i.e. ranks should also be reduced to this range. Thus, conditions are created for the use of non-linear elements of Legendre polynomials as basis functions of nonlinear elements.

The block for determining the weighting coefficients is implemented by solving the classical problem of finding the eigenvector, which corresponds to the maximum eigenvalue of the matrix of mutual covariance of the responses of the matched filters [6, p. 82].

The operation of the device can be described as follows.

Any narrow-band process y (t) can be represented as:

Where

- комплексная огибающая процесса y(t), которая содержит огибающую A(t) и фазу ϕ(t).

is the complex envelope of the process y (t), which contains the envelope A (t) and the phase ϕ (t).

Any inertialess nonlinear transformation f [y (t)] changes only the envelope A (t) of the process y (t):

Where

u [A (t)] is the amplitude characteristic of the nonlinear converter according to the first harmonic of the frequency ω ₀ .

We represent the amplitude characteristic of a nonlinear transducer in the first harmonic by a generalized polynomial

Where

- вектор столбец параметров настройки нелинейного преобразователя,

- система линейно-независимых базисных функций.

- a vector column of non-linear converter settings,

- a system of linearly independent basis functions.

Then

Replace the continuous functions of time with the set of their discrete reports at time instants t _i = iΔt, then y (t _i ) = y _i , A (t _i ) = A _i , ϕ (t _i ) = ϕ _i where i∈ [1 ... N ] is an integer, the number of signal reports.

The reports of the processes at the outputs of the basis functions can be combined in a rectangular N × Q matrix D with elements

, то совокупность отчетов на выходе нелинейного преобразователя

можно представить в виде f=Dh, а энергию на его выходе, равную энергии на входе СФIf the matrix D with elements d _{i, k} is represented by a set of vectors

, then the set of reports at the output of the nonlinear converter

can be represented in the form f = Dh, and the energy at its output, equal to the energy at the input of the SF

We assume that the matched filter has a complex impulse response g (t _j ) = g _i , where j∈ [1 ... M] is an integer, the number of impulse response reports.

Due to the linearity of u [A (t)] with respect to the settings h, the response of the narrowband interference suppression device z (t _i ) = z _i can be represented by the sum of the responses received from each basis function ν _k (A _i )

We denote the set of reports at the outputs of the matched filters by a rectangular matrix R with elements

на выходе устройства подавления узкополосных помех может быть представлена в виде Z=Rh, а энергию на его выходе, т.е. на выходе СФThen the totality of reports

at the output of the narrowband interference suppression device can be represented as Z = Rh, and the energy at its output, i.e. at the output of the SF

The device must adapt to the received interference in such a way as to ensure the maximum ratio of the energy at the output of the matched filter to the energy at its input, if you imagine the device as a simple equivalent of the form a nonlinear converter plus a matched filter. Then the energy at the input of the matched filter is the energy at the output of the weighted sum of nonlinear converters, the amplitude characteristics of which form a system of orthonormal functions. Thus, the efficiency indicator K _E settings of the nonlinear converter is equal to the ratio of the energy at the output E _o and input E _{I of the} matched filter, which are determined by the expressions:

The optimal values of the settings of the nonlinear converter are considered to be those h _onm , at which the maximum of this indicator is achieved.

,

, тогда показатель эффективности примет вид обобщенного отношения Релея:We denote

,

, then the performance indicator will take the form of a generalized Rayleigh relationship:

Its maximization is equivalent to finding the maximum eigenvalue λ _max and the corresponding eigenvector h _opt , which are a solution to the equation

In particular, if the basis functions ν (A) are orthonormal, namely, they are a Legendre polynomial, then the matrix B becomes the identity and the search for optimal parameters h _opt reduces to solving the classical eigenvalue problem.

Thus, the use of the unit for obtaining equally probable reports and the use of the Legendre polynomial as basic functions makes the device insensitive to the type of interference distribution and its power, which ultimately gives a stabilization of the detection threshold and a stable probability of false alarm. Therefore, it is required to calculate once the threshold of correct detection based on the desired value of the false alarm, and it will be constant with any changes in the probability density of the envelope of the interference.

To evaluate the effectiveness of the narrowband interference suppression device, mathematical modeling of its operation was carried out under the following types of interference: harmonic and from an excited sea surface.

Table 1 shows the gain μ for the device for suppressing narrowband interference and the maximum possible gain μ _max (the values in the table are given in brackets) for the conditions of harmonic interference [2, p. 301, Fig. 5.13], with a different “noise / noise” ratio α (basis functions are the Legendre polynomial of order m).

Table 2 shows the gain μ for the device for suppressing narrow-band interference and the maximum possible gain μ _max (the values in the table are given in brackets) calculated according to the procedure [7] for marine noise conditions with a K-distribution of probability density [8, p. 113 , formula 4.33], with a different “noise / noise” ratio α (two variants of the interference parameters: ν = 0.5 and b = 0.5, ν = l and b = 0.5, the basis functions are the Legendre polynomial of order m).

According to tables 1 and 2, we can conclude that the effectiveness of the proposed device for suppressing narrow-band interference is basically insignificant (1-2 dB) differs from the maximum possible efficiency.

LITERATURE

1. RF patent No. 2360360 C1, IPC Н04В 1/00, publ. 06/27/2009.

2. P.A. Bakut. Theory of signal detection. M .: Radio and communications, 1984.440 s. Page 300, fig. 5.12a.

3. RF patent No. 2352063 C1, IPC Н04В 1/10, publ. 04/10/2009.

4. Radio receivers / Ed. A.P. Zhukovsky. M .: Higher school, 1989.342 s.

5. V.I. Shcherbakov, G.I. Grezdov. Electronic circuits on operational amplifiers: Reference. K .: Technique, 1983. 213 p.

6. F.R. Gantmakher. Matrix theory. M .: Nauka, 1966.576 s.

7. V.G. Valeev, A.A. Yazovsky. Adaptive nonlinear converters for suppressing non-Gaussian interference // Kiev Polytechnic Institute. Izv. Russian universities. Radio Electronics 1987. No. 8. Volume 30.S. 62-64.

8. K. Ward, R. Tough, S. Watts. Sea clutter: scattering, the K distribution and radar performance. Croydon: CPI Group Ltd, 2006.452 p.

Claims (1)

A narrow-band interference suppression device containing an amplitude detector, characterized in that the device is equipped with a unit for obtaining equally probable samples, Q non-linear elements whose amplitude characteristics form a system of linearly independent orthonormal functions - Legendre polynomials of order 0, 1, ..., Q-1, 2 * Q multipliers, Q matched filters, a weight calculation unit and an adder, while the output of the amplitude detector is connected to the input of the equally probable reports unit, the output of which is connected to by the course of each non-linear element, the output of each non-linear element is connected to the first input of the corresponding multiplier, the second input of which is connected to the input of the amplitude detector, and the output is connected to the input of the corresponding matched filter, the output of which is connected to the corresponding input of the weight coefficient calculation unit, as well as to the first input the corresponding multiplier, the second input of which is connected to the corresponding output of the unit for calculating the weight coefficients, and the output is connected to the corresponding the general progress of the adder.

Images