RU2759216C1 - Method for telecommunications for achieving potential productivity, interference resistance and speed of connection - Google Patents

Abstract

FIELD: telecommunications.

SUBSTANCE: to realise the described result, a method for telecommunications is proposed for achieving potential productivity, interference resistance, and speed of connection, consisting in the fact that in case of errors in analogue communication channels the information is transmitted in portions. Located in front of the portion is the header - pilot signal - followed by the body of the portion. The body of the portion is formed by a first convolution, wherein reduced copies of the pilot signal are used as said convolution. The formed portion is then read into the communication channel at the maximum permissible speed. At the output of the communication channel, the following tasks are performed using a synchroniser on AS-transformation: a portion is detected and divided into a header and a body; a first scan of the header and the body of the portion is performed, returning the inaccurate information with errors; a second scan is performed, returning distortion and interference; the result of the second scan is subtracted from the result of the first scan.

EFFECT: increase in the productivity, interference resistance, and speed of connection.

1 cl, 5 dwg, 2 tbl

Description

The claimed METHOD relates to the field of telecommunications, EXCEEDS the state of the art in this field, has no analogues and is BASED on the previously proposed AS-conversion of signals (see patent SU 684747).

I. INTRODUCTION

The claimed METHOD is an algorithm for the operation of a new DEVICE, replacing modems in telecommunications in order to CARDINALLY improve their characteristics - "compliance" with the Shannon-Hartley formula (limit).

(This algorithm is applied in the computer program of the same name, registered by the Federal Service for Intellectual Property No. 2019666319.)

The specific (per 1 Hz frequency band of the communication channel) Shannon-Hartley formula determines the potential performance of telecommunications as a function of one argument - additive Gaussian noise (provided that subscribers are instantly connected).

Obviously, modems could "correspond" to the discussed formula if they eliminated ALL other distortions and interference (of a certain intensity) in telecommunication channels under the above condition.

But, modems, as will be shown in chapters II and III of the DESCRIPTION, in principle cannot do this. However, the possibility of telecommunications "correspondence" to the specific Shannon-Hartley formula is OPENED in THIS APPLICATION.

FIG. 1 (see. Graphic images) shows a block diagram of modern telecommunications, where the DEVICE is marked in red instead of a modem, operating according to the proposed algorithm (method), which solves the following tasks: IMPLEMENTATION in analog channels ELIMINATION by invariant methods of all errors from the action of multiplicative linear and_nonlinear distortions, as well as errors from the action of additive INTERFERENCE (without Gaussian noise) and errors in binary communication channels, regardless of their origin. We will CALL the procedure for solving the LISTED problems - the RUIP method.

In the RUIP method, a message (information) is transmitted over telecommunication channels in Portions of multivalued sequences (a signal "constellation"). Ahead of the portion is the HEADER (pilot signal) and then the BODY of the portion containing the message being transmitted.

On the transmission of the BODY of the portion, the 1st CONVOLUTION of the message is formed with complex carrier signals with an effective duration of T and a base D >> l, which are mixed copies of the pilot signal (HEADER). Further, through the DAC (or code converter), the PORTION formed in this way enters the communication channel, where its 2nd CONVERSION with the channel impulse response (IHC) takes place - channel distortions are introduced into the PORTION.

At the reception after the ADC (or code converter) using a high-speed synchronizer based on AS-conversion [3], the following is performed: 1) DURING THE 1ST PERIOD (instantaneously) channel PORTION; 2) SHE is divided into HEADER and BODY. Next, the 1st HEADER and BODY SCAN is performed, returning invalid information with errors from the actions of the above distortions and interference, including errors from Gaussian noise.

NOTE: we point out that all CONVOLUTIONS (CONVOLUTIONS) and DEVELOPMENTS (DEVELOPMENTS) in the RUIP method are not classic convolutions and their inversions.

Along with the 1st SCAN (by analogy with the "classic" convolution inversion), there should also be a 2nd SCAN of unreliable information and a HEADER of a portion (acting here in the role of IHC) at the reception, which returns distortions of the communication channel in the form of intersignal interference, which determines all errors listed ABOVE.

We managed to find this 2nd SCAN (nonlinear operation) and further, subtracting in a certain way (shown in detail in Chapters IV and V) the interference obtained from the result of the 1st SCAN, we obtain quasi-reliable (almost reliable) information - all errors are corrected, except errors from the action of Gaussian noise.

FIG. 2 (see. Graphic images) shows the final product of the proposed METHOD in the form of RED graphs, obtained by modeling it in the MatLab environment under the influence of all distortions and interference. At point "A" (Fig. 1), only errors from Gaussian noise (Psh) are possible. then. in telecommunications, POTENTIAL performance, noise immunity and connection speed are ACHIEVED under the condition of an acceptable probability of error from the action of noise Pm.

The green graph in FIG. 2 shows the specific performance of modems [2] with all the same distortions and interference for the probability of error from the action of noise Psh 1.0e-007.

WE WILL PROVE ALL OF THE ABOVE ...

II. FOREWORD LINEAR SECTION

Igor Baklanov writes in a recently published book [1]: “Since the perestroika period, domestic communication systems are equipped with imported equipment, and in most cases, on a turnkey basis. Figuratively speaking, Russian communications are firmly "put on the needle" of imported technical solutions, which are implemented without deep analytical research. As a result, Russian operators are simply unable to make any qualitative changes in technologies. "

If one can agree with the last phrase, adding that these “qualitative changes” that are currently inaccessible have been lost, first of all, for the country's defense, one cannot entirely agree with the first one. Suffice it to recall the time of the beginning of our "march in the wake" of American IBM technologies, which ended with the ES computers, ASVT and others sunk into oblivion. With what persistence our modem developers began to follow the CCITT (now ITU-T) recommendations in the 70s of the last century by adopting "recommendations" as law for immediate implementation. Bottom line: now there are no domestic developers or domestic telecommunication systems.

It is known that practically all the recommendations in the field of telecommunications (telecommunications) were proposed by the United States, and we are now seeing a paradox: along with all those who are hooked on the “American needle”, the Americans themselves turned out to be: they can no longer make “sudden movements” and “dig a hole” deeper and deeper. and deeper.

Are there they "dig" about our characteristics?

The above question was relevant in the early 70s of the last century at two departments of the former Leningrad Institute of Communications named after prof. M.A. Bonch-Bruevich (now St. Petersburg State University of Telecommunications) - at the Department of TPS, where outstanding scientists, Professor A.M. Zaezdny (head of the department), professor L.M. Fink and studied at the graduate school of the Department of Thermal Power Plant the author of these lines. The first department carried out scientific developments based on analytical studies of local characteristics, the second - implemented the recommendations of the CCITT. We were fortunate enough to join the scientific school of the TPN department and offer (Application No. 2004875/09) a new data transmission solution adequate to domestic communication channels. The then leadership of science at the institute supported our BEGINNING.

Soon, changes took place at the institute: Professor M.A. Zaezdny, Professor L.M. Fink, the leadership of science has changed. BEGINNING closed securely. Needle took over.

Analysis of the current state of the physical level of telecommunications [2] showed that the above-mentioned BEGINNING not only did not lose its relevance in 50 years, but also increased it. What are the reasons for this effect:

1) we "dug a hole" in another place;

2) currently, a big step forward has been made in telecommunications, which has significantly increased their quality - the transition to the transfer of information in digital form; we propose to improve the achieved quality by the second step - to go from simple signals with base ≅1 to complex signals with base D >> 1;

3) from the created theories, know-how and their modeling in the MatLab environment, we made sure that our works contain the best (potential) results in the field of telecommunications (telecommunications);

4) we managed to significantly develop our ideas and invest them in the WAY OF ACHIEVING POTENTIAL TELECOMMUNICATIONS. PRODUCTIVITY. IMMUNITY and CONNECTION SPEED (RUIP method) and extend them to all modern communication channels;

5) we have proposed a new signal transformation called AS-transformation [3], which made it possible to create synchronization for instant connection of subscribers, which provides increased efficiency and secrecy of information transfer, and in addition, this invention served as the basis for the considered innovation as a whole (see below );

6) the practical implementation of this method mainly consists of convolution operations, the potential (best) algorithm of which was also proposed by us [4];

7) on the basis of time-frequency duality [5], we have developed a synthesis of signals and filters that is adequate to this method.

From the above and given that this innovation is built on a different platform and has no analogues, it is appropriate to assume that it will be in demand in NEW GENERATION telecommunications with its most effective application in wire, radio, space, hydroacoustic communication channels and television.

So, the RUIP method is proposed, which provides TELECOMMUNICATIONS with the elimination of all errors by invariant methods that are insensitive:

1) to multiplicative (modulating) distortions such as linear distortions, multipath and echo interference, fading associated with non-stationarity of the channel, Doppler effect, frequency divergence in the channel, to intrinsic nonlinear distortions associated with inaccurate implementation of algorithms (for example, arising when replacement of infinite limits of integration by finite ones);

2) to additive concentrated disturbances (impulse and harmonic);

3) to errors in binary channels, regardless of their origin.

Let us consider the LINEAR part of the RUIP method with respect to modulating interference. We represent the signal (information carrier) u (t) at the input of the communication channel by a pair of Fourier transforms:

Then the signal at the output of the channel, distorted by modulating distortions, can be expressed as follows:

- амплитудно-временные искажения;where

- amplitude-time distortion;

фазово-временные искажения;

phase-time distortion;

амплитудно-частотные искажения;

amplitude-frequency distortion;

фазово-частотные искажения.

phase-frequency distortion.

Functions

completely characterize the parasitic signal modulation caused by distortion (modulating noise). These functions are called interference modulation functions.

If in the communication channel, in addition to the modulating interference, there is a normal additive interference (for example, Gaussian noise)

where n '(f) is the spectrum of realization n (t), then the algorithms of ideal receivers with input signals X (t) or Y (f) and equiprobable signal elements with known ξ1 (t) or ξ2 (f) have the form:

- correlation technique;

- matched filtering,

where T and F are respectively the effective duration and width of the signal spectrum;

- принимаемый сигнал;

- received signal;

- спектр принимаемого сигнала;

- the spectrum of the received signal;

{Ar * (t) ⋅ξl * (t)} - set of reference oscillations;

{Br * (f) ⋅ ξ2 * (f)} - a set of matched filters;

Li, k is the alphabet of signals, and the sign (*) denotes complex conjugacy.

At present, on the basis of algorithm (6), using adaptive reception methods, high-speed information transmission systems (modems) have been created, which to one degree or another approach the ideal in conditions of interference ξ2. At the same time, according to publications, the possibility of practical implementation of the digital IDEAL telecommunications using the algorithm (5) is not known.

The possibility of implementing algorithm (5) can be approached by assuming the function of noise modulation ξ1 (t) is completely unknown and synthesizing a telecommunication system based on the theory of invariant noise immunity [6].

In what follows, we will use the normalized analytical signal in the form of its complex envelope U (t) and its spectrum G (f), i.e. such that

Let the synthesized telecommunication system is intended to operate on channels of width Fk, the transfer functions of which satisfy the expression:

where Тск is the stationarity interval of the communication channel.

Let us introduce a significant limitation, which will be removed further in the nonlinear part of the considered RIE method at τ≤Тcк:

соответственно и на приеме неизвестны, то количественная характеристика помехоустойчивости Р системы электросвязи является функцией обеих помех.If random realizations n (t) and ξ1 (t) belong to two sets of noise N and

respectively, and at the reception are unknown, then the quantitative characteristic of the noise immunity P of the telecommunication system is a function of both interference.

выполняются равенства:An absolutely invariant telecommunication system with respect to interference ξ1 is a system in which for all ξ1 ∈

the equalities hold:

The error probability when performing (10) can be represented as:

where Pdop is the permissible error probability at the specific transmission rate Cy not less than Pdop.

We will use the synthesis of invariant systems, called the invariant operator method:

a) find a linear transformation (operator) Φinvar ξ1 satisfying condition (10), that is, such that for all ξ1 the identity

the found operator is comparable in noise immunity with the ideal one (5), as a result of which we will find out the possibility of satisfying condition (11);

c) we describe the found operator in the form of an algorithm for a real telecommunication system;

d) choose such a waveform U (t) to minimize the probability of error Р

The required invariant operator Φinvarξ; obtained by us on the basis of Parseval's equality and under constraint (9) has the form:

(t)=U(t)⋅ξ(t) - искаженный помехой ξ1 сигнал U(t) на приеме;where

(t) = U (t) ⋅ξ (t) is the signal U (t) distorted by the interference ξ1 at the reception;

- амплитудный спектр неискаженного сигнала U(t);

- amplitude spectrum of the undistorted signal U (t);

to - elementary interval.

Comparing the first part of (14) with (5), it is easy to find their coincidence if in (14) we introduce an additive noise n (t) and replace the infinite limits of integration with finite ones.

In its pure form, operator (14) is not physically realizable due to the need for integration over an infinite time interval. With a finite integration interval T, the telecommunication system based on (14) will be ideal for linear systems, where the normalized energy of intersignal (intersymbol) interference is allowed due to a finite T (see below).

The considered RUIP method can be used both for digital-frequency and digital-time division of channels (subscribers). In addition, he has access to subscriber division - digital-frequency-time, which is not considered here.

In all cases of separation, information is transmitted by quadrature multi-valued sequences Li⋅Uk (t) and Lj⋅Yk (t) [7], which form a signal "constellation" on the Uk, Yk plane. In what follows, to simplify the presentation, one of the quadratures, Li⋅Uk (t), is considered, which corresponds to operation in the LF channel.

On the transmitting side, with digital-partial division, the 1st convolution of information Li with the pilot-signal Lo⋅U is performed and the result in the notation [8] will be the signal:

where {Li, k} = {Li}; i = 1,2,3, ..., A is the multiplicity of the convolution of signals U (t-k⋅to);

=Е - допустимая энергия на одно измерение;

= E - permissible energy per measurement;

D = 2⋅F⋅T - number of measurements during T (signal base);

to = l / (2⋅F) - time shift between adjacent signals U (t);

а »1 - coefficient depending on the message length.

Taking the functions U (t-k⋅to) orthonormal, i.e.

we have:

where δk, r is the Kronecker symbol.

The last equality means that the square of the amplitude spectrum of the signal U (t) in (17) must satisfy the Nyquist condition:

where F = 1 / (2⋅to) is the spectrum width | G (f) | at the 0.5 level.

When (18) is fulfilled, the functions U (t-k⋅to) will have the property:

- допустимая нормируемая энергия межсигнальных помех из-за конечного Т.where

is the permissible normalized energy of intersignal interference due to the final T.

From the above algorithm of the transmitter it follows: information is transmitted over the communication channel in Portions (discretes) of analog multi-valued quadrature sequences (signal "constellation"), forming the body of the portion (we will call these sequences - aBytes).

Such data transfer is adequate to the discrete form of any information that has developed in nature at any level of its presentation and transformation:

- in a conversation: discrete - a thought, a sound word, an alphabetic code of a word;

- in television: discrete - line, frame;

- in data transmission [9]: discrete session, block, fragment, packet, frame;

- in facsimile communication, discrete line, page, etc.

Ahead of the discrete is the header of the portion (space, pause, address, etc.), which serves to identify (recognize) this discrete. In our case, a header (pilot signal) with a duration of T:

providing synchronism of the entire system, instant connection of subscribers, coherent reception of signals, elimination of multiplicative distortions and additive concentrated interference in them in the entire information transmission path. Let us determine the duration of the portion Tp:

where Nb is the number of aBytes per chunk.

Obviously, in order to neglect the loss of system performance due to the presence of a pilot signal in a portion, the following condition must be met:

T << to⋅Nb.

Thus, for example ;, a seven-level block diagram of packet data transmission in IP networks, when using the considered method in them, will take a complete form:

Further, formed in this way, the portion through the DAC enters the communication channel, where the 2nd convolution takes place with the IHC, introducing distortions of the transmission path into the portion.

Let's move on to considering a system with digital-time division of channels. From the principle of time-frequency duality we have: if during frequency division one subscriber receives Na aByte of information during the duration of a portion Тn, then with time division Na subscribers in time Тn receive at least 1 aByte of information. In this case, the signal for transmitting "m" aBytes of information to the k-th addressee in the notation (15) will be as follows:

where Na is the number of channels (subscribers) in a portion;

k - address of the recipient of information (number of aByte, counted from the sync pulse).

With digital-time division, the communication of all subscribers with each other can occur autonomously without any operators, and in some cases - in the same frequency range.

One of the distinctive features of the considered innovation is the use of the SAME RECEPTION ALGORITHM for systems (networks) as with digital-frequency. and with digital-time division of subscribers. This is due to the peculiarity of the digital-time division. If the transmitting side knows the address of the receiver (receiver), then the latter has its own aByte of information and the address of the transmitter, only by demodulating all the previous aBytes. In addition, this operation is needed to prevent contention of subscribers, i.e. operation of several transmitters on one receiver.

And so, in both types of separation in the receiver, from the oscillations at its input after the input filter and the ADC, a PORTION

pilot tone - portion HEADER

and BODY portions - information signal

Next, the 1st sweep of the distorted pilot signal (24) and the information signal (25) is performed, the result of which will be unreliable information Lr, containing errors from additive Gaussian interference and admissible modulating distortions ξ1:

additive noise (normal random variable with zero mean and variance

where No is the spectral density of the additive fluctuation noise n (t)).

Expression (28) shows that coherent reception of information regarding additive noise interference is provided here.

From expressions (19) and (26) it can be seen: the invariance of reception to modulating (multiplicative) distortions (excluding admissible) will take place if

=σдоп - допустимое среднеквадратичное отклонение интеграла в (29) от единицы.where

= σadd is the permissible root-mean-square deviation of the integral in (29) from unity.

The fulfillment of (29) is ensured by the choice of the function U (t) and the time T. Assuming for a sufficiently long integration time T:

where ε is a normal random variable with zero mean and variance σ ⋅ Е ² , we require a minimum of the probability of an error of element-by-element reception:

where F (a) is the Laplace function;

Rr - distance between adjacent letters

- суммарная дисперсия помех.

is the total noise variance.

For a given Rr, requirement (31) can be satisfied under two conditions:

These conditions are contradictory, since for an unknown function of interference modulation ξ1 (t) to satisfy (32), the effective duration of the signal U (t) should not exceed T. On the other hand, with a limited peak power of the communication channel for the best satisfaction (33), it is necessary to have U (t) with a uniform envelope over the entire interval T.

A compromise solution can be found if we turn to the class of complex signals with a base D >> 1. For a signal U (t) with a large base D, a pair of Fourier transforms of which is

the differential equation [10] is valid:

where t = -dψ (f) / df = T (f) is the variance of the signal U (t);

f = dϕ (t) / dt = F (t) is its instantaneous frequency.

For single-valued t = T (f) and f = F (t), the following equality holds:

Taking into account constraint (9), we represent the signal at the output of the communication channel as follows:

(t) в (36) с учетом (34), (35) и (37) запишется следующим образом:Differential equation for

(t) in (36) taking into account (34), (35) and (37) is written as follows:

dT (f) / df - uneven dispersion of the signal U (t);

(f)/df - приращение неравномерности дисперсии сигнала

(t) за счет модулирующих помех.

(f) / df - increment of signal variance unevenness

(t) due to modulating interference.

If you choose

в (38) будет мало отличаться от единицы и можно принятьthen fraction

in (38) will differ little from unity and one can take

which is equivalent to satisfying condition 1 (32).

When inequality (39) is fulfilled, the increase in the unevenness of the signal dispersion U (t), taking into account the constraint (9), fully characterizes the influence of modulating interference and is a characteristic of the communication channel - the specific unevenness of its group deceleration time (GDT):

- фазово-частотная характеристика (ФЧХ) канала связи;where

- phase-frequency characteristic (PFC) of the communication channel;

- его ГВЗ.

- his GVZ.

The specific non-uniformity of the GDT of the communication channel is approximately determined as [11]

where Тк is the channel memory, and inequality (41) can be represented as:

(f) наиболее разумен выборIt is easy to show that for an unknown form

(f) most reasonable choice

where T is the effective duration of the signal U (t).

In this case, equation (38), taking into account (40), will be as follows:

From (45) it follows:

and

which satisfies criterion 2 (33).

Expressions (44), (46) and (47) determine the optimal form of the complex signal U (t) according to criterion 1 (32). For example, a chirp signal with a base D >> 1 possesses such a shape.

III. CONCLUSIONS FROM THE FOREWORD

1). The above in section II was simulated by the MatLab environment [12, 13, 14], as a result of which it was possible to achieve invariance to distortion and interference ONLY for 4-fold KAM-4 sequences, due to the forced replacement of infinite integration limits by finite, as well as the simultaneous limitation of the duration of the signals and the frequency band occupied by them in the channel. These actions in LINEAR systems, which include modern adaptive modems [2], causes FATAL distortions, limiting the performance capabilities of such systems (see Fig. 2, Graphic images, green graph).

2) The possibility of transmitting information over telecommunication channels WITHOUT ERRORS from ALL multiplicative distortions and additive noise (except for Gaussian noise) at ANY (within "reasonable" limits) MULTIPLICITIES of multivalued sequences (see Fig. 2, Graphic images, RED graphs), implemented in NONLINEAR part of RUIP mode. It is described in detail in Sections IV and V of the ALGORITHM of the method in mathematical language and in MatLab (with explanations in Russian), where assumption (9) and all other assumptions and approximations made in the LINEAR part are removed.

3) Section IV provides a SIMPLIFIED BLOCK DIAGRAM OF THE METHOD ALGORITHM, where, in addition to a brief summary of Chapter II, the NONLINEAR part of the RUIP method is described in detail - the 2nd Sweep, which returns to the receiver QUASIDOFFICIENT information (transmitted information that may contain errors ONLY from Gaussian noise ). It also shows the EFFICIENCY of the 2nd sweep, where, in addition to eliminating errors from multiplicative distortions and additive noise (without Gaussian noise), the intrinsic distortions in the entire transmission path obtained from replacing the integration limits from infinite to finite ones, simultaneous limitation of the signal and its spectrum, etc. ...

The concept of "reasonable" (maximum allowable) INTENSITY (power) of distortion and interference is introduced and the limit of "reasonableness" is defined, when exceeded, an emergency signal is generated - the current portion becomes lost due to a malfunction of the communication channel.

IV. SIMPLIFIED BLOCK SCHEME OF NONLINEAR METHOD ALGORITHM.

Let us recall that in the RUIP method, information is transmitted over communication channels by a multi-valued quadrature sequence of bytes (a signal "constellation").

FIG. 3 (see. Graphical images) shows a simplified block diagram of the algorithm of the RIPM method, where vectors of digital arrays are highlighted in purple.

At the first input of the 1st CONVERSION (1CB) we have the transmitted information IUY in decimal code (the encoder is not shown in Fig. 3), the second input of which receives orthogonal analog carrier waves U and Y obtained from the working signal Uc. As a result, we have the BODY of the portion S, to which is attached the HEADER of the portion (AU, AU) a pilot signal consisting of two increased signals Uc, and we obtain the PORTION SK = [Uo, AU, AU, S], where Uo is an empty space.

The generated PORTION SK through a DAC or a code converter (not shown in Fig. 3) are sent: 1) OR to an analog communication channel with a maximum speed determined by the highest frequency of the frequency response (AFC) of this channel, and the indicated AFC should be matched taking into account Doppler effect with the amplitude spectrum of the Uc signal; 2) OR into a binary channel at a speed determined by it.

In the communication channels, the 2nd CONVERSION (2CB) of a portion of SK with IHK occurs, at the output of which we have a distorted portion of SKk. On the receiving side (ADC or code converter in Fig. 3 are not shown) a portion of SКk in the synchronizer (-) by AS-conversion (not shown in Fig. 3) for the TIME of the 1st PERIOD of the ADC clock frequency or for the TIME of the 1st PERIOD of the clock frequency set by the binary channel is divided into distorted BODY portions (Sk) and HEADER portions (AUk, AYk), which are fed to the 1st SCAN (1PB). The result is invalid information (X).

(We indicate here the requirements for the relative stability of the CLOCK FREQUENCIES of the above-mentioned DACs and ADCs - the difference between their phases during the transfer of PORTION should not exceed 2-3 degrees.)

Further, to simplify the presentation, consider one of the quadrature subchannels (IUk). Omitting the constant, we have at the output of the 1st SCAN (1PB):

- массив (вектор) искаженных отчетов сигналов Ukr(t) в максимумах их автокорреляционных функций;

- an array (vector) of distorted reports of signals Ukr (t) at the maxima of their autocorrelation functions;

- искаженный пилот-сигнал AUk

- distorted pilot signal AUk

The work of the 2nd SCAN (2RB), as a result of which ALL multiplicative distortions and additive noise in the form of intersignal interference arrays are returned, is performed in TWO stages:

FIRST STEP:

- массив

после его j-й коррекции;

- array

after its j-th correction;

ΔJ1 is the j-th array of sums in the interval T of intersymbol interference in the TELE portion, which is calculated by the initial 2PB of functions F1 and F2, respectively, from the pilot signal and the result of 1PB:

SECOND PHASE:

- массив

после его j+1-й коррекции;

- array

after its j + 1st correction;

ΔJ2 - j + 1st array of sums in the interval T of intersymbol interference in the TELE portion, which is calculated by the final 2 РВ of functions F1 and F2, respectively, from the pilot signal and the result 1РВ corrected by the first stage:

NOTE: the above functions F1 and F2 are defined further in section V - METHOD ALGORITHM ...

Further, from the FIRST and SECOND STAGES, a CYCLE is organized with the condition of leaving it:

With (4) in (2) and (3), the correcting vectors ΔJ1 and ΔJ2 will be zero, which indicates that the errors in the transmitted information from those listed in Ch. I and III distortion and interference (other than Gaussian noise) will be corrected.

Consideration of both quadratures (subchannels) simultaneously would make it difficult to understand the principle of operation of the distortion compensator, because: 1) procedures (2), (3) and (4) would additionally become more complicated; 2) it would be necessary to take into account inter-channel interference - add THREE more procedures similar to those discussed above.

ALL the mentioned procedures for 2 subchannels are considered further in section V - ALGORITHM OF THE METHOD ...

Once again, we draw your attention to the fact that we are not talking about correcting ALL of the above errors, but only about correcting errors from distortions and interference of "reasonable" INTENSITY, determined by the Graph of Fig. 4 (see. Graphic images), where "reasonableness" is determined by the probability Roche of information errors at the output 1РВ, and КС - the multiplicity of 1СВ. For each multiplicity of CC, ALL errors (except for errors from Gaussian noise) are corrected if telecommunications are in the blue area. 4, where the above CYCLE converges. In the purple area, this CYCLE diverges, and in this case, the RUIP method provides an alarm about an emergency (see Chapter V-ALGORITHM OF THE METHOD ...). Exit from it requires improving the quality of the communication channel or reducing the COP performance.

Graph Fig. 4 confirms the NONLINEARITY of 2РВ: with a LINEAR increase in the productivity of the KS, the blue area under the curve of the Graph increases NONLINEARLY (accelerated), which determines the correcting ability of the RUIP method.

V. METHOD ALGORITHM IN MATLAB LANGUAGE

Vi. FEATURES OF SYNCHRONIZATION.

Synchronization in the RUIP method is performed by an energy-consuming pilot signal. that after its nonlinear AS-transformation at the reception it provides:

1) "INSTANT" COMMUNICATION;

2) extraction of the REFERENCE OSCILLATION from the pilot signal for coherent processing of information signals;

3) INVARIANCE of the system to the DOPLER EFFECT and other frequency discrepancies in the communication channel based on the properties of the AS-transformation);

4) D times greater noise immunity of synchronization relative to information signals, which characterizes its high reliability.

Vii. FIGHTING THE METHOD WITH MULTI-BEAM DISTORTIONS

The RUIP method can effectively deal with multipath distortions by creating a protective (empty) gap between the signals on the transmission (see above METHOD ALGORITHM).

If the delays to the field beams relative to the main beam do not exceed the "width" of the guard gap, then the PORTION of the transmitted information during multipath propagation of electromagnetic or acoustic waves will remain SUCH, but with a changed operating signal Uc equal to the sum of ALL beams (signals), including the main ...

But, with a high intensity of additional beams, interference between the frequencies of the beams will be observed, at which some frequencies of the PORTION spectrum may turn out to be zero (addition of frequencies with opposite phases and equal amplitudes), which means COMPRESSION of the operating frequency band of the communication channel, leading to an EMERGENCY state of the system.

The way out of THIS state is to REDUCE the width of the spectrum of frequencies occupied by a PORTION by REDUCING the information transfer rate by INCREASING the parameter "n" (see V. METHOD ALGORITHM ...).

Modeling in the MatLab environment of the above-described properties of additional rays showed that the EMERGENCY state of the channel will occur less often with an INCREASE in the number and intensity of additional rays

CONCLUSION: the more INTENSIVE the multipath, the more RELIABLE the connection!

VIII. FIGHTING THE METHOD WITH ADDITIVE CONCENTRATED INTERFERENCE

In the RUIP method, information carriers are signals with a base D >> 1. They are D times better resistant to additive impulse and harmonic interference in comparison with simple (D = 1) carrier signals, which are mainly used in modern telecommunication systems.

Concentrated interference in time. Let impulse noise or short-term interruption of communication v occur in the channel no more often than once per time T, having a uniform spectrum in the band F. The ratio of the energy E of the signal U and the energy Ev of the interference v at the receiver input is:

At the output of the correlator of the 1st sweep, each spectral component of the interference will be decomposed in terms of Hill's functions and will have a spectrum with the width:

Therefore, the width of the energy spectrum of the entire interference:

Since the receiver integrator bandwidth is:

where D is the base of the signal U, then only 1 / D of the energy of the impulse noise will fall into it.

Thus, at the output of the correlator, we have an improvement in the signal-to-noise ratio in comparison with its input by a factor of D, i.e.

Concentrated interference in frequency. Let the frequency of the harmonic interference μ be equal to the average frequency of the band F, and the ratio of the signal power to the interference power at the receiver input is

where β is the signal-to-noise ratio when transmitting information by simple signals, i.e. with base Do = l.

At the output of the correlator of the 1st sweep, the harmonic noise (sinusoid) will be expanded in terms of Hill's functions and the width of its energy spectrum will be equal to:

Based on (7), the 2 / D part of the interference power will fall into the integrator band, and at the output of the receiver integrator we will have:

Thus, in the signal / harmonic noise ratio at the output of the correlator, in comparison with the input (6), we have a gain of D / 2 times.

It is easy to show that if the frequency of the sinusoid lies at the edge of the range F, then the gain in the signal-to-noise ratio will be approximately D times.

It should be noted here: 1) the considered interference, in general, is eliminated by the 1st scan in the receiving part of the RUIP method with the correct choice of the base D of the signal and the multiplicity of the 1st convolution: 2) additive high-intensity interference distorts the shape of the transmitted signals, i.e. .. are at the same time multiplicative distortions and are eliminated by the 2nd scan.

IX. FIGHTING WAY WITH ERRORS IN BINARY CHANNELS

The RUIP method, in addition to working on analog communication channels through a DAC and ADC, can EFFECTIVELY serve BINARY communication channels, including fiber-optic ones, through time multiplexing.

The binary channel is a "black box", to the input of which the transmitted information is received from the transmitter in the form of "1" - there is a pulse and "0" - there is no pulse, and before the information there is a synchronization signal in the form of an alternation of "1" and "0". The receiver receives the transmitted information from the output of the "black box" in the form of a sequence identical to the above, but with possible errors.

The ability of the RUIP method to correct errors in binary channels, as well as to combat impulse noise (see VIII above), is based on the use of signals with a large base D as a carrier of digital data. It can be assumed that the error in the 1st digit of numbers corresponds to the smallest intensity, and in the 8th digit - the highest intensity of the impulse noise in the interval T of the signal.

The information intended for transmission through the "black box" in decimal code comes from the RIO to the RUIP (see Fig. 1), where the 1st convolution of the complex signal (with the base D >> l) with the incoming data takes place and the connection to him a pilot signal. The result of these operations is converted from decimal to binary and fed to the input of the "black box".

At the reception from the output of the "black box", the binary sequence with errors, having passed the binary-to-decimal converter, is processed by the receiving part of the RUIP method, where in the 1st scan, signal distortions in the channel are eliminated, leading to errors, and in the 2nd scan, "Cleaning up" the work of the 1st sweep. The result of these operations in decimal code goes to the receiving part of the RUIP.

In order for the error correction to be the most effective, it is necessary to be guided by the graph of FIG. 5 (see. Graphical images), obtained as a result of modeling in the MatLab environment: it is necessary to correctly select the signal base D and the multiplicity of the 1st convolution according to the schedule, taking into account the statistics of the probability of "black box" errors (non-observance of this manual may lead to the multiplication of errors) ...

SUCH SCHEME OF CONSTRUCTION OF TELECOMMUNICATIONS IS CAPABLE WITHOUT LOSS OF PERFORMANCE TO SIGNIFICANTLY REDUCE THE FREQUENCY OF ERRORS, REGARDLESS OF THEIR PHYSICAL ORIGIN.

For example, in fiber-optic channels, this will significantly increase their length to re-reception or amplification.

Modeling the above in the MatLab environment showed that in the interval T at D = 512 and KM = 6, 1 BIT error with a weight of 8 or all erroneous Bits are CORRECTED if the SUM OF WEIGHTS CORRECTED BITS DOES NOT EXCEED 8 (for example, 2 Bits are corrected with with weights 2 and 6, or 3 Bits with weights 1, 2 and 5, or 3 Bits with weights 1, 3 and 4, and 8 Bits with weights 1).

In the RUIP method, the portion length is usually Tn >> T, then the total number of corrected Kosh (8) errors in 8 digits of portion numbers can be:

Kosh (8) >> Тn / Т.

From FIG. 5 that, in principle, it is possible to correct all errors by increasing the base D of the signal or decreasing the multiplicity of the 1st convolution of the CS.

CONCLUSION: CORRECTING ABILITY has a place HERE due to the REDUNDANCY in the transmitted information (blue part of Fig. 4) or an increase in the cost of equipment - D. or a decrease in productivity - KS.

X. PROTECTION OF INFORMATION

Telecommunications with the RUIP method, in addition to the usual work with one exchange protocol, are capable of working in the limit with an unlimited number of original exchange protocols, which gives the connection a new property - to almost completely exclude the INTERCEPTION of the transmitted information by "competitors" even if they have the "hardware" of the RUIP method and powerful decryption tools.

This is achieved due to the presence on the transmitting and receiving sides of the same banks of numbered protocols. In the transmitter, each piece of information randomly selects from the protocol bank the exchange protocol number assigned to it and transmits this number to the receiver along with the useful information with the previous piece. The receiver, having this number, selects from the protocol bank the exchange protocol corresponding to this number to receive the next portion.

The complete elimination of INTERCEPT is achieved by the fact that the decryption of the transmitted information by the "competitors" is possible only if they have decryption items, i.e. portions. But, in turn, there can be no decryption items, since The "competitor" will not be aware of the batch exchange protocols.

Obviously, all of the above will work if the exchange protocol of the 1st portion of the communication session is known at the reception. A secure solution to this problem is not provided here.

XI. PROSPECTS FOR USING THE METHOD OF RUIiP

"If the idea does not seem absurd at first, then it is hopeless ..."

Albert Einstein

In addition to using the RUIP method at the physical level of telecommunications, we see the prospect of introducing it at the highest level - applied.

At the beginning of the communication session between the machines (see above the seven-level architecture of data transmission in networks [9]), the PROCESS is packed into the PAC by the transmitting part of the modern telecommunication system (see Fig. 1, Graphic images), including the RIPM method.

At the reception, the PAK is unpacked by the receiving part of the above system, including the RUIP method.

Such an organization of telecommunications computers has the ability to eliminate ALL PROCESS ERRORS when transmitting it with potential performance over the networks mentioned above.

The formulated idea can be applied to IP networks of IPv6 versions (and higher) and to projected networks based on fiber-optic communication channels.

All of this can be extended to transferring data INSIDE machines.

XII. CONCLUSION

The most effective use of the RUIP method is in a hydroacoustic environment.

The ocean is a heterogeneous medium with strong stochastic properties (unpredictability, chaos). This makes stable data transmission in an underwater environment very problematic, including through an underwater sound channel (SSC) The stochasticity of the ocean with respect to sound propagation in it is generated by dispersion (changes in the speed of sound frequencies), refraction (curvature of the trajectory), reverberation (reflections), diffraction (scattering). The high dynamics of the listed anomalies makes the impulse response of the SSV substantially non-stationary. All this leads to a lot of multiplicative (modulating) distortions in the transmission of information over the PZK, such as significant frequency distortions (amplitude and phase), multipath, Doppler effect, fading, and others. With a small time interval of stationarity of the impulse response of the PZK, the fight against distortions in it on the basis of adaptive methods of signal processing (modems) becomes very controversial.

In the discussed method of RUIP, an invariant method of generating signals in transmission and processing them in reception is proposed, which is not sensitive to all known multiplicative (modulating) distortions of "reasonable" intensity. Considering all the contents of the discussed RUIP, we are sure that this innovation is the ONLY CHANCE to achieve stable data transmission, including speech, in a hydroacoustic environment.

Nowadays, RUIP can be used effectively in long-range and ultra-long wave radio communications, in space electromagnetic and underground acoustic telecommunications, as well as in the implementation of the “last mile” in peripheral communications and wired MILITARY communications in electronic warfare conditions.

The RUIP method can be widely used in taiga In systems of operational, covert and reliable CONTROL of space, air, ground, surface and underwater objects.

TELEVIDENCE will receive absolutely noise-immune frame and line synchronization from the METHOD and all the advantages described in this work.

Literature

1. Baklanov I. Multiparametricity and the Renaissance of Analog Thinking, Metrotek, M. 2004.

2. Lagutenko OI Modern modems, "Eco-trends", M. 2002.

3. Usanov A.S. Harmonic data transmission corrector, patent SU 684747.

4. Grigoriev V.G., Usanov A.S., Ivanov Yu.A. Accumulating multiplier, patent SU 987618.

5. Usanov A.S., Finkelstein E.Z. The use of time-frequency duality to establish some relationships between complex signals and their spectra. "Proceedings of educational institutions of communication", vol. 67, 1974.

6. Okunev Yu.B. Communication systems with invariant characteristics of noise immunity, "Communication", M. 1973.

7. Sergienko A. B. Digital signal processing, "PETER", M., 2006.

8. Vosencraft D., Jacobe I. Theoretical foundations of communication technology, "Mir", M., 1969.

9. Yakubaytis E.A. Computer network architecture. "Statistics", M., 1980.

10. Wackman D.E., Sedletsky P.M. Issues of synthesis of radar signals, "Soviet radio", M., 1973.

11. Varakin L.E. Complex signal theory. "Soviet radio", Moscow 1970.

12. Anufriev I. Self-instruction manual MatLab / 5/3/6 / x. "BHV-Petersburg", 2003.

13. Dyakonov V.P. MatLab 6 / 6.1 / 6.5 + Simulink 4/5 in Mathematics and Simulation. "SOLON-Press", M., 2003

14. Rudakov P.I., Safonov V.I. Signal and image processing (MatLab 5x). "DIALOGMEPHI", M. 2000.

APPENDIX TO THE DESCRIPTION A way to achieve potential performance, noise immunity, and connection speed in communications.

Multipath signal propagation - 5 beams

EYE DIAGRAM at the OUTPUT of the 1st SCAN

EYE DIAGRAM at the OUTPUT OF 2ND SCAN

AS-SIGNAL CONVERSION

Consider the block diagram in Fig. 1, where 1 delay lines of signals for the time T with taps, 2 - multipliers, 3 - adder:

The scheme works as follows. If the input receives a signal A (t) of duration T, having a spectrum of width F, then a signal S (t) is formed at its output:

where Δt≤1 / 4⋅F is the signal delay time between adjacent taps of delay lines 1;

N = 2⋅T / Δt - for Δt≤1 / 4⋅F, where N is the number of taps of the delay lines;

N = 2⋅F⋅T = D - for Δt = 1 / 4⋅F, where D is the base of signals A (t) and S (t}.

Expression (1) is called AS-signal transformation [1]. Its short form:

In integral form, AS-transform (1) can be written as a pair of Fourier transforms as follows:

where A (t) is the converted signal at the input of the converter;

S (t) - signal at the output of the AS-converter;

C (f) - spectrum (complex) signal A (t) at the output of the AS-converter.

Representing a general signal A (t) as its sum of even and odd parts and using (3), we obtain the AS-transformation:

where Rch (t) is the autocorrelation function of the even part of Ah (t);

Rн (t) - autocorrelation function of the odd part An (t);

Rчн (t) - cross-correlation function Ач (t) and Ан (t).

AS-transform (4) in the frequency domain looks like this:

where Gch (f) is the amplitude spectrum of the even part of Ah (t);

GH (f) - amplitude spectrum of the odd part An (t);

Gchn (f) - mutual frequency spectrum (complex) Ach (t) and An (t).

For symmetric signals Ac (t) with even or odd symmetry and with a given duration T from (3), (4) and (5) we have:

where Sc (t) is the signal symmetry function Ac (t);

Sc (f) is the amplitude spectrum of the Ac (t) signal.

Then, from (6), based on Parseval's equality, we obtain:

where R (t) is the autocorrelation function of the Ac (t) signal.

Thus, the AS-transformation of any symmetric signals (or symmetrical parts of signals) calculates the halves (in amplitude) of their autocorrelation functions, compressed in time by a factor of 2.

From what has been said, it follows important: AS-transformation is a nonlinear function (the width of the spectrum of the output signal is 2 times the width of the input signal).

1. AS-conversion of simple signals.

Simple signals are called signals in which the base D = 1, and the phase and phase-frequency spectrum are equal to zero, i.e.

Simple signals and their spectra are only even or odd functions and to them, as to symmetric functions, relation (7) is fully applicable.

For the case when the autocorrelation functions of the signals coincide in shape with the signals themselves, the following AS transformations take place:

Thus, the AS-transformation of the above signals is a deformation of the signal ambiguity plane (“time-frequency” plane), which changes inversely proportionally to the time scale of the t and f axes by 2 times.

For periodic simple signals, expression (6) will be written as follows:

where N is the number of periods of the signal A (t) in the interval of its processing.

Expanding А (t), in a Fourier series, we write for signals with even symmetry:

where

Thus, AS-conversion of periodic simple signals allows simultaneous shaping, frequency multiplication and amplification.

2. AS-transformation of complex signals.

Complex signals are signals in which the base D >> 1 and the phase and phase spectrum are not equal to zero, i.e.

Consequently, complex signals are arbitrary waveforms and in the processing interval T can contain: 1) both symmetrical and asymmetrical parts - we will consider them as interference (note that based on (4) there are no completely asymmetric signals); 2) only symmetrical parts - useful signals created specially (Rh = 0 or Rn = 0).

Let us give an example of a diagram of a symmetrical signal As (t) with a base D = 1024 - Fig. 1:

The diagram shows that the signal Ac (t) consists of two identical halves facing each other in the horizontal plane.

Below are the diagrams of the AS-converter operation with this signal.

As mentioned above, each arbitrary signal contains both balanced and unbalanced parts. We call these signals interference. The ideal interference in this sense is "white" Gaussian noise. Below we will investigate its influence on the operation of the AS-converter for signal A (t) with base D.

As shown by tests of the mathematical model of the AS-converter, the transmission coefficient of the implementation of noise n (t) through it will be as follows:

The useful signal A (t) converted into the symmetry function S (t) has an amplitude transfer coefficient:

From (10) and (11) follows the value of the gain of the signal-to-noise ratio V at the output of the AS-converter in comparison with the signal-to-noise ratio at its input:

The above is illustrated in FIG. 8, 9, 10 and 11. So, if the signal base D = 1000, then with the signal-to-noise ratio at the input of the AS-converter 0 dB (the noise and signal powers are equal) and with the signal-to-noise ratio - 6 dB (the noise power in 2 times the signal strength) AS-conversion gives a 26 dB gain.

It is obvious that by increasing the signal base it is possible to achieve an arbitrarily large gain in the signal-to-noise ratio at the output of the AS-converter.

Let us formulate the main properties of the AS-transformation.

I. In the class of complex signals, AS-transformation has pronounced selective properties. From the whole variety of signals, the AS-converter is adequate only for symmetrical signals: it “selects” them, amplifies them by D / 2 times and compresses them by 2-D times. The interference is amplified only by the square root of D times and is compressed due to the expansion of their spectrum only 2 times.

II. In contrast to the matched filtering, which calculates the autocorrelation function of the input signal having the main lobe width T, the output signal (symmetry function) of the AS-transform has, for the same base D, the main lobe width T / 2.

III. The input signal and the output signals of the AS-transformations, in contrast to the autocorrelation functions, have the same dimension, which allows the latter to be switched on in stages. In this case: a) each cascade decreases Tl by a factor of 2; b) increases the amplitude of the main lobe proportionally to D; c) proportionally to D, it reduces the side lobes of the symmetry function.

IV. It is obvious that the AS-transformation is invariant to the frequency divergence in the communication channel.

In the RUIP method, the AS-transformation of the complex symmetric signal As (t), which is here the PILOT SIGNAL (see the DESCRIPTION), is used to synchronize the instantaneous transportation of the PORTION from the transmitter to the receiver (without taking into account the transmission medium delay + the duration of the pilot signal) - message reception (the body of the portion) starts at the moment of calculating Sc (t) in (7).

The above is illustrated by the following Fig. 2, where "P-S" is the pilot signal:

Let us give two more possibilities of using the AS-transformation in radio engineering. 1. In telecommunications. Asynchronous transmission-reception of low-speed digital information in accordance with (7): "+, 1" - positive symmetry, "-, 0" - negative symmetry (see Fig. 3):

2. At the location (see Fig. 4):

A distinctive feature in all three examples of AS-conversion application, the "P-S" transmitter, in essence, radiates the signal itself into the channel together with its reference oscillation, thereby preparing the correlation processing of the signal at reception (the loss relative to coherent reception is only 3 dB due to " noisiness "of the reference oscillation - the second half of the pilot signal).

In addition, all three examples provide invariance (insensitivity) to all multiplicative (modulating) distortions.

The aforementioned is possible provided that both halves of the signal are equally distorted and there is no group deceleration time (GDT) - "pull-in". Unfortunately, not all types of multiplicative (modulating) distortions satisfy the above condition, for example, multipath and linear.

In Fig. 5 shows the AS-conversion circuit free from the indicated drawback - the delay line 4 is introduced:

Then Figures 2, 3 and 4, where the distorted signal is marked in red, will take the form (see, respectively, Fig. 5, Fig. 6 and Fig. 7):

From the last 4 figures it can be seen: 1) instead of a symmetrical signal, the concept of IDENTICAL SIGNAL Aa (t) was introduced, consisting of TWO consecutive identical signals through a protective TIME INTERVAL greater than the maximum allowable delay time of the signal in the channel (delay line "4" in Fig. 5); 2) at the reception, the SIGNAL SYMMETRATION block "SS" is introduced.

Consider FIG. 12 and FIG. 13. The black graph shows an identical signal on the transmission "d - d" with a base D = 1024 and duration "d - d" T = 2048 and its AS-transformation calculating the symmetry function Sc at the point T / 4 + 1 (see. 3- line on the graph). It can be seen that the Sc function has twice the width of the spectrum than the input signal and is positioned on the graph in the center of the "gravity" of the undistorted signal. Moving on to the red graph, we see that the Sc function did not change its position at the reception and is also in the center of gravity of the already distorted signal.

CONCLUSIONS. AC-transformation of IDENTICAL signals: a) invariant (insensitive) to ALL multiplicative (modulating) distortions, if these distortions - GVZ does not exceed the protective gap; b) as for the location, in case of many ray distortions, it sums up all the beams in power that have delays relative to the main beam <= GVZ and determines the coordinate of the range of the center of gravity of the beam sources, including the main beam.

The Appendix contains diagrams explaining the formations: a) symmetry functions S (t); b) autocorrelation function R (t).

Literature:

Usanov A.S. Harmonic data transmission corrector, patent SU 684747.

Claims (1)

The method of telecommunications to achieve potential performance, noise immunity and connection speed under conditions of errors in analog communication channels from the action of all multiplicative distortions and additive interference, as well as errors from nonlinear distortions, in addition, in conditions of errors in binary communication channels, regardless of their origin, is in that information is transmitted via communication channels in portions of multi-valued quadrature sequences of bytes; on transmission in each portion: create a working signal Uc with a guard gap from the harmful effects of multipath distortion; from the information vector arriving for transmission in the decimal code IUY, two quadrature components IU and IY are formed, over which the 1st convolution is performed with the orthogonal carrier oscillations U and Y, obtained by the Hilbert transform of the signal Uc, as a result, the body of the portion S (signal "constellation" ); a header is attached to the body of the portion, consisting of two oscillations of increased power AU, AU and a portion is formed with an empty space Uo - SK = [Uo, AU, AU, S]; this portion is read either through a DAC into an analog communication channel at a maximum speed corresponding to the highest frequency of its amplitude-frequency characteristic, matched, taking into account the Doppler effect, with the amplitude spectrum of the signal Uc, or through a code converter into a binary communication channel with a speed determined by its clock frequency ; then, in the above communication channels, the 2nd convolution of the portion with their impulse characteristics, introducing the above-mentioned distortions of the transmission path into the transmitted information, occurs; at the reception in each portion: each period from the communication channels through the ADC or the inverse converter of the code receive a digital signal in decimal code and for one cycle of the clock frequency of the ADC or the clock frequency of the inverse converter of the code calculate the AS-conversion of the symmetric header of the portion [AUk, fliplr (AUk) ] and get a compressed autocorrelation function, which separates the header of the portion from the body of the portion; then from the header form the oscillation AU1 = (AUk + AUk) / 2, it is transformed according to Hilbert with the result of two orthogonal oscillations AUk and AYk, using which, they perform the 1st scan of the portion body, which returns the quadrature components IU1 and IY1, and decoding them, have distorted transmitted information IUY1; then the function F1 from the header of the portion and the function F2 from the result of the 1st scan IUY1 are subjected to the 2nd scan in two stages - the initial 2nd scan and the final 2nd scan, and, combining these stages, the distortions of the communication channels are returned in the form 2 vectors of intersymbol and 2 vectors of interchannel interference; By subtracting the obtained vectors from the result of the 1st sweep and decoding the differences, an exact copy IUY5 of the transmitted information IUY is obtained - all the errors described above are corrected, except for errors from the action of the admissible additive Gaussian noise.

Images