RU2550086C1 - Method of decoding discrete signals propagating in multibeam channel - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: instead of the prototype-valid estimation an information-bearing cyclic time shift of an m-sequence in each beam separately and averaging of results of said estimation across all beams, the present invention includes coherent accumulation (each separately) of an information pulse arriving on all beams, and generating the desired estimate of the time shift of the m-sequence based on the single-beam information pulse generated as a result of said coherent accumulation.

EFFECT: low decoding error probability.

3 dwg

Description

The invention relates to the field of transmission of digital information and is intended for use in decoders of communication systems operating in a multipath channel.

The main problem facing the developer of a device for decoding discrete signals transmitted through a multipath communication channel is the dissipation of the signal energy over time, which, in the absence of technical measures to compensate for this effect, leads, firstly, to a decrease in the signal-to-noise ratio on the device forming the result of decoding, and secondly, to intersymbol interference, which reduces the quality of decoding even with an arbitrarily large signal to noise ratio.

To solve this problem, an a priori knowledge of the current form of the channel impulse response (IRF) is necessary. With the known KFM, the entire alphabet of transmitted characters can be predicted, i.e. recounted to the point of reception. This recalculation (prediction) is carried out as a convolution of each of the transmitted symbols from the KFM. In this regard, almost all known solutions related to the design of multipath signal decoders are somehow based on radiation along with information (i.e., unknown at the receiving end of the communication system) symbols or pulses (or signals) of test or test signals (pulses) of a known form by which the current KFM is evaluated; the shape of the test signals at the receiving end is a priori known. Such a transmission principle is referred to as a “system with a test pulse and prediction” (or SIIP) (see, for example, [1], p.3.1, in particular, the footnote on p.109). In this case, the KFM is estimated from a series of test pulses when these pulses are transmitted between the information pulses (see, for example, [1], Fig. 3.1. On p.108; [2], Fig. 3.4 on p.123 and [3], section 15.7.1, p.1013). Moreover, the accuracy of estimating the KFM is low due to the small energy of the sequence of test pulses, since they are transmitted briefly, and most of the time is occupied by the transmission of information pulses.

This drawback is deprived of the method of decoding discrete signals described in [4], which provides, like the aforementioned analogue, the initial estimation of the KFM by the test pulse preceding the series of information pulses of the transmitted message, with a further transition (in contrast to the specified analogue) to the estimation of the KFM by generating particular estimates of this KFM for each of the information impulses, iterative formation of the resulting KFM estimate by accumulating (averaging) the sequence of these particular prices with the subsequent determination of the relative delay times of the rays (i.e., the delay times relative to the ray arriving first; hereinafter, for brevity, the word "relative" is omitted). Such a transition becomes possible as information pulses are received and the form (as part of the decoding task) of each of them is determined. After such a determination of the shape of each n-th information impulse, based on the assessment of the KFM generated before processing this pulse, this information impulse from the point of view of the possibility of using it to estimate the KFM becomes equivalent to the test pulse. Directly for processing multipath information signals in this object, a set of arrays of beam delay times obtained from the estimated instantaneous KFMs is used. This method is further considered as a prototype. The block diagram of the prototype is shown in figure 3, where indicated (below used the numbering of the features of the prototype, corresponding to the continuous numbering of the corresponding features of the proposed method (see figure 1)):

- 1 - buffering of received pulses;

- 2 - formation of the assessment of the KFM by the test impulse;

- 3 - formation of an averaged assessment of the KFM;

- 4 - formation of an estimate of the array of ray delays;

- 5 - formation of an assessment of the KFM for each information impulse;

- 7 - determining the shape of each of the information impulses;

- 8 - the formation of the decoding result of each information pulse;

- 18 - determination of the time shift (OT) of the information signal in each beam;

- 19 - accumulation of estimates of the OT information signal for all beams.

The prototype is designed to work with a synchronous communication system in the following time diagram of the transmitted message. First, the so-called leading sync signal (hereinafter referred to as the test pulse) is transmitted, and then the information block (hereinafter referred to as the set of informational pulses) is transmitted (see Fig. 5 in [4]). For simplicity of presentation, below all objects are described with reference to the situation of using a single test pulse. In the general case, this impulse can be a series or a pack of impulses composing it, which, in principle, does not change anything in the description of these objects.

An essential (from the point of view of achieving a positive effect in the claimed object) feature of the principle of operation of the prototype is that it includes the implementation of the evaluation of the carrier of information OT in each of the received information pulses for each beam individually, followed by the accumulation of these estimates of OT for all rays. This feature predetermines the relatively low quality of decoding, since this quality directly depends on the accuracy of estimation (determination) of the airspace, and for an error in determining the airspace by the totality of beams, it is practically enough in the address specifier shown in [4] in Fig. 7 to mistake at least one emission of interference for a signal in the beam (more precisely, in a kind of pseudo-beam). When this error is committed, a correspondingly erroneous determination of the OT of the information signal (supposed, but actually absent) in this pseudo beam occurs. The indicated error, expressed in units equal to the correlation interval of the transmitted pulse with phase shift keying, is limited from above only by the value of the base of the m-sequence B = Δf · T (where T is the duration of the used m-sequence, Δf = τ ^-1 , τ is the period of possible switching phase during phase manipulation, equal to the correlation interval of the transmitted pulse); the latter in practice is B = 2 ⁹ ... 2 ¹⁰ . Moreover, due to the high probability of a significant error in determining the EO in a pseudo ray (since this EO in this situation is determined by the implementation containing only noise), a so large error in the result of the accumulation of EO in all beams is quite probable that it is enough for an error in decoding a symbol. To make an error in decoding a symbol, an error is sufficient as a result of determining the EO in only one correlation interval of the transmitted pulse equal to τ (as noted above, with a possible error of one of the results of the EO estimation in a separate (pseudo) ray up to (2 ⁹ ... 2 ¹⁰ ) τ). Thus, the disadvantage of the prototype is the low quality of decoding or a high probability of errors.

The aim of the proposed method is to improve the quality (reduce the likelihood of errors) decoding.

на основе каждой сформированной оценки ИРК H_n(k), а также формирование результата декодирования каждого информационного импульса на основе результата определения его формы, осуществляется накопление сигналов во всех лучах по каждому буферизованному n-му информационному импульсу в отдельности с использованием результата формирования оценки текущего массива задержек лучей $${\overrightarrow{K}}_{n-1}$$

, а последовательное определение формы каждого из информационных импульсов осуществляется на основе результатов указанного накопления сигналов во всех лучах.The goal is achieved by the fact that in the method of decoding discrete signals propagating in a multipath channel, including buffering the received pulses, forming an estimate of the channel impulse response (IRF) H _{n = 1} (k) (where k is the argument of the discrete time) from the buffered received first test pulse , the shape of which is known in advance, determining the shape of each of the next n-th (n≥2) pulses, which are informational, and forming, taking into account the result of this determination, for each received and buffered n-th (for n≥2) the informational impulse of the estimate of the KFM h _n (k), as well as the average estimate of the KFM H _n (k) by weight summation of the similar estimate of H _n-1 (k) generated when the (n-1) -th pulse is received and the current estimate of the KFM h _n (k), forming an estimate of the array of ray delay times $${\overrightarrow{K}}_n$$

on the basis of each generated estimate of the KFM H _n (k), as well as the formation of the decoding result of each information pulse based on the result of determining its shape, signals are accumulated in all beams for each buffered n-th information pulse separately using the result of the formation of the estimate of the current array ray delays $${\overrightarrow{K}}_{n-\mathrm{one}}$$

and sequential determination of the shape of each of the information pulses is based on the results of the indicated accumulation of signals in all beams.

The above indication that k is an argument of discrete time means that at the sampling frequency of the signal f _d taken at a discrete time moment, the count of the function (signal) x (t _k = k / f _ä ) corresponds to the notation x (k), t .e. the recording of the time argument k is the abbreviation of the recording of the time argument t _k = k / f _ä = k · Δ, where Δ is the period of the sampling frequency.

The block diagram of the proposed method is shown in figure 1, where indicated:

- 1 - buffering of received pulses;

- 2 - formation of the assessment of the KFM by the test impulse;

- 3 - formation of an averaged assessment of the KFM;

- 4 - formation of an estimate of the array of ray delays;

- 5 - formation of an assessment of the KFM for each informational and impulse;

- 6 - the accumulation of signals in all beams upon arrival of each n-th information pulse;

- 7 - determining the shape of each of the information impulses;

- 8 - the formation of the decoding result of each information pulse.

At the input of the device implementing the inventive method, the received signal mixture (as in the prototype, first a series of test pulses, and then a combination of information pulses) with noise is received in real time. The following is a description of an embodiment (one of the totality of equivalent options) of the proposed method and its implementing device. To explain the essence of the operations of the proposed method, as well as the principle of its action, figure 2 shows a block diagram of a device that implements this method, where are indicated:

- 9 - block buffer memory;

- 10, 14 - respectively, the first and second correlator;

- 11 - drive correlation function;

- 12 - block delay detection of rays;

- 13 - block accumulation of signals in the rays;

- 15 - block forming the reference signal;

- 16 - block determining the magnitude of the OT;

- 17 - block for the formation of the decoding result of each information pulse.

In the inventive method, operation 1 of the buffering of received pulses is performed by block 9 (figure 2) with the same name. This block contains two memory areas. Buffering, as in the prototype, undergoes the implementation of the received pulses (signals) of duration T + t _irk , where t _irk is the expected duration of the KFM (or the expected time delay interval). In this case, the first signal implementation (containing the first pulse arriving on all the rays) in the time interval 0 ... T + t _irk (hereinafter for zero time, relative to which the time is counted, is sequentially buffered into the first memory area of block 9, the moment of arrival of the leading edge test signal). Then, to the second memory area, the second implementation (containing the second impulse) in the time interval T ... 2 · T + t _irk , then again to the third memory area, the third implementation, etc .; The nth implementation corresponding to the processing time interval of the nth pulse is in the time interval (n-1) · T ... n · T + t _irk . Each received multipath pulse of duration T + t _irk is located in M = (T + t _irk ) / f _d conditional (i.e., storing multi-bit words) memory cells of block 9.

The two outputs of block 9 are shown in FIG. 2 (and, accordingly, the two outputs of operation 1 — in FIG. 1) and appear conditionally in the present description to illustrate that when processing each (i.e., test and each information) pulse to the correlator input 10, all stored in the corresponding one (the first when processing odd n-th pulses and the second when processing even odd pulses) from the departments of block 9 receives a signal of duration T + t _irk , and then L is transmitted to the input of block 13 from the same department of block 9 9 signal implementation duration Stu T start points are determined formed estimates of ray unit 12 delays. In fact, block 9 has a hardware-only output on which a signal implementation is generated, which is read at a particular moment from this block 9 to the inputs of blocks 10 and further 13.

The operation 2 of forming the estimate of the KFM by the test pulse is performed, as in the prototype, by the first correlator 10. Moreover, in this phase of the claimed object, the reference function of the first correlator 10 is the test pulse, the shape of which is a priori known. The correlator 10 in all situations calculates a linear correlation between the implementation of the received signal read from the buffer memory unit 9 and its own reference function. The version of the reference function, which coincides with the test pulse, is stored in the long-term memory included in the correlator 10. An embodiment of the first correlator 10 is described, for example, in [5, the block diagram in Fig. 5.14, p. 295]. In this case, the signal input of this correlator is the lower input in the indicated Fig. 5.14, to which the received signal x (n) is supplied; the reference function of the correlator (indicated in Fig. 5.14, it is designated as h (n)), as noted above, is stored in its memory (generally long-term or operational), not shown in Fig. 5.14 for simplicity. When processing each n-th pulse, the entire signal implementation stored at that moment in block 9 is read to the input of the correlator 10. The discrete Fourier transform (DFT) operation is performed over it in the correlator 10, then the result of this operation is multiplied by the result of the DFT operation on the correlator support function, after which the inverse DFT operation is performed from the product result array. When performing the inverse DFT in half the spectrum (i.e., when the symmetric and complex conjugate parts are discarded), the output result is presented as an analytical signal that does not contain carrier oscillations, which is fundamentally important in light of the need for further implementation of coherent signal accumulation in the rays ( see below). When performing DFT operations on the implementation of the input signal and the reference function, these arrays are supplemented with zero samples in accordance with general rules to ensure the calculation of linear (aperiodic) correlation (see [6], sections 2.23 and 2.24).

The response of the correlator 10 to a pulse of an a priori known shape transmitted through the multipath channel is actually an estimate of the KFM (in the working frequency band).

The operation 3 of generating an averaged estimate of the KFM is performed by the correlation function accumulator 11 shown in FIG. 2. In the phase of receiving the test pulse, the correlation function accumulator 11 only translates the result of the KFM estimation by the test pulse H _{n = 1} (k) to the input of the ray delay determination unit 12, since on the time interval 0 ... T + t _irk there is still nothing to accumulate. Subsequently, when processing every n-th one with n≥2 (informational) momentum, drive 11 calculates the formation of an averaged estimate of the KFM H _n (k), for example, by the formula

where the parameters α and β determine, respectively, the scale of formation of the averaged estimate of the KFM and the value of the time interval of averaging. For example, the following choice of the indicated parameters is possible: α = (β + 1) ^-1 , β = 0.9.

размерности L. Оценивание массива задержек лучей производится для того, чтобы в итоге обеспечить когерентное сложение сигналов (от каждого n-го импульса в отдельности), пришедших во всех L лучах. Для достижения указанного эффекта необходимо оценивание задержек со стандартными погрешностями порядка долей интервала корреляции каждого импульса; при этом принципиально важно то, что коррелятор 10 обеспечивает формирование оценки ИРК в виде аналитического сигнала. В противном случае требовалось бы обеспечение погрешностей оценивания задержек со стандартными погрешностями порядка долей периода несущей частоты, что было бы проблематично.The operation 4 for generating an estimate of the array of ray delays is performed by the block of the same name 12. In this block, all samples of the signal generated by the drive 11 (i.e., a set of samples of the signal, which is a function of discrete time), are compared with a threshold, and those time arguments k _l to which correspond the amplitudes of the samples of the estimate of the KFM H _n (k _l ) that exceeded the threshold are fixed as the delay times of these rays. Their set formed according to the estimate of H _n (k _l ) is a vector $${\overrightarrow{K}}_n$$

dimension L. The array of ray delays is estimated in order to ultimately ensure coherent addition of signals (from each nth pulse separately) that came in all L rays. To achieve this effect, it is necessary to estimate delays with standard errors of the order of fractions of the correlation interval of each pulse; it is fundamentally important that the correlator 10 provides the formation of the KFM estimate in the form of an analytical signal. Otherwise, it would be necessary to ensure errors in estimating delays with standard errors of the order of fractions of the carrier frequency period, which would be problematic.

; при этом массив, соответствующий каждому лучу, содержит $$B=T/f_{\ddot{a}}$$

(точнее $$B=T/f_{\ddot{a}}+1$$

, но это уточнение при больших В несущественно), отсчетов сигнала. Расчеты указанных границ каждого из L диапазонов номеров ячеек блока 9, в которых хранятся временные реализации n-го импульса, пришедшего в каждом из L лучей, могут производиться либо в блоке 12 и выдаваться в блок 9, либо непосредственно в блоке 9 входящим в его состав управляющим процессором (хост-машиной).Operation 6 of the accumulation of signals in all beams upon arrival of each n-th information pulse is implemented by block 13 as follows. As noted above, block 12 (see Fig. 2) before the processing of the nth (n≥2) pulse begins, an array of the results of estimating the delays k _{l of} each l-th beam (i.e., l-th when l = 1 ... L is the argument of the time estimate of the KFM H _n-1 (k _l )), which exceeded the threshold). Each parameter k _l defines a range of memory cells of block 9, in which the implementation of the nth pulse arriving in the 7th ray is stored; the boundaries of this range are $$k_{\mathrm{one}}...k_{\mathrm{one}}+T/f_{\ddot{a}}$$

; the array corresponding to each ray contains $$B=T/f_{\ddot{a}}$$

(more precisely $$B=T/f_{\ddot{a}}+\mathrm{one}$$

, but this refinement at large V is not significant), of the signal samples. Calculations of the indicated boundaries of each of the L ranges of cell numbers of block 9, in which temporary implementations of the nth pulse arriving in each of the L beams are stored, can be performed either in block 12 and issued in block 9, or directly in block 9 managing processor (host machine).

и запоминаются в памяти блока 13, затем с номерами в диапазоне $$k_2\dots k_2+T/f_{\ddot{a}}$$

и т.д. При этом их временные аргументы корректируются посредством вычитания из их исходных значений величины k_l, в результате чего все они «приводятся» к диапазону временных отсчетов $$0\dots T/f_{\ddot{a}}$$

, По мере указанного чтения реализации сигналов, пришедших в разных лучах, в блоке 13 реализуется суммирование их одноименных по времени (с учетом указанной коррекции времени) отсчетов. Как это имеет место в прототипе, указанное суммирование может предваряться умножением временной реализации сигнала в каждом луче на видоизменяющую последовательность G(k), если такое же умножение реализуется и при формировании информационных импульсов для их передачи. В этом случае реализуемое при обработке умножение читаемой из блока 9 реализации сигнала в каждом луче на видоизменяющую последовательность G(k) эффект упомянутого умножения информационного сигнала на эту же последовательность, реализованную при формировании передаваемого импульса, «снимает» (компенсирует). Указанное использование видоизменяющей последовательности G(k) существенно снижает негативный эффект влияния на погрешность определения в дальнейшем ВЗ, имеющий место в связи с тем, что сигналы в разных лучах друг для друга являются помехами, потенциально способствующими формированию откликов второго коррелятора на информационный импульс при аргументах времени, существенно не совпадающих с искомым ВЗ. Данное явление обусловлено тем, что коррелированны с опорной функцией информационные сигналы, пришедшие по всем лучам, но все эти сигналы в разных лучах характеризуются разными задержками.Actually, the operation of accumulating the signals related to the nth pulse that came along L beams is implemented by sequentially reading the samples of the signal from block 9 to the input of block 13, and these samples are read from memory cells with numbers in the range $$k_{\mathrm{one}}...k_{\mathrm{one}}+T/f_{\ddot{a}}$$

and stored in the memory of block 13, then with numbers in the range $$k_{}$$$${}_2...{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="00000017.png"\; state="\{\{state\}\}">k</figure-callout>}}_2+T/f_{\ddot{a}}$$

etc. At the same time, their time arguments are corrected by subtracting k _l from their initial values, as a result of which they are all “reduced” to the range of time samples $$0...T/f_{\ddot{a}}$$

, As the reading of the implementation of the signals arriving in different beams is indicated, in block 13, the summation of their samples of the same name (taking into account the specified time correction) is implemented. As this takes place in the prototype, the specified summation can be preceded by multiplying the temporal implementation of the signal in each beam by a mutating sequence G (k), if the same multiplication is also realized when generating information pulses for their transmission. In this case, the multiplication of the signal read from block 9 realized in processing in each beam by the mutating sequence G (k) that modifies the information signal by the same sequence implemented in the formation of the transmitted pulse is “removed” (compensates) by compensating. The indicated use of the modifying sequence G (k) significantly reduces the negative effect of the influence on the error in the determination of the OT in the future, which occurs due to the fact that signals in different beams for each other are interferences that potentially contribute to the formation of responses of the second correlator to the information pulse for time arguments that do not substantially coincide with the desired OT. This phenomenon is due to the fact that information signals that came along all the beams are correlated with the reference function, but all these signals in different beams are characterized by different delays.

As a result, at the end of block 13, after the processing of the nth information pulse, a useful one is formed (accurate to noise), i.e. carrying the transmitted information a result of the form

where S _n (k) is the form of the transmitted n-th information pulse, A _l is the signal amplitude in the l-th beam.

The response Z (k) of the form (2) is a single-beam signal that coincides in shape with the transmitted nth information pulse and is characterized by a level corresponding to the result of coherent signal addition in all beams.

. (Справка: под циклической корреляционной функцией понимаем циклическую свертку, вычисляемую при чтении опорной функции без инверсии времени (см., например, [6], раздел 2.24).) Указанная опорная функция хранится в долговременной памяти коррелятора 14.The operation 7 of determining the shape of each of the information pulses is implemented as shown in figure 2 by the second correlator 14 and the block 16 determining the magnitude of the OT. The second correlator 14 calculates a cyclic (or, what is the same thing, periodic or circular) correlation function between the implementation of the signal Z (k) coming from the signal accumulation unit 13 and the reference functions equal to the initial (transmitted) information signal at its zero time cyclic shear $$S_{\mathrm{and}}^{\left(0\right)}\left(k\right)$$

. (Reference: by a cyclic correlation function we mean a cyclic convolution calculated when reading the reference function without time inversion (see, for example, [6], Section 2.24).) The indicated reference function is stored in the long-term memory of the correlator 14.

форма фактически принятого информационного импульса однозначно определяется величиной введенного в него ВЗ, оценка которой по n-му информационному импульсу i_{max n} выработана. При этом выполнена и операция 6 определения формы текущего информационного импульса.Block 16 determining the magnitude of the cyclic OT corresponding to the maximum of the correlation function, according to the function performed, coincides with the similar block of the prototype. It is a programmable hardware and contains, for example, memory for storing an array of time samples of the result of calculating the correlation function by the correlator 14, with which the samples of this array are alternately read to the comparison unit; as a result of comparison, for example, of all samples of the array, the maximum level is revealed and its number in the k _{max n} array is determined; this number is associated with the desired value of the cyclic OT i · τ as in k _{max n} = i _{max n} · τ / Δ. When determining from this ratio the value corresponding to the maximum of the correlation function of the time index i _{max n} , the result is rounded to the nearest whole, i.e. i _{max n} = k _{max n} · Δ / τ], where the sign of square brackets means the specified rounding to the nearest whole. With a known shape of the information pulse at zero airspace $$S_{\mathrm{and}}^{\left(0\right)}\left(k\right)$$

the shape of the actually received informational impulse is uniquely determined by the value of the OT introduced into it, the estimate of which is generated from the nth informational impulse i _{max n} . In this case, the operation 6 of determining the shape of the current information impulse was also performed.

при i_n=i_{max n}. По выполнении функции формирования опорного сигнала первого коррелятора 10 $$S_{\dot{e}}^{\left(in\right)}\left(k\right)$$

указанная функция записывается оперативную память этого коррелятора, и далее это коррелятор, начиная с момента приема n=1 импульса (т.е. первого информационного импульса), в качестве опорного сигнала использует текущий результат формирования этого сигнала $$S_{\dot{e}}^{\left(in\right)}\left(k\right)$$

блоком 15. Отличие операции 5 формирования оценки ИРК по каждому информационному импульсу от упомянутой выше операции 2 формирования оценки ИРК по испытательному импульсу состоит лишь в используемой при этом формировании опорной функции первого коррелятора 10. В остальном содержание этих двух операций полностью совпадает. В итоге выполнения операции 5 формируется оценка h_n(k), используемая для формирования осредненной оценки ИРК H_n(k) (1) при выполнении операции 3.The operation 5 for generating an estimate of the KFM for each information pulse is implemented by the reference signal generating unit 15 and the first correlator shown in Fig. 2 and the first correlator 10. Block 15 generates a reference signal of the first correlator 10 by introducing into the m-sequence used in the communication system a cyclic OT characterized by values of k _{max n} in seconds, or i _{max n} in time quanta equal to the momentum correlation interval τ. Then, on the basis of the indicated m-sequence with the indicated OT introduced into it by the same block 15, a phase-shifted signal is formed $$S_{\dot{e}}^{\left(in\right)}\left(k\right)$$

when i _n = i _{max n} . By performing the function of forming a reference signal of the first correlator 10 $$S_{\dot{e}}^{\left(in\right)}\left(k\right)$$

the indicated function records the operational memory of this correlator, and then this correlator, starting from the moment of receiving n = 1 impulse (i.e., the first information impulse), uses the current result of the formation of this signal as a reference signal $$S_{\dot{e}}^{\left(in\right)}\left(k\right)$$

block 15. The difference between the operation 5 for generating an estimate of the KFM for each information impulse from the above-mentioned operation 2 for generating the estimate for the KFM for the test impulse consists only in the support function of the first correlator 10 used in this process. Otherwise, the content of these two operations completely coincides. As a result of operation 5, an estimate of h _n (k) is formed, which is used to form an average estimate of the KFM H _n (k) (1) during operation 3.

The operation 8 of generating the decoding result of each information pulse is performed by the block 17 for generating the decoding result of each information pulse shown in FIG. It contains memory for storing the correspondence table of time indices of cyclic OT information signal (i) to the alphabet of characters of the discrete communication system {A _i } for i = 1 ... N. When applying to its input the result of evaluating the cyclic OT of the next (n-th) pleasant information impulse i _{max n} in block 17, the symbol A _{imax n} corresponding to this shift is read from the indicated table and issued to the consumer.

All blocks that implement the inventive decoding method are digital programmable hardware.

The inventive decoding method is designed for use in a synchronous communication system. In such a system, at the receiving end, the moments of the beginning of arrival of each information signal and test signals are known. In principle, it is possible, for example, to implement synchronization with the implementation of the transmitter and receiver in a single time system; while the propagation time of the signal from the transmitter to the receiver is known. In this case, the device implementing the inventive method includes a timer that issues a synchronization signal to all blocks 9 ... 17 at the time of arrival of each pulse (test or informational). Hardware synchronization in the device that implements the inventive method is not included, since the vast majority of digital (discrete) communication systems are synchronous and therefore are implemented as standard.

The technical effect achieved in the claimed object — the reduction in the probability of decoding errors — is due to the fact that in it the operation of estimating the magnitude of the OT in each information pulse is carried out based on the result of coherent accumulation of this information pulse for all rays. At the same time, the situation of evaluating the information carrier that carries information on an implementation that does not contain an information signal (which was the case in the prototype and increased the probability of decoding errors) is practically excluded.

Literature

1. Klovsky D.D. Transmission of discrete messages over the air. M .: Communication. 1969.

2. Nikolaev B.V. Sequential transmission of discrete messages on continuous channels with memory. M .: Radio and communications, 1988.

3. Sklyar B. Digital communication. Theoretical foundations and practical application. 2nd edition, 2003.

4. Krantz V.Z., Sechin V.V. Using information symbols to synchronize a communication system with complex signals. Scientific and technical collection "Hydroacoustics", 2012, issue 15, p. 36-41.

5. The use of digital signal processing. Ed. E. Oppenheim. M .: World. 1980.

6. Rabiner L., Gould B. Theory and application of digital signal processing. M .: World. 1978. 848 p., Ill.

Claims (1)

A method for decoding discrete signals propagating in a multipath channel, including buffering the received pulses, generating a current channel impulse response (TFI) estimate H _{n = 1} (k) (where k is the discrete time argument) from the buffered received first test pulse, the shape of which is known in advance , determining the shape of each of the subsequent n-th (n≥2) pulses, which are informational, and forming, taking into account the result of this determination, for each received and buffered n-th (for n≥2) information Nome pulse evaluation KFM h _n (k), and estimates the averaged KFM H _n (k) by the weighting summation analogous estimate H _n-1 (k), formed by, when receiving the (n-1) -th pulse and the current IGC estimates h _n (k), forming an estimate of the array of ray delay times

based on each generated estimate of the KFM H _n (k), as well as the formation of the decoding result of each information pulse based on the result of determining its shape, characterized in that the signals are accumulated in all beams for each buffered n-th information pulse individually using the result forming an estimate of the current array of ray delays

, and determining the shape of each of the information pulses is based on the results of the indicated accumulation of signals in all beams.

Images