RU2251811C2 - Device for signal resolution on random noise background - Google Patents

Abstract

FIELD: electrical communications; radio-controlled signal receivers, radar stations, mobile-object communication systems.

SUBSTANCE: proposed device for signal resolution on random noise background has signal processing channels l₁ - l_n, number of signal processing channels being equal to number of resolved-signal superposition combinations; clock generator 2; and comparison unit 3; each signal processing channel has signal correlation processing unit 4 and adder units 5₀ - 5_L, unit 6 for calculating summary likelihood ratio functional of current resolved-signal superposition combination, unit 7 for calculating a posterior probabilities of number combinations of resolved-signal Gaussian components, and extrapolation unit 8. Parallel-conveyer method is implemented in device for processing arriving signal including recursive calculation of likelihood ratio functional which ensures its uncriticality to noise parameters and high signal processing speed.

EFFECT: enhanced reliability of reception in the environment of variable-parameter and structural noise, and echo signals.

1 cl, 3 dwg

Description

The invention relates to telecommunication technology and can be used in receivers of radio signals, radar stations, communication systems with mobile objects.

A device for noise suppression in a communication system (patent RU 2169992 C2 “Method and device for noise suppression in a communication system”, published on June 27, 2001), which is designed to transmit information using information frames in channels, is selected as an analogue, moreover, information frames contain noise from which the channel noise estimate is obtained, the device comprising means for estimating channel energy in the current information frame, means for estimating the total channel energy in the current information frame based on channel energy estimates ala, means for estimating the power of spectra of the current information frame based on an estimate of the energy of the channel, means for estimating the power of spectra of a plurality of past information frames based on an estimate of the power of spectra of the current frame, means for determining the deviation between the estimate of the power of spectra of the current frame and estimating the power of the spectra of many past frames and means for updating the channel noise estimate based on the total channel energy estimate and the resulting deviation. The device in question also contains correlation processing units, however, the signal is processed under the assumption that the channel contains only noise, structural noise and reflected signals are not taken into account, which does not allow obtaining the required technical result in the form of a decision on which signals and in which combination present at the entrance.

As a prototype, a device for distinguishing signals against a background of arbitrary interference (copyright certificate SU 1596469 A1 “Device for distinguishing signals from a background of arbitrary interference”, publ. 09/30/1990, Bull. No. 36) was selected. The signal discrimination device against the background of arbitrary interference contains signal processing channels, the first outputs of each of which are connected to the first inputs of the comparison unit, the outputs of which are the outputs of the device, the information input of which is connected to the information inputs of the signal processing channels, and each of the signal processing channels contains a correlation block signal processing and summing blocks, the outputs of which are the first outputs of the signal processing channel, the information input of which is information the input of the correlation signal processing unit, the first and second outputs of which are connected to the inputs of the summing blocks, a clock generator is introduced, the output of which is connected to the clock inputs of the signal processing channels, the second outputs of which are connected to the second inputs of the comparison unit, the control inputs of the signal processing channels connected to the control input of the device, and in each signal processing channel, a maximum signal selection block is introduced, the output of which is the second output of the signal processing channel, t the act input of which is connected to the clock inputs of the summing unit, the signal correlation processing unit and the maximum signal selection unit, the signal inputs of which are connected to the outputs of the signal correlation processing unit.

The device, considered as a prototype, after each cycle of operation, at the outputs of the comparison block contains information about which, from the considered set of signals, was input, but does not allow to obtain the required technical result in the form of a decision on which signals and in which combinations are present at the input. In addition, the distinction device uses a vector of input samples of a sufficiently large dimension to make a decision, which significantly increases the computational complexity of processing, and with large samples it generally makes it practically impossible.

The technical problem to be solved is to increase the fidelity of reception under conditions of interference with varying parameters, structural interference, the presence of re-reflected signals by solving the problem of resolving signals.

The technical problem to be solved in a device for resolving signals against arbitrary interference, containing signal processing channels, a clock generator whose output is connected to the clock inputs of signal processing channels, the outputs of each of which are connected to the corresponding inputs of the comparison unit, the outputs of which are the outputs of the device, the information input of which connected to the information inputs of the signal processing channels, and the control inputs of the signal processing channels are connected to the control input of the device, and Each signal processing channel contains a signal correlation processing block and summing blocks, it is achieved by the fact that the number of signal processing channels is equal to a given number of combinations of overlapping allowed signals, each signal processing channel additionally contains a block for calculating the final likelihood ratio functional (BFOP) of the current combination of resolving allowed signals, the output of which is the output of the signal processing channel, and the corresponding signal inputs of the BFOP are connected to the corresponding signals the total outputs of the summing blocks, which are parallel connected to the corresponding inputs of the unit for calculating the posterior probabilities of the combinations of numbers of Gaussian components of the allowed signals, the outputs of which are respectively connected to the inputs of the extrapolation unit, whose signal outputs are connected to the corresponding inputs of the calculation unit of the final likelihood ratio functional and the corresponding inputs of the a posteriori calculation unit probabilities of combinations of numbers of Gaussian components of the allowed signal s.

Figure 1 shows the structural diagram of a device for resolving signals against arbitrary interference; figure 2 is a structural diagram of a block of correlation signal processing, which is part of each channel signal processing; figure 3 is a block diagram of a unit for calculating the final likelihood ratio functional.

The device for resolving signals against arbitrary interference, contains channels (1 ₁ -1 _n ) for signal processing, a clock generator 2, the output of which is connected to the clock inputs of the signal processing channels, the outputs of each of which are connected to the corresponding inputs of the comparison unit 3, the outputs of which are outputs device, the information input of which is connected to the information inputs of signal processing channels, and the control inputs of signal processing channels are connected to the control input of the device, and each of the channels is processed heel signals comprises a correlation processing unit 4 and the first signal blocks 5 ₀ -5 _L summation, the number of signal processing channels equal to a predetermined number of resolvable signals overlay combinations, each signal processing channel further comprises a calculation unit 6 functional outcome likelihood ratio of the current combination overlay enable signal the output of which is the output of the signal processing channel, and the corresponding signal inputs of the calculation unit of the final likelihood ratio functional I of the current combination of overlapping allowed signals are connected to the corresponding signal outputs of the summing blocks, which are parallel connected to the corresponding inputs of the block 7 for calculating the posterior probabilities of the combinations of numbers of the Gaussian components of the allowed signals, whose outputs are respectively connected to the inputs of the extrapolation block 8, whose signal outputs are connected to the corresponding inputs of the block calculating the final likelihood ratio functional and the corresponding inputs of the computation unit posterior probabilities of combinations of numbers of the Gaussian components of the allowed signals. The system that provides power to the units is not shown in the drawing.

Block 4 of the correlation signal processing, which is part of each channel of the signal processing, shown in figure 2, contains blocks 9 ₁ -9 _N calculation of private Gaussian likelihood ratio functionals for each component of interference and white Gaussian noise and blocks 10 ₀₁ -10 _NL calculation of private Gaussian likelihood ratio functionals for all combinations of superposition of the components of the allowed signals present in the mixture and the interference component, as well as white Gaussian noise, the information inputs of which are connected to the information the input input of the corresponding signal processing channel, and the information outputs of blocks 9 ₁ -9 _N and 10 ₀₁ -10 _{NL are} connected to the information inputs of the corresponding summing blocks 5 ₀ -5 _L.

Block 6 for calculating the final likelihood ratio functional of the current combination of resolving allowed signals, which is part of each signal processing channel shown in Fig. 3, contains the first multiplication blocks 11 ₁ -11 _L , the first information inputs of which are connected to the corresponding outputs of blocks 5 ₁ -5 _L summation, and the second information inputs of blocks 11 ₁ -11 _L multiplication connected to the corresponding outputs of block 8 extrapolation, and the corresponding outputs of blocks 11 ₁ -11 _L multiplication connected to the corresponding inputs of W the second summing block 12, the output of which is connected to the second input of the division block 13, the first input of which is connected to the output of the summing block 5 ₀ , while the output of the division block is connected to the first input of the second multiplication block 14, the output of which is the output of the signal processing channel and the input of the block 15 delay, the output of which is connected to the second input of the multiplication unit 14.

The first blocks 11 ₁ -11 _{L of the} multiplication, the second summing unit 12, the division unit 13, the second multiplying unit 14 and the delay unit 15 can be performed according to standard schemes published in the literature.

A device for resolving signals against a background of arbitrary interference works as follows.

(всего J типов сигналов), импульсной помехи и белого гауссовского шума в виде m - отсчетов n-го элемента поступает на информационные входы блоков 4 корреляционной обработки сигнала всех каналов обработки сигналов. Задача полного разрешения сигналов формулируется следующим образом: предполагается, что во входном колебании в произвольный момент времени может присутствовать произвольное число сигналов из некоторого набора; по результатам наблюдения в присутствии шумовых и импульсных помех с учетом априорной информации о параметрах сигналов требуется определить, сколько и в какой комбинации из всех возможных сигналов присутствуют на входе устройства. Сигналы разных типов представляют собой дискретные многоэлементные сигналы (МЭС) - последовательности К элементарных сигналов:The input signal, which is the sum of r signals of different types

(total J types of signals), impulse noise and white Gaussian noise in the form of m - samples of the nth element is fed to the information inputs of signal correlation processing units 4 of all signal processing channels. The problem of full resolution of the signals is formulated as follows: it is assumed that an arbitrary number of signals from a certain set may be present in the input oscillation at an arbitrary moment of time; according to the results of observation in the presence of noise and impulse noise, taking into account a priori information about the parameters of the signals, it is necessary to determine how many and in what combination of all possible signals are present at the input of the device. Signals of different types are discrete multi-element signals (MES) - sequence K of elementary signals:

where I is the number of types of elementary signals, J is the number of signals in the ensemble (the number of types of allowed signals) specified by Markov-mixed poly-Gaussian (MS-PG) models:

и

- элементы матрицы переходных вероятностей

и вектора начальных вероятностей

.Where

and

- elements of the transition probability matrix

and initial probability vectors

.

может содержать сумму случайного числа r сигналов

Кроме того, во входной смеси присутствует белый гауссовский шум n_ш(t) и ординарный поток независимых импульсных помех n_п(t) с вероятностью Р_п накладывающихся на k-ю временную позицию МЭС. При условии наложения импульсной помехи на k-ю временную позицию ее распределение имеет следующий вид:Signals can appear on the observation interval (0, T) in random different combinations. Therefore input fluctuation

may contain the sum of a random number r of signals

In addition, the input mixture contains white Gaussian noise n _w (t) and an ordinary flux of independent impulse noise n _p (t) with probability R _p overlapping the k-th time position of the MES. Provided that impulse noise is applied to the kth temporary position, its distribution has the following form:

- вероятность

-й компоненты импульсной помехи.Where

- probability

th components of impulse noise.

The joint distribution of interference and noise at the k-th position has the form:

также представляется МС-ПГ моделью следующего вида:When superimposing r signals with numbers j ₁ , j ₂ , ..., j _r , due to the invariance of Gaussian distributions with respect to linear transformations, the vector observed at the kth time position

also appears to be an MS-PG model of the following form:

- набор типов элементарных сигналов на k-й временной позиции, соответствующих набору сигналов в конкретной комбинации наложения сигналов из ансамбля j₁,j₂,...,j_r. Для простоты рассуждений введем понятие гипотезы

, соответствующей этой комбинации, причем число гипотез

. Нижние индексы в блоках 5₀-5_L, 9₁-9_N и 10₀₁-10_NL, 11₁-11_L представляют собой натуральный ряд чисел, где N представляет собой число гауссовских компонент в полигауссовом разложении импульсной помехи, a L определяется как число всевозможных сочетаний номеров гауссовых компонент в полигауссовом разложении разрешаемых сигналов, присутствующих в текущей гипотезеWhere

- a set of types of elementary signals at the kth time position corresponding to a set of signals in a particular combination of the superposition of signals from the ensemble j ₁ , j ₂ , ..., j _r . For simplicity, we introduce the concept of a hypothesis

corresponding to this combination, and the number of hypotheses

. The subscripts in blocks 5 ₀ -5 _L , 9 ₁ -9 _N and 10 ₀₁ -10 _NL , 11 ₁ -11 _L are a natural series of numbers, where N is the number of Gaussian components in the poly-Gaussian expansion of impulse noise, a L is defined as the number of all possible combinations of the numbers of the Gaussian components in the poly-Gaussian decomposition of the allowed signals present in the current hypothesis

, а набор комбинаций сигналов, соответствующий гипотезам имеет видIn particular, for the case corresponding to J = 2 (the number of signals in the ensemble), R = 3 (the maximum number of signals from the ensemble in the input oscillation), the number of hypotheses

, and the set of signal combinations corresponding to hypotheses has the form

those. the hypothesis h = 1 corresponds to a combination of the complete absence of signals, h = 2 corresponds to a combination when only a type 1 signal is present (from the ensemble), h = 8 corresponds to a combination when there is one type 1 signal and one type 2 signal (from the ensemble), etc. Thus, the set of signals corresponds to hypothesis h

The Gaussian components N {•} in expression (4) are determined by convolutions of the distributions corresponding to the Gaussian densities, linear combinations of which represent the poly-Gaussian signal and noise densities in (1).

We transform expression (4) to the form

Where

соответствующего комбинации наложения сигналов ансамбля

(или, по-другому гипотезе h).- respectively, the mathematical expectation vector and the covariance matrix of the GH distribution of the vector

the appropriate combination of overlay ensemble signals

(or, alternatively, hypothesis h).

и вектора набора номеров, реализовавшихся компонент ансамбля сигналовBased on the analysis of the joint distribution of the sample vector observed at the kth step (kth time position)

and the dialing vector of the realized components of the ensemble of signals

taking into account the property of Markov property, we can write an expression for joint posterior distributions of components and signal numbers:

представляет собой элемент матрицы переходных вероятностей МС-ПГ распределения комбинации сигналов

, которая определяет зависимость вектора номеров компонент ансамбля сигналов, реализовавшегося на k-й временной позиции от вектора номеров, реализовавшегося на (k-1)-й временной позиции (так как вектор номеров, реализующихся в процессе приема гауссовских компонент

по позициям

, образует также дискретную однородную цепь Маркова, которая характеризуется вышеуказанной матрицей переходных вероятностей и вектором начальных вероятностей):In the expression (9), the component

represents an element of the matrix of transition probabilities MS-GH distribution of a combination of signals

, which determines the dependence of the number vector of the components of the ensemble of signals realized at the k-th time position on the vector of numbers realized at the (k-1) -th time position (since the vector of numbers realized in the process of receiving Gaussian components

by position

also forms a discrete homogeneous Markov chain, which is characterized by the above matrix of transition probabilities and the vector of initial probabilities):

moreover, due to the statistical independence of the signals, the transition probability matrix and the initial probability vector are determined through similar characteristics of the ensemble signals as follows:

The optimal algorithm for resolving MS-GH signals against a background of a complex of interference based on the Bayesian criterion with the same fees for errors is:

,

сведем определяются следующими выражениями:The resulting likelihood relationship functionals

,

will be determined by the following expressions:

представляет собой v разрядный двоичный код и поступает на вход устройства одновременно с сигналом стробирования

. Функционал плотности распределения произвольной импульсной помехи представляется выражением (2), где N_п - число гауссовских случайных процессов как с одинаковыми ковариационными матрицами

и различными средними

.Each input sample

represents a v bit binary code and enters the device input simultaneously with the strobe signal

. The distribution density functional of an arbitrary impulse noise is represented by the expression (2), where N _p is the number of Gaussian random processes with identical covariance matrices

and different averages

.

определяются следующим образом: (19а)Particular Gaussian Likelihood Functionals

defined as follows: (19a)

In these expressions, we introduce the following notation:

вычисляющих взвешенные суммы выборочных значений входного процесса, взаимной энергии сигналов

соответствующих

гауссовским компонентам импульсной помехи и

сигналов и

энергий входной реализации. Подставляя принятые обозначения, получим:which make sense of correlation integrals

calculating the weighted sums of sample values of the input process, the mutual energy of the signals

relevant

Gaussian components of impulse noise and

signals and

energies of the input implementation. Substituting the accepted notation, we obtain:

формируют на своих выходах сигналы, пропорциональные частным гауссовским функционалам отношения правдоподобия, входящих в выражения (18)-(19). Эти значения с соответствующих выходов каждого блока 4 корреляционной обработки сигнала поступают одновременно на информационные входы блоков 5₀-5_L суммирования, где формируются сигналы, пропорциональные линейной комбинации частных гауссовских функционалов отношения правдоподобия по всем помеховым компонентам.Block 4 of the correlation signal processing of each channel, consists of blocks 9 ₁ -9 _N , 10 ₀₁ -10 _{NL for} calculating the private Gaussian likelihood ratio functionals, which, after receiving the input signal

generate signals at their outputs that are proportional to the particular Gaussian functionals of the likelihood ratio included in expressions (18) - (19). These values from the corresponding outputs of each block 4 of the correlation signal processing are sent simultaneously to the information inputs of the blocks 5 ₀ -5 _L summation, where signals are generated that are proportional to the linear combination of the private Gaussian likelihood ratio functionals for all interfering components.

определяемый выражением (18), где

- 1-я комбинация номеров компонент разрешаемых сигналов в рамках текущей гипотезы.In this case, at the output of block 5 ₁ summation, a signal of the form

defined by expression (18), where

- The 1st combination of the numbers of the components of the allowed signals within the framework of the current hypothesis.

The signals from the output of the blocks 5 ₀ -5 _L summation are fed simultaneously to the information inputs of the block 6 for calculating the final functional of the likelihood ratio and the block 7 for calculating the posterior probabilities of the combinations of components of the allowed signals. At the output of the a posteriori probability calculation block 7, signals are generated determined by expression (17), which are then fed to the corresponding inputs of the extrapolation block 8, where the weight estimates of the combinations of Gaussian components of the allowed signals are calculated in accordance with expression (16), which, in turn, are fed to the corresponding inputs of the calculation unit 6 of the final likelihood ratio functional. Block 6 of calculating the total likelihood ratio functional forms the final likelihood ratio functionals as follows. The signals received from the outputs of blocks 5 ₁ -5 _L summation are multiplied with the corresponding signals received from the outputs of block 8 extrapolation in blocks 11 ₁ -11 _L multiplication, then fed to the inputs of block 12 summation, where the sum included in the expression (15 ), after which, in the division block 13, the value received from the output of the summing block 12 is divided between the value received from the output of the summing block 5 ₀ . The obtained value goes to the input of the multiplication block 14, the output of which is connected to the delay block 15, the output of which, in turn, is connected to the second input of the multiplication block 14, where the final likelihood ratio functional is calculated in strict accordance with expression (15). The received signal from the output of the block 6 for calculating the final likelihood ratio functional of each signal processing channel is fed to the inputs of the comparison block 3, where by choosing the maximum value in accordance with expression (14) and deciding in favor of a certain hypothesis about the combination of superimposed allowed signals. As a result, after n cycles of operation of the device at the outputs S ₁ , S ₂ , ..., S _n there is information about the receipt of a certain combination of signals at the input of the device. Thus, in comparison with the prototype, which determines the presence of an appropriate signal from the distinguished ones at the input, the present invention allows one to determine how many and which of the possible signals are present at the input.

Claims (1)

A device for resolving signals against arbitrary interference, containing signal processing channels, a clock generator whose output is connected to the clock inputs of signal processing channels, the outputs of each of which are connected to the corresponding inputs of the comparison unit, the outputs of which are the outputs of the device, the information input of which is connected to information inputs signal processing channels, and the control inputs of the signal processing channels are connected to the control input of the device, and each channel of the signal processing c contains a signal correlation processing unit and summing units, characterized in that the number of signal processing channels is equal to a predetermined number of allowed signal superposition combinations, each signal processing channel further comprises a final function likelihood ratio calculation functional of the current allowed signal superposition combination, the output of which is an output of the processing channel signals, and the outputs of the correlation signal processing unit are connected to the inputs of the respective summing units, and the corresponding signal inputs of the calculation block of the final likelihood ratio functional of the current combination of resolving signals are connected to the corresponding signal outputs of the summing blocks, which are parallel connected to the corresponding inputs of the calculation of the posterior probabilities of the combinations of numbers of the Gaussian components of the allowed signals, the outputs of which are respectively connected to the inputs of the extrapolation block, the signal outputs which are connected to the corresponding inputs of the calculation unit AkzoNobel functional likelihood ratio corresponding inputs and calculating a posteriori probabilities of combinations of numbers permitted for Gaussian component signals.

Images