RU2582874C1 - Adaptive computer for interference rejection - Google Patents

Abstract

FIELD: computer engineering.

SUBSTANCE: invention relates to computer engineering and can be applied in automated coherent-pulse systems for separation of signals from movable targets on passive interference background. Adaptive computer for rejecting interference comprises auto compensator, first and second delay units, main and additional units for measuring correlation coefficient, unit for calculating weight coefficients, main and additional weight units, main adder, clock generator, digital delay line.

EFFECT: technical result is implementation of rejecting passive interference with a priori unknown spectrally correlation properties when selecting signals of movable targets.

2 cl, 18 dwg

Description

The invention relates to computer technology and can be used in automated coherent-pulse systems to isolate signals of moving targets against a background of passive interference with a priori unknown spectral-correlation properties during wobble of the repetition period of probe pulses.

A digital device for suppressing passive interference [1] is known, which contains two channels, each of which contains three main multipliers, an adder, the first and second memory blocks, the first and second additional multipliers, as well as a phase measurement unit, a functional converter, a computational unit, and a unit measuring the correlation coefficient and the unit for calculating the weight coefficient. However, this device has a low efficiency of signal extraction of moving targets due to the narrowing of the interference delay band during the wobble of the repetition period, which is a consequence of the inability to adapt the weight coefficients of the device to non-stationary time intervals within the wobble period.

Another known device is the correlation auto-compensator [2], which contains a number of delay units, two multipliers, an adder and a unit for estimating the parameters of the correlated noise. This device also has a number of drawbacks, among which the most significant are: low efficiency of signal extraction of moving targets due to narrowing of the passive jamming delay band during the wobble of the repetition period and poor suppression of extended interference edges, which is a consequence of the large time constant of the adaptive feedback circuit.

Closest to the proposed technical solution is the digital adaptive calculator for interference rejection selected as a prototype [3], which contains an autocompensator for the Doppler phase of passive interference, the first and second delay units for the repetition period, the main unit for measuring the correlation coefficient, the unit for calculating weight coefficients, the main weight block, main adder and sync generator. However, this device does not provide sufficient efficiency for distinguishing signals of moving targets against passive interference with a priori unknown spectral-correlation properties due to the narrowing of the passive interference delay band during the wobble of the repetition period and the lack of adaptation of the filter weight coefficients to the interference properties during non-stationary time intervals.

The problem to be solved in the present invention is to increase the efficiency of rejecting passive interference and isolating signals of moving targets during a wobble of the repetition period against a background of passive interference with a priori unknown spectral correlation properties.

To solve the problem, an adaptive calculator for interference rejection, comprising an auto-compensator, first and second delay units, a main unit for measuring the correlation coefficient, a unit for calculating weight coefficients, a main weight unit, a main adder and a clock generator, while the outputs of the auto-compensator are connected to the inputs of the same unit delays, the first inputs of the main unit for measuring the correlation coefficient and the first inputs of the main adder; the outputs of the first delay unit are connected to the inputs of the second delay unit of the same name, the second inputs of the main unit for measuring the correlation coefficient and the first inputs of the main weight unit, the outputs of which are connected to the same inputs of the main adder; the output of the main unit for measuring the correlation coefficient is connected to the first input of the unit for calculating the weight coefficients, the first output of which is connected to the second input of the main weight unit; the output of the sync generator is connected to the sync inputs of the auto-compensator, the first and second delay units, the main unit for measuring the correlation coefficient, the unit for calculating the weight coefficients, the main weight unit and the main adder; introduced an additional unit for measuring the correlation coefficient, a digital delay line and an additional weight unit, while the first inputs of the additional unit for measuring the correlation coefficient are connected to the first inputs of the same name for the main unit for measuring the correlation coefficient; the second inputs of the additional unit for measuring the correlation coefficient are connected to the same outputs of the second delay unit and the first inputs of the additional weight unit, the outputs of which are connected to the same third inputs of the main adder; the input of the digital delay line is connected to the output of the main unit for measuring the correlation coefficient; the output of the digital delay line is connected to the second input of the weighting unit, the third input of which is connected to the output of the additional unit for measuring the correlation coefficient, and the second output is connected to the second input of the additional weighting unit; the sync inputs of an additional block for measuring the correlation coefficient, a digital delay line and an additional weight block are connected to the output of the sync generator; moreover, the inputs of the adaptive calculator for rejecting interference are the inputs of the auto-compensator, and the outputs are the outputs of the main adder.

Additional blocks introduced into the proposed device, and their connection with other blocks of the calculator allow adaptation to the spectral-correlation properties of the interference, taking into account the wobble of the repetition period, which is a necessary and sufficient condition for its effective rejection.

Comparison with technical solutions known from published sources of information shows that the claimed technical solution contains a new set of distinctive features and, therefore, has novelty and an inventive step.

The claimed solution is technical in nature, feasible, reproducible and, therefore, is industrially applicable.

The essence of the proposed technical solution is illustrated by the relevant drawings and graphs.

In FIG. 1 is a structural electrical diagram of an adaptive calculator for interference rejection (AVRPP); in FIG. 2 - a private embodiment of the implementation of the AUDP; in FIG. 3 - auto-compensator; in FIG. 4 - delay unit; in FIG. 5 - main and additional units for measuring the correlation coefficient; in FIG. 6 - weight block; in FIG. 7 - the main adder; in FIG. 8 - additional adder; in FIG. 9 - block retention; in FIG. 10 - phase measurement unit; in FIG. 11 - complex multiplier; in FIG. 12 - block complex conjugation; in FIG. 13 - integrated drive; in FIG. 14 - block combining quadratures; FIG. 15 - drive; in FIG. 16 - complex adder; in FIG. 17 - integrated inverter; in FIG. 18 shows the dependencies of the gain in the signal-to-noise ratio averaged over the Doppler phase of the signal of the proposed adaptive calculator in comparison with the prototype.

The adaptive calculator for interference rejection (Fig. 1) contains an auto-compensator 1, first 2 and second 3 delay units, a main unit 4 for measuring the correlation coefficient, a unit 5 for calculating weight coefficients, a main weight unit 6, a main adder 7, a clock generator 8, an additional unit 9 measuring the correlation coefficient, a digital delay line 10 and an additional weight unit 11; while the outputs of the auto-compensator 1 are connected to the inputs of the first delay unit 2, the first inputs of the main unit 4 for measuring the correlation coefficient and the first inputs of the main adder 7, the outputs of the first delay unit 2 are connected to the inputs of the second delay unit 3, the second inputs of the main unit 4 for measuring the coefficient correlation and the first inputs of the main weight unit 6, the outputs of which are connected to the second inputs of the main adder 7, the output of the main unit 4 for measuring the correlation coefficient is connected the first input of the weight coefficient calculation unit 5, the first output of which is connected to the second input of the main weight unit 6, the first inputs of the additional correlation coefficient measurement unit 9 are connected to the first inputs of the correlation coefficient measurement unit 4 of the same name, the second inputs of the additional correlation coefficient measurement unit 9 are connected to the same outputs of the second delay unit 3 and the first inputs of the additional weight unit 11, the outputs of which are connected to the third inputs of the main bag ator 7, the input of the digital delay line 10 is connected to the output of the main unit 4 for measuring the correlation coefficient, the output of the digital delay line 10 is connected to the second input of the unit 5 for calculating the weight coefficients, the third input of which is connected to the output of the additional unit 9 for measuring the correlation coefficient, and the second output with the second input of the additional weight unit 11, the output of the clock 8 is connected to the clock inputs of the auto-compensator 1, the first 2 and second 3 delay blocks, the main unit 4 for measuring the correlation coefficient, block 5 compute weighting coefficients weighting the primary unit 6, the main adder 7, an additional unit 9, the correlation coefficient measurement, the digital delay line 10 and additional weighting block 11.

A particular embodiment of an adaptive calculator for interference rejection (FIG. 2) is characterized in that in the ATSD in FIG. 1, a third delay unit 12 and an additional adder 13 are additionally introduced, while the inputs of the third delay unit 12 are connected to the same outputs of the auto-compensator 1 and the first inputs of the additional adder 13, the second inputs of which are connected to the outputs of the third delay unit 12, and the outputs to the inputs of the same name delay unit 2, the first inputs of the main unit 4 for measuring the correlation coefficient, the first inputs of the additional unit 9 for measuring the correlation coefficient and the first inputs of the main adder 7, sync output generator 8 is connected to the sync inputs of the third delay unit 12 and the additional adder 13.

The auto-compensator 1 (Fig. 3) comprises a delay unit 14, a first and second complex multiplier 15, a phase measurement unit 16, and a first and second delay unit 17; block 2, 3, 12, 17 delay (Fig. 4) contains two random access memory (RAM) 18; the main 4 and additional 9 blocks for measuring the correlation coefficient (Fig. 5) comprise a complex conjugation block 19, a complex multiplier 20, a complex storage 21, three quadrature combining units 22, two storage 23, two dividers 24 and a square root extraction unit 25; the weight unit 6, 11 (Fig. 6) contains two multipliers 26; the main adder 7 (Fig. 7) contains two complex adders 27; additional adder 13 (Fig. 8) contains a complex inverter 28 and a complex adder 29; block 14 detention (Fig. 9) contains two RAM 30; the phase measuring unit 16 (Fig. 10) comprises a complex conjugation unit 31, a complex multiplier 32, a complex drive 33, a quadrature combining unit 34, a square root extracting unit 35, and two dividers 36; complex multiplier 15, 20, 32 (Fig. 11) contains two channels (I and II), each of which consists of two multipliers 37 and the adder 38; block 19, 31 complex pairing (Fig. 12) contains an inverter 39; complex drive 21, 33 (Fig. 13) contains two drives 40; block 22, 34 combining quadratures (Fig. 14) contains two multipliers 41 and the adder 42; the drive 23, 40 (Fig. 15) contains a channel consisting of n delay elements 43 for the interval t _d and n adders 44; complex adder 27, 29 (Fig. 16) contains two adders 45; complex inverter 28 (Fig. 17) contains two inverters 46.

Adaptive calculator for rejection interference works as follows.

A pack of coherent radio pulses, consisting of passive interference significantly exceeding the signal from the target, is fed to the input of the receiving device, in which it is amplified, is transferred to the video frequency in quadrature phase detectors, and then subjected to analog-to-digital conversion (the corresponding blocks in Fig. 1 are not shown) .

The readings are received at time instants separated by p non-equidistant or unequal time intervals T ₁ , T ₂ , ..., T _i , ..., T _p , and form the core of the wobble, repeating with a constant period of wobble:

,

,

where p is the number of repetition periods in the core of the wobble.

Digital codes (x _kl , y _kl ) of both quadrature projections following through non-equidistant intervals T ₁ , T ₂ , ..., in each range resolution element (range ring) of each repetition period form a sequence of complex numbers

,

,

where k is the number of the current period, l is the number of the current range ring, θ _kl is the current phase (usually interference, due to its significant excess over the signal), and

,

,

where φ _0l is the initial phase; φ _jl is the Doppler phase shift of the interference over the period T _j equal to φ _jl = 2πf _jl T _j , here f _jl is the Doppler frequency of the interference.

.Digital readings in the inventive calculator (Fig. 1) are fed to the inputs of the auto-compensator (AK) 1, in which adaptive compensation is performed directly for the Doppler shift of the interference spectrum. To realize this in the time domain, the total Doppler phase shift of the interference phase is measured over the incoming number of periods. This uses the current data of two adjacent repetition periods T _k-1 and T _k coming from n + 1 adjacent range resolution elements and forming a training sample

.

The block diagram of AK is shown in FIG. 3. In the phase measurement block 16 (Fig. 10), according to the input samples U _kl and the samples U _{k-1, l} delayed in the first delay block 17 _, estimates of the Doppler phase shift of the interference phase for the kth repetition period are calculated (k = 1, 2, ...) for each l-th element of range resolution (l = 1, 2, ...). At the same time, in block 31 of complex conjugation using inverter 39 (Fig. 12), the sign of imaginary projections is inverted. In the complex multiplier 32, the multiplication of the corresponding complex numbers occurs, realized by operations with the projections of these numbers in accordance with FIG. 11. Formed values

,

,

с n+1 смежных элементов разрешения по дальности временного строба, кроме элемента с номером j=l, для чего выходные величины элемента 43 задержки с номером n/2 поступают только на последующий элемент 43 задержки (фиг. 15). При этом на выходах накопителя 33 (фиг. 13) образуются величиныenter the complex drive 33 (Fig. 13), consisting of drives 40 (Fig. 15), which, using delay elements 43 and adders 44, summarize the products sliding along the range in each repetition period

with n + 1 adjacent resolution elements in the range of the time strobe, except for the element with the number j = l, for which the output values of the delay element 43 with the number n / 2 go only to the subsequent delay element 43 (Fig. 15). Thus at the outputs of the drive 33 (Fig. 13) values are formed

- оценка сдвига фазы помехи за период T_k для l-го элемента разрешения по дальности, усредненная по n смежным элементам разрешения; n - объем обучающей выборки.Where

- an estimate of the phase shift of the interference for the period T _k for the l-th element of the range resolution averaged over n adjacent resolution elements; n is the size of the training sample.

, в блоке 35 извлечения квадратного корня - величины

, а затем на выходах делителей 36 (фиг. 10) - величины

, поступающие на первые входы второго комплексного перемножителя 15 (фиг. 3). В результате их перемножения с выходными отсчетами второго блока 17 задержки образуются величиныIn block 34 combining quadratures (Fig. 14) are determined by

, in block 35 extraction of the square root - values

, and then at the outputs of the dividers 36 (Fig. 10) - values

entering the first inputs of the second complex multiplier 15 (Fig. 3). As a result of their multiplication with the output samples of the second delay unit 17, values are formed

,

,

, задержанными блоком 14 задерживания (фиг. 3) с целью временного согласования вводимых и компенсируемых фазовых сдвигов на интервал т, равный задержке оценок по отношению к среднему элементу обучающей выборки.In the first complex multiplier 15 (Fig. 3), these values are multiplied with the original samples

delayed by the delay unit 14 (Fig. 3) in order to temporarily coordinate the introduced and compensated phase shifts by the interval m equal to the delay of the estimates with respect to the middle element of the training sample.

The value of the interval τ is determined by the expression

where t _at is the calculation time of the estimation of the phase of the interference, n is the number of elements of the training sample, t _d is the interval (period) of time sampling.

с точностью до погрешностей измерения оценки

не содержат доплеровского сдвига фазы помехи, что позволяет осуществлять последующее режектирование помехи вычислителем (фильтром) с действительными весовыми коэффициентами.Formed at the output of the auto-compensator 1 (Fig. 1, 3) values

accurate to within measurement error

do not contain a Doppler phase shift of the interference, which allows subsequent rejection of the interference by the computer (filter) with actual weight coefficients.

позволяет адаптироваться к аргументу реальной корреляционной функции помехи, что является необходимым условием ее эффективного режектирования.Using current ratings

allows you to adapt to the argument of the real correlation function of the interference, which is a necessary condition for its effective notching.

.Basically 6 and an additional 11 weight blocks (Fig. 1), the projections are scalarly multiplied by the weighting coefficients g ₁ and g ₂ (Fig. 6), which come respectively from the first and second outputs of the weighting coefficient calculation unit 5 (Fig. 1). In the main adder 7 (Fig. 1) there is a separate summation (Fig. 7, 16) of the same name projections of the weighted sequence of processed samples and the formation of the output value of the calculator

.

When choosing weighting coefficients according to the adaptive algorithms of the prototype calculator [3], the limit possibilities of isolating the signal against the background of incoming interference with a constant repetition period are realized. However, when these algorithms are used to calculate weighting coefficients during the wobble of the repetition period, the effect of reducing the noise suppression coefficient takes place, the greater the higher the index (depth) of the wobble. In the case of a wobble of the repetition period, the filter weights must be calculated according to new adaptive algorithms, in particular for a second-order calculator (filter) (m = 2), having the form

,

и

^_ оценки коэффициентов межпериодной корреляции помехи.Where

,

and

^_ estimates of inter-period correlation coefficients of interference.

вычисляется в основном блоке 4 измерения коэффициента корреляции (фиг. 1, 5). Цифровая линия 10 задержки оценки

позволяет получить оценку

где k - номер текущего периода. Оценка р₁₃ вычисляется в дополнительном блоке 9 измерения коэффициента корреляции (фиг. 1, 5). Блоки 4, 9 измерения коэффициента корреляции выполняются в соответствии с фиг. 5 и реализуют алгоритм оцениванияRating

calculated in the main unit 4 of the measurement of the correlation coefficient (Fig. 1, 5). Digital line 10 evaluation delay

allows you to get an estimate

where k is the number of the current period. The score p _{13 is} calculated in the additional unit 9 for measuring the correlation coefficient (Fig. 1, 5). The correlation coefficient measurement units 4, 9 are performed in accordance with FIG. 5 and implement the estimation algorithm

.

.

,

и

блок 5 вычисления весовых коэффициентов (фиг. 1) реализует алгоритмы (3). Данный блок представляет собой арифметико-логическое устройство или сигнальный процессор.Based on the obtained estimates of the inter-period interference correlation coefficients

,

and

block 5 calculation of weighting coefficients (Fig. 1) implements algorithms (3). This unit is an arithmetic logic device or signal processor.

Each of the delay blocks 2, 3, 12, 17 (Figs. 1, 2, 4) consists of RAM 18 connected in parallel. Moreover, each RAM 18 serves to store the values of the samples from the range rings of one quadrature channel for a period. Moreover, each RAM 18 delay blocks 2, 3, 12, 17 is used to store only N first samples in each repetition period, where N = T _min / f _d is the number of range rings corresponding to the minimum repetition period T _min = min (T ₁ , T ₂ , ...).

The delay unit 14 (Fig. 3, 9) delays the input samples in real time by the interval τ determined from expression (2) and is equal to the delay of the estimates with respect to the middle element of the training sample excluded in the drive 40 (Fig. 13, 15 ) in accordance with the expression (1). Then, in the case of a signal commensurate in magnitude with the interference, or discontinuous interference during the subsequent rejection of the interference samples from the resolution element containing the signal, the possibility of attenuation or suppression of the signal due to its influence on the estimates used is excluded. In this case, the RAM 30 are used for the "sliding" storage of τ / t _d samples.

The synchronization of the adaptive calculator for noise rejection is carried out by applying to all the blocks of the calculator a sequence of synchronizing pulses generated by the synchro-generator 8 (Fig. 1), controlled by the pulses of the radar synchronizer (not shown in Fig. 1), following each at intervals T ₁ , T ₂ , .... The repetition period of the synchronizing pulses is equal to the time sampling interval t _d selected from the condition of the required range resolution.

The advantage of the proposed technical solution is, firstly, the ability to adapt to the argument and the modulus of the real correlation function of the interference, without resorting to approximating its shape, taking into account the wobble of the repetition period of the probe pulses and, secondly, the short duration of the adaptation process, which ends within the transition process in an adaptive calculator for rejecting interference.

In FIG. Figure 18 shows the dependences of the gain Δµ in the signal-to-noise ratio averaged over the Doppler phase of the signal of the proposed calculator compared to the prototype versus the wobble depth mod (in percent) for two values of the normalized noise spectrum width β = ΔfT _min (β = 0.05 - curve 1 and β = 0.1 - curve 2). The curves are plotted for the case of a double wobble of the repetition period (p = 2) and the training sample size n = 5.

Thus, an adaptive calculator for interference rejection increases the efficiency of rejecting passive interference and isolating the signals of moving targets during the wobble of the repetition period against the background of passive interference with a priori unknown spectral-correlation properties.

In FIG. 2 shows a particular embodiment of an adaptive calculator for rejecting interference. A third delay unit 12 and an additional adder 13 are introduced into it. It is known that these blocks are independently used to suppress passive interference. An additional inclusion of these blocks makes it possible to increase the efficiency of rejecting passive interference, to reduce the length of the bit grid of digital arithmetic devices (multipliers, dividers, and adders) in subsequent blocks of the adaptive calculator, without reducing the requirements for accuracy of calculations, or to increase the accuracy of calculation with the same length of the bit grid .

Bibliography

1. A.S. 809018 USSR, IPC G01S 7/36. Digital device for suppressing passive interference / D.I. Popov. - No. 2755228; declared 04/16/1979; publ. 02/28/1981, Bull. No. 8. - 5 sec.

2. Radio-electronic systems: fundamentals of construction and theory. Reference book / Ya.D. Shirman, S.T. Baghdasaryan, A.S. Malyarenko, D.I. Lekhovitsky [et al.]; edited by Y.D. Shirman. - 2nd ed., Revised. and add. - M .: Radio engineering, 2007; from. 439, fig. 25.22.

3. A.S. 1098399 USSR, IPC G01S 7/36. Device adaptive rejection of passive interference / D.I. Popov. - No. 3299959; declared 06/12/1981; publ. 12/20/1998, Bull. Number 35. - 16 p.

Claims (2)

1. An adaptive calculator for interference rejection, comprising an auto-compensator, first and second delay units, a main correlation coefficient measurement unit, a weight calculation unit, a main weight unit, a main adder and a clock generator, while the outputs of the auto-compensator are connected to the inputs of the first delay unit of the same name, the first the inputs of the main unit for measuring the correlation coefficient and the first inputs of the main adder, the outputs of the first delay unit are connected to the inputs of the second block of the same name by the second inputs of the main unit for measuring the correlation coefficient and the first inputs of the main weight unit, the outputs of which are connected to the same second inputs of the main adder, the output of the main unit for measuring the correlation coefficient is connected to the first input of the unit for calculating the weight coefficients, the first output of which is connected to the second input of the main weight unit, the output of the sync generator is connected to the sync inputs of the auto-compensator, the first and second delay units, the main unit for measuring the correl coefficient the unit, the calculation unit of the weight coefficients, the main weight unit and the main adder, characterized in that an additional unit for measuring the correlation coefficient, a digital delay line and an additional weight unit are introduced, while the first inputs of the additional unit for measuring the correlation coefficient are connected to the first inputs of the same unit for measuring correlation coefficient, the second inputs of the additional unit for measuring the correlation coefficient are connected to the outputs of the second delay unit and by the first inputs of the additional weight unit, the outputs of which are connected to the third inputs of the main adder of the same name, the input of the digital delay line is connected to the output of the main unit for measuring the correlation coefficient, the output of the digital delay line is connected to the second input of the unit for calculating the weight coefficients, the third input of which is connected to the output of the additional unit measuring the correlation coefficient, and the second output with the second input of the additional weight unit, the output of the sync generator is connected to the sync inputs of the additional tional correlation coefficient measurement unit, a digital delay line and additional weighting unit, and inputs for calculating an adaptive noise rezhektirovaniya automatic compensator are inputs, and the outputs - the main outputs of the adder. 2. The adaptive calculator for noise rejection according to claim 1, characterized in that a third delay unit and an additional adder are introduced into it, while the inputs of the third delay unit are connected to the outputs of the autocompensator and the first inputs of the additional adder, the second inputs of which are connected to the outputs of the same name the third delay unit, and the outputs with the same inputs of the first delay unit, the first inputs of the main unit for measuring the correlation coefficient, the first inputs of the additional unit for measuring the coefficient correlation coefficient and the first inputs of the main adder, the output of the clock is connected to the clock inputs of the third delay unit and the additional adder.

Images