RU2624005C1 - Method of processing super-wide-band signals - Google Patents

Abstract

FIELD: radio engineering, communication.

SUBSTANCE: received input signal is first multiplied by a reference signal matched to the transmitter signal to form two quadrature channels; then, in each quadrature channel of all N ranging channels, processing is performed matching the bursts of sub-impulses, forming a matrix of complex signals in the form of two quadrature components and further, starting from the matrix S, the phase distortion Δψ_ko in each sub-impulse is performed, and these corrections are added to the corresponding sub-impulses of all channels, which are then summed in each distance channel, forming the resulting N complex samples of the output signal.

EFFECT: compensating phase distortions of linear frequency modulation.

7 cl, 19 dwg

Description

The invention relates to methods for processing ultra-wideband signals (SSS) with linear frequency modulation (LFM) in radio and acoustic systems for location, navigation and communication, taking into account the distortions of these signals when they pass through the transceiver paths and the propagation channel from the transmitter to the receiver.

The characteristics of location, navigation and communication systems are largely determined by the width of the frequency spectrum of the signal used. The expansion of the spectrum width, the transition from narrow-band signals to complex broadband, and even more so the transition to the NW, when the signal spectrum width ΔF becomes comparable or even greater than the minimum frequency F _{min of} this spectrum, increases the resolution and accuracy of determining the coordinates of objects during location and navigation, the transmission speed information in communication systems, noise immunity of systems, secrecy of their work, etc. At the same time, the use of a long-duration signal (DB) not only improves noise immunity, stealth, etc., but also provides long-range systems.

. При этом, например при согласованной фильтрации в отсутствии доплеровского смещения используют фильтр, импульсная переходная функция которого в комплексном представлении

, где точка над переменной означает, что она комплексная, а знак (*) - комплексное сопряжение.In all location, navigation and communication systems, in order to increase the signal-to-noise ratio in the receiving devices, as a rule, processing of the input signal is used, which is consistent with the signal generated in the transmitter. This can be a coordinated filtering (see, for example, “Radar Handbook” edited by M. Skolnik, Moscow, Sov. Radio, 1976, volume 1, p. 8 and 111), correlation processing (see, for example, “Theoretical Foundations of Radar”, under the editorship of Ya. D. Shirman, publishing house “Sov. Radio”, Moscow, 1970, p. 100) or correlation-filtering processing (see, for example, Bryzgalov’s article A .P., Karaulovoj E.V., Hnykina A.V. “Analog-digital data processing in synthesized aperture radars using ultra-wideband signals with linear frequency mode “in the journal“ Digital Signal Processing ”No. 4, 2004). Moreover, in all cases, it is believed that the useful signal at the input of the matched processing is known and, up to parameters such as delay and Doppler frequency shift, corresponds to the signal generated in the transmitter

. In this case, for example, with matched filtering in the absence of Doppler shift, a filter is used, the pulse transition function of which in the complex representation

, where the dot over the variable means that it is complex, and the sign (*) is complex conjugation.

However, to obtain consistent processing, it is necessary not only to know the shape of the signal generated in the transmitter, but it is also necessary that the phase-frequency characteristics of the transceiver path and the entire signal propagation path in the signal band are linear. Only in this case, the waveform, its modulation does not change after the signal passes this path and can be incorporated into a matched filter. But the transceiver path (including antennas, amplifiers, filters, etc.) and the propagation channel have the indicated linearity over a limited frequency range. Typically, in practice, the linearity of the transceiver path can be ensured in the frequency interval not exceeding ± (10-15)% of the value of the carrier frequency of the signal.

In the case of nonlinearity, the phase-frequency characteristic of the path passed by the signal from the moment of its formation in the transmitter to the coordinated processing in the receiver, the signal at the output of such processing is distorted. As a rule, its amplitude decreases, the duration of the main lobe of the processing response to a single point reflector expands, the level of side lobes increases, etc. For example, in the book by C. Cook and M. Bernfeld “Radar Signals” (M. Publishing House “Soviet Radio” ”, 1971) shows the appearance of side lobes in the form of a“ paired echo ”. To account for signal distortions in the path, a transition to quasi-consistent adaptive processing is necessary.

To ensure the linearity of the path, in general, in addition to the linearity of the phase-frequency characteristic of the path passed by the signal, the constancy of its amplitude-frequency characteristic is required. Often this is not the case for UWB. In some cases, the nonlinearity of the amplitude-frequency characteristic has practically no effect on the waveform at the output of the matched filter and therefore may not be taken into account during processing. In some cases, the initial compensation of the inconstancy of the amplitude-frequency characteristic can be performed, including a priori data, in some cases, it may be necessary to adaptively adapt the processing to a priori unknown non-uniformity of the amplitude-frequency characteristic of the path in the NW band. However, in practically important cases, the signal distortion at the output of the matched processing for amplitude and phase distortions of the input LFM signal can be considered independent (see, in the above book by C. Cook and M. Bernfeld, p. 399), and therefore, methods for their compensation can be considered separately.

In any case, coordinated processing that does not take into account phase distortions of the signal can not only degrade the characteristics of the system using it, but also lead to its complete inefficiency.

As an analogue of the input signal processing system, taking into account the distortion of the signal during the passage of the path, we can consider a communication system with the simultaneous transmission of information and reference signals. It is assumed that the distortions of both signals during the passage of the path are the same. The received reference signal is used in correlation or correlation-filter quasi-consistent processing of the received information signal. For narrow-band signals, in order to reduce their mutual influence, these two signals, as a rule, are slightly spaced in frequency, but stored in a frequency band in which the distortions of both signals are the same. This is not feasible for the SSS DB, and special methods are required to ensure the reception of the information signal even when it overlaps in the frequency range with the reference signal. In particular, such a communication system can be performed by shifting one of the parameters of the reference signal relative to the information signal in accordance with the patent for the invention (Bryzgalov AP, Volkov PV “Method for transmitting and receiving information.” RF patent No. 2106066 and Eurasian patent No. 000732 with priority of 2/27/1996).

For the prototype of the invention adopted one of the closest to the claimed technical solution for the technical nature of the methods described in the previously mentioned article Bryzgalova A.P., Karaulova E.V. and Hnykina A.V. This is a method of coordinated correlation-filter processing of a coherent burst of LFM pulses in synthesized aperture radar (SAR). In accordance with this method, the received input signal is multiplied by the reference SSB (local oscillator signal) in the form of a packet of chirp pulses. In this case, the local oscillator signal is matched with the transmitter signal, and is tuned to certain parameters of this signal associated with the passage of the path signal, which in the considered prototype is determined by the relative positions and speeds of the aircraft (LA) - the SAR carrier and the located area. In particular, the local oscillator signal can be configured to receive a signal reflected from a point on the ground.

When multiplying, they form two quadrature channels (see, for example, the indicated book under the editorship of Ya.D. Shirman, p. 108). In each quadrature channel, the signals are digitized and then processed in N distance channels. Each range channel is tuned to a specific point on the ground. To do this, the signal after the multiplier is processed in x N channels, taking into account the difference in the signal parameters of each channel (delay and time transformation coefficient p = 1-2v _r / c, where v _r is the radial component of the PCA carrier velocity relative to the location, c is the speed of light) from the parameters of the local oscillator signal. At the same time, N samples of the output signal of the receiver are formed for the given terrain points, thereby forming a radar image (RLI) of the located terrain.

At the same time, it should be noted that the considered method of correlation-filter processing can be used to solve other problems of radar, as well as navigation and communication, and the number of pulses in a packet can be reduced to one pulse. In other tasks, processing parameters may be other parameters related, for example, to spatial, temporal, frequency, and other differences in the range channels.

As can be seen from the description of the prototype, two pulses: the input signal corresponding to a certain range channel, and the local oscillator pulse can be offset relative to each other and partially not overlap in time. Such a shift leads, in particular, to energy losses, reducing the effective duration of the probe pulses. For these losses to be insignificant, it is necessary that the displacement under consideration be significantly less than the pulse duration.

As already noted, such consistent processing works well in the absence of distortions of the signal generated in the transmitter, in particular due to the nonlinearity of the phase-frequency characteristic of the path passed by this signal. In the presence of distortions, these distortions are, as a rule, a priori unknown and, in the general case, require the use of adaptive methods.

An object of the invention is to develop a method for adaptive quasi-consistent processing of the BSS in the form of a single LFB DB pulse or a packet of such pulses, taking into account the distortions that this signal receives when it passes the entire path from generation in the transmitter to processing in the receiver due to the non-linearity of the phase-frequency characteristic of the path. The creation of such a method will solve one of the main problems associated with the use of SSH - the problem of distortion of SHS in the path due to the nonlinearity of this path.

, каждый импульс опорного сигнала разбивают на К_подымп узкополосных подымпульсов, и в каждом квадратурном канале всех N дальностных каналов осуществляют обработку, согласованную с К_подымп пачками из N_пач подымпульсов, формируя матрицу S комплексных сигналов

в виде двух квадратурных составляющихThe essence of the proposed method ultrawideband signal processing in a stack of N _Pace pulses with linear frequency modulation (LFM) for _Pace N ≥1, implemented using the filter-correlation method, according to which a received input signal is first multiplied by a reference signal matched with the signal formed in the transmitter, with the formation of two quadrature channels, in each of which the signals are digitized and then processed in N distance channels, forming N pairs of quadrature components of the complex processing output signal for specified processing parameters, such as a time delay of the signal, its frequency offset, etc., and then find the amplitudes of these signals, which determine N samples of the output module of the processing signal, while processing the input signal distorted due to nonlinearity of the phase-frequency characteristics of the path traveled by these signals from formation in the transmitter to processing at the receiver, with an interval of time distortion correlation

Each pulse of the reference signal is split into K _podymp narrowband podympulsov and in each quadrature channel of all N channels, the processing of range, matched with K _podymp batches of N _Pace podympulsov forming a matrix of complex signals S

in the form of two quadrature components

S _nkc , S _nks ,

where k = 1, ..., K _sub - _pulse is the serial number of the sub-pulse,

n = 1, ..., N-number of the ranging channel,

всех каналов, которые затем в каждом дальностном канале суммируют, формируя результирующие N комплексных выборок выходного сигнала.and further, proceeding from the matrix S, the phase distortions Δψ _ko are estimated in each subpulse and these corrections are introduced into the subpulse signals corresponding to the number k

all channels, which are then summed in each range channel, forming the resulting N complex samples of the output signal.

, но не менее чем в 5 раз больше интервала заданных значений задержек входных сигналов по времени, а отношение сигнал/шум при согласованной обработке пачки подымпульсов значительно (не менее чем в 5 раз) превышало 1, а при оцифровывании сигнала такт временного квантования выбирают таким, чтобы число выборок сигнала на один подымпульс было не меньше 10.Moreover, the duration of the subpulse τ _{subpump is} chosen so that it is not less than 5 times less

, but not less than 5 times the interval of the specified values of the input signal delays in time, and the signal-to-noise ratio during the coordinated processing of the sub-pulse train significantly (not less than 5 times) exceeded 1, and when digitizing the signal, the temporal quantization cycle is chosen as so that the number of signal samples per subpulse is not less than 10.

осуществляют оценку фазовых искажений Δψ_ko.Moreover, to estimate the phase distortions Δψ _ko, first, starting from the matrix S, the initial output signal is formed for all N channels and the Mth range channel is selected from the initial signal, the amplitude of the output signal of which is at least 5 times the amplitude of the output signals of other long-range channels that do not coincide with the M-th channel in time delay or are not allowed with it in any of the processing parameters, and further according to the signals

assess phase distortion Δψ _ko .

всех K_подымп подымпульсов, или сигналов

.In this case, the formation of the initial output signal is carried out by summing in each n-th range channel or signal modules

all K _sub- pulses of _sub- pulses, or signals

.

, оценивая фазу ψ_Mk по каждому подымпульсу как ψ_Mk=arctg(S_Mks/S_Mkc) и формируя вектор ψ_м=(ψ_М1, … ψ_{MKподымп}), и далее зависимость ψ_Mk от k аппроксимируют полиномом ψ_ko=а+bk+ck², где a, b и с находят по методу наименьшего квадратичного отклонения, исходя из минимизации значенияIt is advisable to evaluate the phase distortions Δψ _ko in the selected Mth channel based on the signals

, estimating the phase ψ _Mk for each subpulse as ψ _Mk = arctan (S _Mks / S _Mkc ) and forming the vector ψ _m = (ψ _M1 , ... ψ _{MK subnumber} ), and then approximate the dependence of ψ _Mk on k by the polynomial ψ _ko = а + bk + ck ² , where a, b, and c are found by the least square deviation method, based on minimizing the value

.

.

Next, phase corrections are determined to correct for the distortion of the NWS in each kth subpulse, based on the expression Δψ _ko = ψ _Mk -ψ _ko .

To simplify the processing, the change in ψ _{Mk with} respect to k is also approximated by the straight line ψ _ko = a + bk.

Despite some restrictions on the application of the proposed method, when using it, the task of adaptive quasi-consistent processing, taking into account phase distortions of the LFM SSH in the path of its distribution, is solved in many cases important for practice.

The list of drawings:

In FIG. Figure 1 shows the radar image in the color (brightness) palette in the XY coordinates when locating the SAR of a point reflector in the absence of other reflectors, noise, and signal distortion.

In FIG. 2 shows the same radar image as in FIG. 1, but already three-dimensional when viewed from the side.

In FIG. 3 and 4 show radar images for the same case as in FIG. 1 and FIG. 2, but already in the presence of phase distortion of the signal after passing through the path.

In FIG. 5 and FIG. Figure 6 shows the radar image for that case, but with distortion compensation.

In FIG. Figure 7 shows the radar image with a decrease in the number of subpulses from 15 to 10.

In FIG. 8, the case of coordinated processing when locating two reflectors is considered, when there are no distortions and no compensation is performed.

In FIG. 9 shows the results of the case as for FIG. 8, but when there is a phase distortion of the input signal in the form of a sine wave with a span of 60 degrees and a period of 0.5 τ _imp . No compensation.

In FIG. 10 shows the simulation results of the case that in FIG. 9, but in the presence of distortion and compensation.

In FIG. 11 ... 13 shows the simulation results under boundary conditions for range resolution.

In FIG. 14 ... 16 shows the simulation results at a distance between two reflectors less than the resolution in range.

In FIG. 17 ... 19 shows the results of modeling processing with equal ranges to two reflectors.

The technical implementation of the proposed method and its effectiveness can be explained by the example of applying it to a prototype - to spatio-temporal processing of a coherent burst of chirp pulses in PCA. The output of this processing is a radar image, consisting of many N individual image elements - pixels, corresponding to reflections from surface areas. Therefore, coordinated processing for each section should be carried out taking into account its parameters related to its position and speed relative to the SAR antenna with each sounding in the synthesis interval, i.e. Coordinated processing should be carried out in N channels, conditionally called range. In correlation-filter processing, this corresponds to its three stages.

At the first stage, the input signal is multiplied by the local oscillator signal with the formation of two quadrature channels. At the same time, the local oscillator signal corresponds in form to the probing signal, and its parameters are selected based on the location problem, for example, from a given location for location. This is an analog processing step. It ends with band-pass filters that separate the differential frequency signals, and analog-to-digital converters (ADCs). At the second stage, the processing agreed upon for each probe pulse in all N distance channels is completed. To do this, in each channel of the sample, the digitized signal for each pulse is multiplied by the corresponding factor, taking into account the difference between the parameters of the local oscillator signal and the signal reflected from a particular section, and the samples are summed per pulse. Thus, by the sum of two stages, the input pulses are convolved with the corresponding pulses of the reference signals (coordinated processing of all pulses). Then, at the third stage, inter-period processing of bursts of pulses obtained in the synthesis interval is carried out. For this, for each range channel, the convolution results at the second stage are multiplied by the coefficients calculated for the corresponding terrain element (range channel), and are summed. If there are several receiving modules, the 4th stage is introduced, at which summation of the processing results in different receiving modules should be carried out.

The influence of phase distortion in the prototype can be illustrated by the results of numerical simulation of SAR. The motion of the SAR carrier was simulated, the continuous radiation of the probing LFM pulses generated in the transmitter, their reflection from point objects located at a specified distance from the SAR (with the given XY coordinates on the ground), and the coordinated processing of the received signals — the synthesis of radar images of the introduced reflectors. If necessary, the intrinsic noise of the PCA receiver was also simulated (with the required signal-to-noise ratio).

The following are the simulation results when simulating the flight of a SAR carrier in XY coordinates at a height of h _n :

- the initial coordinates of the carrier: X _n = 0, Y _n = -500 m, h _n = 1000 m, the only non-zero component of the velocity of the carrier V _un = 100 m / s;

- a single reflector is located at a point with coordinates X _OTR = 10000 m, Y _OTR = 0;

- the local oscillator is tuned to the reflector, i.e. X _g = 10,000 m, Y _g = 0;

- parameters of the pulsed LFM pulses formed in the transmitter and used in the synthesis: pulse duration τ _imp = 1 ms, pulse repetition period T _p = 0.1 s, initial carrier frequency f ₀ = 1 GHz, frequency deviation f _dev = 200 in each pulse MHz;

- the length of the synthesis interval D _synt = 1000 m,

- to suppress the side lobes, in addition to the coordinated processing, an additional Hamming weight treatment is used.

In FIG. Figures 1 and 2 show the response of the coordinated spatiotemporal processing to the action at the input of a coherent burst of LFM pulses with the above parameters in the absence of distortions of this signal during its passage through the entire path, i.e. radar data are shown when locating a SAR of a point reflector in the absence of other reflectors, noise, and distortion. In FIG. 1 is given by X-rays in the color (brightness) palette in XY coordinates. In FIG. Figure 2 shows the same response, but a side view is given in which the main lobe of the response and the low level of the side lobes are clearly visible. In FIG. Figures 3 and 4 show the same response, but in the presence of phase distortions in the received signal, simulated by introducing an additional phase change of the LFM pulse along a sinusoid with an amplitude of 30 degrees with a period of τ _imp /2.266 s. As can be seen from FIG. 3 and 4, distortions of the input signal led to a certain decrease in the main lobe, to its expansion, but, most importantly, to the appearance of significant side lobes, which will distort the radar data of the locality and, it can be considered, completely disrupt the PCA.

In FIG. Figures 5 and 6 show the radar image for the case under consideration when introducing phase distortion compensation in accordance with paragraphs. 4 and 7 of the proposed formula. From a comparison of FIG. 2 and FIG. Figure 6 shows that after compensation, the main lobe of the response has not changed, and the level of the side lobes, although it has increased (from about -42 dB to -35 dB), is quite consistent with the required level in practical implementations of weight processing.

To compensate for phase distortions of incoming reflected signals at the second stage during intra-pulse processing in each range channel, the reference pulse is divided into K _sub- pulses of the _sub- pulses, thereby dividing the CWB into the sum of narrow-band signals, since the deviation of the sub-pulse compared to the deviation of the original LFM pulse decreases by K _sub-pulse time. At the same time, we can assume that the distortions in the subpulse are insignificant, and they can be neglected - therefore, for each subpulse and for a pack of such subpulses of the same frequency range (subpulse number), coordinated processing is performed. It is important to compensate for distortions between sub-pulses, which is achieved by calculating and entering, when summing the signals of the sub-pulses, the correcting phase correction Δψ _ko . At the same time, it can be shown (see, in particular, the earlier article by A.P. Bryzgalov) that the input phase during the coordinated processing of the signal per pulse by summing the subpulse signals in each range channel (including the M-th channel) in the absence of distortion is approximated by a polynomial not higher than the second degree. The difference from this approximation is caused by distortion of the input signal. Therefore, assuming that the channel with the largest amplitude (Mth channel) and, accordingly, with the best signal / (noise + interfering reflection) ratio can be taken as a standard, we can write that Δψ _ko = ψ _Mk -ψ _ko , where ψ _ko is an approximation of the change in the measured values of ψ _Mk for K _sub- pulses of _sub- pulses. It can be noted that from the point of view of the influence of noise and interference, such an approximation and determination of Δψ _ko should be carried out after coordinated processing not of individual subpulses, but of packs of such subpulses for the synthesis interval. However, especially when modeling in the absence of noise and interference, the results of the coordinated processing of subpulses of a single pulse can be used.

_, можно проиллюстрировать фиг. 7, на которой приведено РЛИ для рассматриваемого случая при уменьшенном количестве подымпульсов с 15 до 10. В рассматриваемом примере искажения носят периодический характер в виде синусоиды. Функция корреляции тоже периодическая. Очевидно, что в этом случае требование по корреляции эквивалентно тому, чтобы длительность подымпульса была много меньше периода корреляции. Видно, что при 10 подымпульсах, когда длительность подымпульса примерно равна 0.23 периода изменения искажений, даже после компенсации искажения в этом случае значительные, в то время как при 15 подымпульсах искажения выходного сигнала приемлемы.The need for the condition that (in accordance with the formula) the duration of the sub-pulse τ of the _{sub-pulse is} less than the interval of the time correlation of distortions of the input signal

_Can be illustrated by FIG. 7, which shows the radar image for the case under consideration with a reduced number of subpulses from 15 to 10. In this example, the distortions are periodic in the form of a sinusoid. The correlation function is also periodic. Obviously, in this case, the correlation requirement is equivalent to the subpulse duration being much shorter than the correlation period. It can be seen that at 10 sub-pulses, when the sub-pulse duration is approximately equal to 0.23 of the distortion change period, even after compensation, the distortions in this case are significant, while at 15 sub-pulses the distortion of the output signal is acceptable.

The requirement that the interval of the set delay values of the input signals in time be significantly less than τ _{sub-amplitudes is} quite obvious. It is assumed that the delay of the local oscillator signal corresponds to this interval, for example, is equal to the average value of this interval. Failure to fulfill this condition will lead to a significant displacement of two pulses - signal and reference - relative to each other and, as a result, to a loss of energy potential and compression ratio, i.e. to reduce the range, resolution, accuracy, etc.

. В частности, при τ_имп=1 мс, К_подимп=20, f_дев=500 МГц разрешение составит 6 м. Для этого случая было проведено моделирование при f₀=500 МГц, Т_п=1 с при двух точечных отражающих объектах с координатами Х_отр1=9990 м, Х_охр2=10010 м, Y_отр1=Y_отp2=0 м, h_отр1=h_отр2=0 и при движении носителя РСА вдоль оси Y со скоростью -100 м/с из начальной точки Х_н=0, Y_н=500 м, h_н=100 м. На фиг. 8 приведено РЛИ для случая согласованной обработки, когда отсутствуют искажения и коррекция не проводится. На фиг. 9 приведены результаты, когда имеются искажения в виде аддитивного дополнительного изменения фазы входного сигнала в форме синусоиды с размахом 60 град и периодом 0.5 τ_имп. Компенсация не проводится. На фиг. 10 приведены результаты моделирования при наличии и искажений и компенсации (К_подымп=20). Из приведенных чертежей видно, что при компенсации практически полностью восстанавливается форма отклика РСА (его РЛИ), только незначительно возрастает уровень боковых лепестков, но он остается на уровне технических лепестков, имеющихся на практике.In the presence of several reflected signals with approximately equal intensity, for effective compensation of distortions it is necessary that these reflectors either coincide in range or be resolved by at least one of the processing parameters. In this case, resolution is possible in range (delay) or in linear resolution in angular coordinates (in azimuth). In this case, the linear resolution in azimuth is determined by the resolution in the SAR obtained during the synthesis. In range, the resolution is determined by the resolution of one subpulse, since compensation is made according to the signals received after processing the packets of subpulses. Taking into account that the frequency deviation in the subpulse during the LFM signal in K, the _{subpulse is} less than the deviation in the pulse, it is easy to obtain that when compensating, the range resolution is determined by the expression

. In particular, at τ _imp = 1 ms, K _subimp = 20, f _dev = 500 MHz, the resolution will be 6 m. For this case, modeling was performed at f ₀ = 500 MHz, T _p = 1 s for two point reflecting objects with coordinates X _sp1 = 9990 m, X _sp2 = 10010 m, Y _sp1 = Y _sp2 = 0 m, h _sp1 = h _sp2 = 0 and when the PCA carrier moves along the Y axis at a speed of -100 m / s from the starting point X _n = 0 , Y _n = 500 m, h _n = 100 m. In FIG. Figure 8 shows the radar image for the case of coordinated processing, when there are no distortions and no correction is performed. In FIG. Figure 9 shows the results when there are distortions in the form of an additive additional change in the phase of the input signal in the form of a sinusoid with a span of 60 degrees and a period of 0.5 τ _imp . No compensation. In FIG. 10 shows the simulation results in the presence of both distortion and compensation (K _subpump = 20). It can be seen from the drawings that, when compensated, the response form of the SAR (its XRD) is almost completely restored, the level of the side lobes only slightly increases, but it remains at the level of the technical lobes available in practice.

In FIG. 11 ... 13 shows the simulation results under boundary conditions for range resolution - the distance between two reflectors is 5 m (X _adr1 = 9997.5 m, X _ad2 = 10002.5 m, Y _ad1 = Y _ad2 = 0 m, h _ad1 = h _ad2 = 0 ) The radar image is given in case of coordinated processing in the absence of distortions, in the presence and at their compensation, by the method proposed by paragraph 7 of the formula. In FIG. 14 ... 16 shows the simulation results at a distance between two reflectors equal to 2 m (X _spr1 = 9999 m, X _spr2 = 10001 m). Approximately the same thing happens when the distance between the reflectors on the X axis is 4 m. It can be seen from the drawings that when the specified distance is reduced to 4 or less, the distortion compensation is violated.

However, with X _OT1 = X _{OT2, the} distortion compensation efficiency is restored. In FIG. 17 ... 19 shows the simulation result of processing with compensation at X _OTR1 = X _OTR2 = 10000 m and Y _OTR1 = -4 m, Y _OTR2 = 4 m. The compensation efficiency is high. The allowable difference in X is determined by the resolution of the LFM pulse.

Claims (13)

1. A method for signal processing ultrawideband (UWB) as a burst of N pulses _Pace with linear frequency modulation (LFM) for _Pace N ≥1, implemented using the filter-correlation method, according to which a received input signal is first multiplied by the reference signal, matched with the signal generated in the transmitter, with the formation of two quadrature channels, in each of which the signals are digitized and then processed in N distance channels, forming N pairs of quadrature components of the complex output signal processing for the given parameters, and then find the amplitudes of these signals, which determine N samples of the output module of the processing signal, characterized in that when processing an input signal distorted due to the nonlinearity of the phase-frequency characteristics of the path traveled by these signals from generation in the transmitter to processing in the receiver, correlation interval time τ distortion _{kor_isk,} each pulse of the reference signal is split into K _podymp narrowband podympulsov and in each quadrature channel all gave N nostnyh channel, the processing, matched with K _podymp batches of N _Pace podympulsov each forming a matrix of complex signals S

in the form of two quadrature components S _nkc , S _nks , where k = 1, ... K _podymp - serial number of the podimpulse, n = 1, ... N is the number of the range channel, and further, proceeding from the matrix S, the phase distortions Δψ _ko are estimated in each subpulse and these corrections are introduced into the subpulse signals corresponding to the number k

all channels, which are then summed in each range channel, forming the resulting N complex samples of the output signal. 2. The method according to p. 1, characterized in that the duration of the sub-pulse τ _{sub-steps is} chosen so that it is not less than 5 times less than τ _{cor_isk} , but not less than 5 times the interval of the set delay values of the input signals in time, and the signal-to-noise ratio during the coordinated processing of the burst of sub-pulses was significantly greater than 1, and when digitizing the signal, the time-quantization cycle was chosen so that the number of signal samples per sub-pulse was not less than 10. 3. The method according to p. 1, characterized in that to assess the phase distortion Δψ _ko, first, based on the matrix S, the initial output signal is generated for all N channels and the M-th range channel is selected from the initial signal, the amplitude of the output signal of which is not less than than 5 times exceeds the amplitudes of the output signals of other long-range channels that do not coincide with the M-th channel in time delay or are not allowed with it in any of the processing parameters, and further according to the signals

assess phase distortion Δψ _ko . 4. The method according to p. 3, characterized in that the original output signal is formed by summing the signal modules

of all K _sub- pulses of _sub- pulses in each nth range channel. 5. The method according to p. 3, characterized in that the initial output signal in each n-th range channel is formed by summing K _sub- signals

. 6. The method according to PP. 4 and 5, characterized in that to assess the phase distortion Δψ _ko in the selected Mth channel, based on the signals

, estimate the phase ψ _Mk for each subpulse as ψ _Mk = arctan (S _Mks / S _Mkc ), forming the vector ψ _M = (ψ _M1 , ... ψ _MKnodim ), then approximate the dependence of ψ _Mk on k by the polynomial ψ _ko = а + bk + ck ² , where a, b, and c are found by the least square deviation method, based on minimizing the value

and then phase corrections are determined to correct the distortion of the NWS in each kth subpulse based on the expression Δψ _ko = ψ _Mk -ψ _ko . 7. The method according to claim 6, characterized in that, to simplify the processing, ψ _{Mk is} approximated by a straight line ψ _ko = a + bk.

Images