RU2761983C2 - Method for determining frequencies of multiple signals in subsampling receiver - Google Patents

Abstract

FIELD: radio engineering.

SUBSTANCE: method relates to the field of radio engineering and can be used for broadband analysis of the electronic situation. A multi-channel receiver is being used. Each channel performs analogue-to-digital conversion, amplitude spectrum computation, and signal detection. The signals detected in the channels are compared in frequency and amplitude. When the signals coincide, a decision is made about their existence at these frequencies. The decision on the coincidence of the signals in amplitude is taken if the ratio of the amplitudes of the output signals for each pair of channels is within the specified limits.

EFFECT: ensuring the possibility of determining the frequencies of time superimposed signals in the receiver with subsampling while reducing the number of processing channels.

3 cl, 4 dwg

Description

The method relates to the field of radio engineering and can be used for broadband analysis of the electronic situation.

If the frequency of the received signal is higher than half the sampling frequency of a digital receiver, then such a receiver operates in the subsampling mode [1]. The advantage of using subsampling lies in the multiple expansion of the range of processed frequencies while maintaining or slightly increasing the amount of equipment used. However, in a complex signal environment, signals at the input of a wideband receiver are likely to overlap in time. In this case, even in the absence of aliasing of the signal spectra, an error occurs in determining the frequency. The nature of this error depends on the specific downsampling receiver circuitry.

From the technical level known receiver with downsampling, described in the patent [2]. The receiver contains three channels, each containing an ADC. One channel can measure the frequency in the first Nyquist zone, i. E. if the frequency of the input signal is less than half the sampling frequency. Accordingly, when using only one channel for frequency measurement in the second and subsequent Nyquist zones, ambiguity arises. Two channels are used to detect an oversampling input frequency. The third channel in a wide frequency range eliminates the ambiguity in determining the frequency arising from the correspondence to one tuple (an ordered set, in which the order of elements is important) of frequencies in the first Nyquist zone of several values of the input frequency.

в многоканальном приемнике с субдискретизацией в общем случае основан на решении системы линейных сравнений [3, 4]:The principle of determining the input frequency

in a multichannel receiver with downsampling, in the general case, it is based on the solution of a system of linear comparisons [3, 4]:

(1)

(one)

- число каналов,

- целое число, модуль которого равен частоте

измеренной в первой зоне Найквиста i-го канала приемника,

- частоты дискретизации, используемые в каналах приемника. Значения

могут принимать как положительные, так и отрицательные значения;

- только неотрицательные.where

- number of channels,

- an integer, the modulus of which is equal to the frequency

measured in the first Nyquist zone of the i-th receiver channel,

- sampling rates used in the receiver channels. The values

can take both positive and negative values;

- only non-negative ones.

можно представить как сумму или разность частоты дискретизации произвольно выбранного i-го канала приемника

, умноженной на некоторое неотрицательное число, и частоты

, определенной в первой зоне Найквиста i-го канала приемника.According to system (1), any input frequency

can be represented as the sum or difference of the sampling rate of an arbitrarily chosen i-th receiver channel

multiplied by some non-negative number, and the frequency

defined in the first Nyquist zone of the i-th receiver channel.

приемник должен по кортежу

построить все возможные кортежи

и решить для каждого из них систему сравнений (1).To determine the actual frequency of the input signal

receiver must by tuple

build all possible tuples

and solve for each of them a system of comparisons (1).

In a particular case, the solution of the comparison system (1) and, accordingly, the determination of frequencies is performed based on the rotation of the amplitude spectra of the signals detected in all channels on a single frequency axis into a set of Nyquist zones in the reverse order of their superposition during sampling. Signals are considered detected at those frequencies that correspond to the coincidence on a single frequency axis of the spectral components detected in all channels. This approach is used in [5].

(фиг. 1). Таким образом, этому сигналу соответствует кортеж значений

Но при наличии на входе нескольких сигналов невозможно понять, каким входным частотам соответствуют те или иные кортежи. Например, сигналы с частотами

и

формируют соответствующие им кортежи частот в первой зоне Найквиста

и

(фиг. 2). На фиг. 2 для примера выделен кортеж, соответствующий сигналу с частотой

. Однако частоты в первой зоне Найквиста могут сформировать и такой кортеж частот, который будет соответствовать несуществующей частоте на входе приемника

(фиг. 3). Найденная таким образом частота

является ложной. Для данного приемника устранение ложных частот решается с помощью технического решения, описанного в [6] и принимаемого за прототип. Для определения M частот наложенных во времени сигналов используется M+1 каналов обработки с различными частотами дискретизации. В каждом канале выполняется аналого-цифровое преобразование, вычисление амплитудного спектра на основе быстрого преобразования Фурье и сравнение амплитуд спектральных составляющих с величиной заранее выбранного порога обнаружения. Значения частот, соответствующие максимальным амплитудам сигнала, используются в алгоритме устранения неоднозначности при оценке нескольких частот. Этот алгоритм определяет истинные частоты на основе необходимого количества их обнаружения в разных каналах. Недостатком данного технического решения является большое количество каналов обработки, которое зависит от количества принимаемых сигналов. Determining the frequencies of time superimposed signals in such a receiver is difficult. So, in a three-channel receiver, one signal in the ith channel corresponds to its own frequency in the first Nyquist zone

(Fig. 1). Thus, this signal corresponds to a tuple of values

But if there are several signals at the input, it is impossible to understand which input frequencies these or those tuples correspond to. For example, signals with frequencies

and

form the corresponding tuples of frequencies in the first Nyquist zone

and

(Fig. 2). FIG. 2, for example, a tuple corresponding to a signal with a frequency

... However, the frequencies in the first Nyquist zone can form such a tuple of frequencies, which will correspond to a non-existent frequency at the input of the receiver.

(Fig. 3). The frequency found in this way

is false. For this receiver, the elimination of false frequencies is solved using the technical solution described in [6] and taken as a prototype. To determine the M frequencies of the time superimposed signals, M + 1 processing channels with different sampling rates are used. Each channel performs analog-to-digital conversion, calculation of the amplitude spectrum based on the fast Fourier transform, and comparison of the amplitudes of the spectral components with the value of the preselected detection threshold. The frequency values corresponding to the maximum signal amplitudes are used in the disambiguation algorithm when evaluating multiple frequencies. This algorithm determines the true frequencies based on the required number of detections in different channels. The disadvantage of this technical solution is the large number of processing channels, which depends on the number of received signals.

The purpose of the proposed method is to provide the ability to determine the frequencies of superimposed signals in time in the receiver with downsampling while reducing the number of processing channels.

The technical result consists in reducing the mass and overall dimensions of the receiver, as well as increasing its reliability.

This result is achieved by using the proposed method for determining the frequency in a digital multichannel receiver with downsampling. In this case, analog-to-digital conversion is performed in each channel, the amplitude spectrum of each digitized sequence is calculated and the amplitudes of the spectral components are compared with the value of the preselected detection threshold. The channels use different sampling rates. Additionally, in the range of operating frequencies of the receiver, the minimum and maximum values of the transmission coefficient of each channel are measured. The frequencies of the detected spectral components are swept in all channels into a plurality of Nyquist zones on a single frequency axis in the reverse order of their superposition during sampling. The signals detected in the channels, deployed in many Nyquist zones, are compared with each other in frequency and amplitude. When the signals coincide in frequency and amplitude, a decision is made about their existence at these frequencies. Moreover, the decision on the coincidence of the signals in amplitude is made if two conditions are met for each pair of channels:

- the ratio of signal amplitudes at the outputs of the channels does not exceed the ratio of the maximum value of the transmission coefficient of one channel to the minimum value of the transmission coefficient of the other;

- the ratio of the signal amplitudes at the outputs of the channels is not lower than the ratio of the minimum value of the transmission coefficient of one channel to the maximum value of the transmission coefficient of the other.

Amplitude comparison can be performed only for one channel in pairs with the others.

When comparing the signal amplitudes at the channel outputs, the ratio of the maximum value of the transfer coefficient of one channel to the minimum value of the transfer coefficient of the other and the ratio of the minimum value of the transfer coefficient of one channel to the maximum value of the transfer coefficient of the other are chosen with margins, taking into account the assumed measurement errors.

In practice, in a wide frequency range, random signals from various sources of radio emission, located at different distances from the receiver and having different power and other radiation parameters, arrive at the input of the receiver. Therefore, with a wide dynamic range of the receiver, which in practice is ensured by the use of appropriate ADCs, the signals at the receiver input differ significantly in power. Accordingly, to determine whether the signal in the first Nyquist zone belongs to one or another input signal, it is possible to compare the signals in the first Nyquist zone in amplitude. Thus, the possibility of determining the frequencies of the signals superimposed in time is ensured by distinguishing the signals by their amplitude.

In practice, however, the receiver has channel diversity caused by the difference in gains of the channels, and unevenness in the gain in each channel, caused by operation in a wide frequency band. When comparing in amplitude, it is required to take into account the diversity of channels and unevenness.

и максимальное

значения коэффициента передачи каждого канала приемника во всем диапазоне рабочих частот.For this, before carrying out frequency measurements, measure the minimum

and the maximum

values of the transmission coefficient of each channel of the receiver in the entire range of operating frequencies.

Consider two receiver channels, the inputs of which receive a signal with amplitude A. At the output of the first channel, the signal amplitude is A _1meas , at the output of the second, the amplitude is A _2meas .

Two extreme cases are possible, which correspond to the largest difference in the transmission coefficients in channels 1 and 2. In the first case, the ratio of the signal amplitudes at the channel outputs is determined as:

In the second case:

Thus, if the condition is met:

then we can assume that the signals at the outputs of channels A _meas1 and A _meas2 refer to the same input signal.

In general terms, taking into account the margin for the measurement error, this condition can be formulated as:

(2)

(2)

и

- коэффициенты запаса, выбираемые с учетом предполагаемых погрешностей измерения, m и n - номера каналов приемника,

а j и p - номера обнаруженных в первой зоне Найквиста сигналов,

M - число сигналов, обнаруженных в первой зоне Найквиста.where

and

are the safety factors selected taking into account the expected measurement errors, m and n are the numbers of the receiver channels,

and j and p are the numbers of signals detected in the first Nyquist zone,

M is the number of signals detected in the first Nyquist zone.

Provided that the frequency of the components of the spectrum of the input signal, deployed in a plurality of Nyquist zones, and the ratio of the amplitudes at the outputs of the channels, satisfying condition (2), the decision is made that the signals at the outputs of the channels correspond to one input signal. If this condition is not met, a decision is made that the signals compared in amplitude correspond to different signals at the receiver input.

Amplitude comparison should be performed for all channels in pairs. Those. for a three-channel receiver, the amplitudes in channels 1 and 2, 1 and 3, 2 and 3 are compared.

. Тогда для трехканального приемника проверка должна быть выполнена, например, для каналов 1 и 2, 1 и 3. Отличие разницы между максимальным и минимальным (минимальным и максимальным) коэффициентами передачи каналов 2 и 3 от разниц тех же величин между другими парами каналов повышает точность попарного сравнения по амплитуде между всеми каналами. Однако отсутствие сравнения между каналами 2 и 3 позволяет сократить объем вычислений и на практике незначительно влияет на достоверность принятого в результате всех сравнений решения, если коэффициенты передачи каналов отличаются незначительно.Amplitude comparison can be performed only for one channel in pairs with the others. One of the channels is considered the first, respectively, for (2) we get:

... Then, for a three-channel receiver, a check must be performed, for example, for channels 1 and 2, 1 and 3. The difference between the maximum and minimum (minimum and maximum) gains of channels 2 and 3 from the differences of the same values between other pairs of channels increases the accuracy of the pairwise comparison in amplitude between all channels. However, the absence of a comparison between channels 2 and 3 allows one to reduce the amount of computations and, in practice, insignificantly affects the reliability of the decision made as a result of all comparisons, if the transmission coefficients of the channels differ insignificantly.

Since the number of digital receiver channels N does not depend on the number of simultaneously processed signals M, then N can be minimal. Accordingly, unlike the prototype, with an increase in the number of signals, the mass and overall dimensions of the receiver do not increase, and its reliability does not decrease either. Thus, the technical result, which the proposed method is aimed at, is achieved.

FIG. 1 shows the signal in the first Nyquist zone for a three-channel receiver, as well as the corresponding true input signal.

FIG. 2, a case similar to that shown in FIG. 1, but for two signals.

FIG. 3 illustrates finding the false frequency when there are two signals at the input.

FIG. 4 shows a diagram of a receiver that implements the proposed method.

An example of implementation of the proposed method is illustrated on the basis of a digital multichannel receiver circuit (Fig. 4). The receiver contains N channels. In each channel, ADC 1, a spectrum calculator (UVS) 2 and a detection device (ID) 3 are connected in series. The channels are connected to a solver 4.

разветвляется на N каналов. В каналах с помощью АЦП 1.1 - 1.N выполняют аналого-цифровое преобразование с различными частотами дискретизации. На практике для расширения полосы рабочих частот перед каждым АЦП 1 могут использовать дополнительное устройство выборки и хранения. После оцифровки в каждом канале с помощью устройства вычисления спектра 2 получают амплитудный спектр оцифрованной последовательности, полученной с учетом своей частоты дискретизации. Устройство обнаружения 3 выполняет обнаружение спектральных отсчетов, превышающих заданный порог, и измеряет соответствующую данному отсчету частоту в первой зоне Найквиста

. Затем в решающем устройстве 4 получают оценки истинных частот входных сигналов. С этой целью для обнаруженных спектральных составляющих развертывают спектры во всех каналах во множество зон Найквиста на единую ось частот в порядке, обратном их наложению при дискретизации. Выбирают частоты, совпадающие во всех каналах по частотному положению и амплитуде. Причем решение о совпадении сигналов по амплитуде принимают, если для каждой пары каналов (или для каждой пары каналов, образованной с одним из каналов) выполняются два условия:The operating frequency range of the receiver exceeds half the highest sampling rate, i.e. the receiver operates in downsampling mode. Previously, before receiving signals, in the operating frequency range of the receiver, the minimum and maximum values of the transmission coefficient of each channel are measured. When the receiver is operating, the mixture of input signals with frequencies

branches into N channels. Analog-to-digital conversion with different sampling rates is performed in the channels using ADC 1.1 - 1.N. In practice, an additional sampling and storage device can be used to expand the bandwidth in front of each ADC 1. After digitization in each channel using a spectrum calculator 2, the amplitude spectrum of the digitized sequence is obtained, obtained taking into account its sampling frequency. Detector 3 detects spectral samples exceeding a predetermined threshold and measures the frequency corresponding to this sample in the first Nyquist zone

... Then, in the solver 4, estimates of the true frequencies of the input signals are obtained. For this purpose, for the detected spectral components, the spectra are scanned in all channels into a plurality of Nyquist zones on a single frequency axis in the reverse order of their superposition during sampling. Select frequencies that coincide in all channels in terms of frequency position and amplitude. Moreover, the decision on the coincidence of the signals in amplitude is taken if for each pair of channels (or for each pair of channels formed with one of the channels) two conditions are met:

- the ratio of signal amplitudes at the outputs of the channels does not exceed the ratio of the maximum value of the transmission coefficient of one channel to the minimum value of the transmission coefficient of the other;

- the ratio of the signal amplitudes at the outputs of the channels is not lower than the ratio of the minimum value of the transmission coefficient of one channel to the maximum value of the transmission coefficient of the other.

принимают решение о существовании на них сигналов.In case of coincidence in frequency position and amplitude for the corresponding frequency estimates

make a decision about the existence of signals on them.

The method shown in FIG. 4 receiver. The ADCs 1.1 - 1.N included in the receiver are selected according to the operating frequency range. UVS 2.1 - 2.N, UO 3.1 - 3.N and solver 4 in practice can be implemented in software, for example, based on FPGA. Thus, the implementation of the method is possible using the available electronic component base.

List of sources

1. Kondakov DV, Lavrov AP Determination of the frequency spectrum of a multicomponent radio signal in a digital receiver with subsampling // Radiotekhnika. 2019.Vol. 83, No. 9 (13). S. 52-61.

2. Patent US 5099194 A. Digital frequency measurement receiver with bandwidth improvement through multiple sampling of real signals / Richard B. Sanderson, James B. Y. Tsui. No. 672310; declared 03/06/1991; publ. 03.24.1992.

3. Huang X., Bai R., Jin X., Fu H. Robust and Efficient Frequency Estimator for Undersampled Waveforms Based on Frequency Offset Recognition / PLoS ONE. 2016. Vol. 11, no. 10. e0163871. doi: 10.1371 / journal.pone.0163871.

4. Xiao L., Xia X. G. Frequency Determination from Truly Sub-Nyquist Samplers Based on Robust Chinese Remainder Theorem // Signal Processing. 2018. Vol. 150, p. 248-258.

5. Patent RU 2516763 C1. Receiving device / A. N. Krenev et al. No. 2012143923/08; declared 10/15/2012; publ. 05/20/2014, Bul. No. 14.

6. Patent US 5293114 A. Frequency measurement receiver with means to resolve an ambiguity in multiple frequency estimation / William S. McCormick, James B. Y. Tsui. No. 997262; declared 12.24.1992; publ. 03/08/1994.

Claims (3)

1. A method for determining the frequency in a digital multichannel receiver with downsampling, in which an analog-to-digital conversion is performed in each channel, the amplitude spectrum of each digitized sequence is calculated, the amplitudes of the spectral components are compared with the value of the preselected detection threshold, and the sampling frequencies in the channels are different, differing in that that in the operating frequency range of the receiver the minimum and maximum values of the transmission coefficient of each channel are measured, the frequencies of the detected spectral components are swept in all channels into a plurality of Nyquist zones on a single frequency axis in the reverse order of their superposition during sampling, the signals detected in the channels are compared by frequency and amplitude, when the signals coincide in frequency and amplitude, a decision is made on their existence at these frequencies, and the decision on the coincidence of signals in amplitude is taken if for each pair of channels the ratio of signal amplitudes to the output ax of the channels does not exceed the ratio of the maximum value of the transmission coefficient of one channel to the minimum value of the transmission coefficient of the other, and the ratio of the signal amplitudes at the outputs of the channels is not lower than the ratio of the minimum value of the transmission coefficient of one channel to the maximum value of the transmission coefficient of the other. 2. The method according to claim 1, characterized in that the amplitude comparison can be performed only for one channel in pairs with the others. 3. The method according to claim 1, characterized in that when comparing the amplitudes of the signals at the outputs of the channels, the ratio of the maximum value of the transfer coefficient of one channel to the minimum value of the transfer coefficient of the other and the ratio of the minimum value of the transfer coefficient of one channel to the maximum value of the transfer coefficient of the other are multiplied by the safety factors , selected taking into account the expected measurement errors.

Images