RU2339958C1 - Method of measuring frequency of sinusoidal signals and device for implementing method - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: method allows for on the results of the calculated effective processing of the phase difference statistics of the signal with use of optimum weighing of data according to the invention, increasing accuracy of measuring frequency and other parameters, coherent on the interval of measuring packets of sinusoidal signals. In particular it is possible to increase accuracy of measuring parameters of signals for locators with a synthetic aperture and coherent Doppler units. The obtained parameter values are the best in the root-mean-square sense of evaluation.

EFFECT: increased accuracy of measuring frequency of sinusoidal signals on the background of noise due to synthesis of statistically optimum, calculated effective evaluation of coherent packets of sinusoidal signals in the measuring interval.

4 cl, 2 dwg

Description

The invention relates to communication technology and can be used to measure the frequency of sinusoidal signals in information-measuring devices.

The task of measuring the frequency of sinusoidal signals constantly attracts the attention of researchers in connection with its fundamental significance in evaluating the parameters of periodic signals mixed with noise to extract information about the Doppler shift in communication, navigation and radar problems [1-3].

A number of known methods for measuring the frequency of sinusoidal signals [4-7], however, they are not statistically optimal from the point of view of measuring the frequency of sinusoidal signals against a background of noise [8, 9].

There are a number of methods for measuring the frequency of sinusoidal signals [8, 10, 11], based on the use of the Fourier transform, in which the maximum power spectral density (PSD) argument of the signal-to-noise mixture is taken as an estimate of the frequency of the sinusoidal signal. The indicated methods of measuring the frequency are statistically optimal from the point of view of measuring the frequency of sinusoidal signals against a background of noise [8, 9]. However, these methods require large computational costs associated with the implementation of one or more Fourier transforms, which does not allow to obtain an estimate of the frequency on a time scale close to real.

A known method of measuring the frequency of sinusoidal signals [12-14] based on the use of differential-phase statistics of the signal, which is also statistically optimal from the point of view of measuring the frequency of sinusoidal signals against the background of noise [12, 13], which reaches the Cramer lower boundary for large sample sizes -Rao. The method [14] is significantly more computationally efficient than the methods [8, 10, 11], since it does not require the implementation of Fourier transforms, is closest to the proposed one and therefore adopted as a prototype.

According to this method:

1) receive the current phase of the signal as an argument of a complex number, in-phase samples are used as the real part, and quadrature samples, converted to digital form, of the filtered components of the quadrature decomposition of the signal are used as the imaginary part;

2) receive the difference ψ _k adjacent current phases of the signal;

3) form a two-component state vector a (k) = [a ₁ (k), a ₂ (k)] ^T , whose components are given by recurrence expressions

в соответствии с формулой:4) determine the estimate of the signal frequency

according to the formula:

where f _s is the sampling frequency of the signal samples.

The prototype device [14] contains connected to its input, through two parallel chains of series-connected multiplier, low-pass filter (low-pass filter) and analog-to-digital converter (ADC), the corresponding inputs of read-only memory (ROM), while the second inputs of the multipliers are connected with a common source of a sinusoidal signal, directly at the first multiplier, and through the phase shifter at the second multiplier, the ROM output is connected to the input of the subtraction device, between which the subtractive input and the output The ROM contains a memory element, and the clock inputs of both ADCs are connected to a common clock generator (TG), the first and second arithmetic devices (AU) are connected to the output of the subtraction device, between the outputs and second inputs of which are included the second and third memory elements, third inputs both ACs are connected via a counter to a clock generator, between the outputs of the first and second ACs a second subtraction device is turned on, the output signal of which is proportional to the measured signal frequency, while the first AC operates in Compliant with the expression

and the second AU - in accordance with another expression

where ψ _k is the difference of adjacent current phases of the signal from the ROM output, and ₁ (k) and a ₁ (k-1) are the current and previous values of the first component of the state vector from the output of the first AU, a ₂ (k) and a ₂ (k-1) is the current and previous values of the second component of the state vector from the output of the second control unit, k is the index variable k = 1, 2, ..., formed by the counter according to clock cycles, indicating the sequence number of the current phase difference ψ _k .

However, as with analogs [8, 10-13], the statistically optimal, from the point of view of measuring the frequency of sinusoidal signals against noise [12, 13], prototype method [14] when measuring the frequency of short packets of sinusoidal signals measurement is limited to the lower boundary of the Cramer-Rao [12, 13]

₀) - дисперсия (или квадрат среднеквадратической ошибки) измеряемой частоты, f_s - частота выборки сигнала, SNR - входное отношение сигнал/шум, N - количество отсчетов, используемых при измерении частоты, для сигнала длительностью Т, равной целой части Т·f_s.where var (

₀ ) is the variance (or squared root mean square error) of the measured frequency, f _s is the sampling frequency of the signal, SNR is the input signal-to-noise ratio, N is the number of samples used in measuring the frequency for a signal of duration T equal to the integer part T · f _s .

In addition, a number of systems work with coherent packets of a sinusoidal signal when there is no signal in the pauses between packets or radiation from other sources is present. Such operating modes occur in coherent radars [15, 23, 24] and in packet communication systems, for example, in a GSM cellular communication system [25]. In these cases, the signal frequency is estimated by the prototype method using only one separate signal packet and the coherence of the signal packets is not used. A problem with the application of the prototype method may also occur in other wireless communication systems in conditions of signal fading.

The technical result of the invention is to improve the accuracy of measuring the frequency of sinusoidal signals against a background of noise due to the synthesis of a statistically optimal, computationally effective estimation for coherent packets of sinusoidal signals on the measurement interval.

The technical result is achieved in that in a method for measuring the frequency of sinusoidal signals, including obtaining the current phase of the signal, as an argument to a complex number, the real part of which is used in-phase samples, and as the imaginary part is quadrature samples converted to digital form, the filtered components of the quadrature decomposition signal, obtaining differences ψ _i the phase related current signal, where i = 1, 2, ... - index variable, meaning the sequence number of the current phase difference ψ _i, in accordance to fig The data blocks of in-phase and quadrature samples of a signal of length N are accumulated, L samples of the filtered components of the quadrature decomposition of a signal are passed through, second blocks of data of in-phase and quadrature samples of a signal of length N are accumulated, the accumulated data blocks of in-phase samples are combined into a single data block of 2N length, and the accumulated data blocks of quadrature samples - to another data block of 2N length, for combined blocks of in-phase and quadrature samples of the signal, each 2N long, for each pair of in-phase x and quadrature signal samples corresponding to a single point in time, the current phase of the signal is obtained, which is formed from (2N-1) difference ψ _i adjacent current phases combined signal for successive current-difference signal phase ψ _i, i = 1, 2, .. ., (2N-V) obtained from a sample of the combined 2N in-phase and quadrature samples of the signal form a weight function in accordance with the following expression

and the estimate of the signal frequency is determined in accordance with the formula

where f _s is the sampling frequency of the signal samples.

Another technical result of the invention is to increase the accuracy of measuring the phase increment of the sinusoidal signals, for coherent packets of sinusoidal signals on the measurement interval.

The technical result is achieved in that in the method for measuring the frequency of sinusoidal signals according to the invention, an additional function is determined for the increment of the phase of the signal in accordance with the following expression

Another technical result of the invention is a device for implementing a method for measuring the frequency of sinusoidal signals, which increases the accuracy of frequency measurement by using a statistically optimal, computationally efficient estimation for packets of sinusoidal signals that are coherent in the measurement interval.

The technical result is achieved in that in a device for implementing a method for measuring the frequency of sinusoidal signals, to the input of which are connected two parallel chains of series-connected multipliers, a low-pass filter (low-pass filter) and an analog-to-digital converter (ADC), while the second inputs of the multipliers are connected to a common the source of the sinusoidal signal, the first multiplier - directly, and the second multiplier - through the phase shifter, the output of a permanent storage device (ROM), forming the current The signal base is connected to the input of the subtraction device, between the subtracting input of which and the output of the ROM a memory element is connected, and the clock inputs of both ADCs are connected to a common clock generator (TG), also connected to a counter that generates an index variable i = 1, 2, ..., an arithmetic device (AU) is connected to the output of the subtraction device, while at the output of the subtraction device, successive current differences ψ _i , i = 1, 2, ... of the phase of the signal are selected from a quadrature sample, according to the invention, the second input of the AU by it is connected to the output of the second ROM, the third input of the AU is connected to the control device, to the outputs of the first and second ADCs, respectively, the first and second inputs of the first ROM are connected through the corresponding series-connected switch and buffer random access memory (BOS), each of which contains 2N elements memory and is connected in the same way as the control unit with the control device, the control inputs of the switches are connected to the counter via the control device, the counter reset input is connected to the TG through another control output of the device control, while the AU, the output signal of which is proportional to the measured frequency of the signal, operates in accordance with the expression

where f _s is the TG frequency, and the second ROM generates a weight function from the (2N-1) coefficient according to the following formula

Another technical result of the invention is a device for implementing the method, increasing the accuracy of measuring the phase increment of the sinusoidal signals, for packets of sinusoidal signals that are coherent in the measurement interval.

The technical result is achieved in that in a device for implementing a method for measuring the frequency of sinusoidal signals according to the invention, a second AU is additionally connected to the output of the AU, the second input connected to the counter via an additional output of the control device, while the output of the second AU is proportional to the phase increment function of the signal with the following expression

Figure 1 presents the structural diagram of a device in which the proposed method for measuring the frequency of sinusoidal signals is implemented.

Figure 2 presents the structural diagram of another embodiment of the device, which implements the proposed method for measuring the frequency of sinusoidal signals.

When co-processing coherent packets of sinusoidal signals, it is necessary to exclude from processing the signals of other sources present in the pauses between packets. This requires skipping data samples during a pause in the radiation of a source whose frequency is measured.

According to the proposed method for measuring the frequency of sinusoidal signals:

1) accumulate data blocks of in-phase and quadrature samples of a signal of length N digitized, filtered components of the quadrature decomposition of the signal;

2) miss L samples filtered components of the quadrature decomposition of the signal;

3) accumulate the second data blocks in-phase and quadrature samples of a signal of length N;

4) sequentially combine the accumulated data blocks of the in-phase samples into one data block 2N long, and the accumulated data blocks of the quadrature samples into another data block 2N long;

5) for the combined blocks of in-phase and quadrature samples of the signal, each 2N long, for each pair of in-phase and quadrature samples of the signal corresponding to one moment of time, the current phases of the signal are obtained as an argument of a complex number, the corresponding part of which is used as the real part, and as imaginary - the corresponding quadrature samples, converted to digital form, of the filtered components of the quadrature decomposition of the signal;

6) from the obtained current signal phases form (2N-1) the difference ψ _{i of} adjacent current phases of the combined signal, for successive current phase differences of the signal ψ _i , i = 1, 2, ..., (2N-1) obtained from the sample from the combined 2N in-phase and quadrature samples of the signal;

7) form a weight function in accordance with the following expression

8) determine the estimate of the frequency of the signal in accordance with the formula

where f _s is the sampling frequency of the signal samples.

The estimate of the signal frequency obtained in this way is the best in the mean-square sense for coherent packets of sinusoidal signals that are coherent in the measurement interval.

Show it.

Consider the reception of coherent packets of a sinusoidal signal when the source does not emit in a pause between packets, as, for example, is the case with the reception of signals of coherent pulsed radars [15, 23, 24] and in communication systems with temporary access [3, 25].

In this case, the source data is a sample of a signal including two packets of a complex sinusoid, the model of which has the form

where z _t = ζ _Rt + j · ζ _It is the sample from the white Gaussian sequence with zero mean and variance σ ² _Z. As can be seen from (1), the signal sampling consists of two signal packets, including N points each, separated by a gap of L points. The initial data can be understood as the product of a segment of a complex sinusoid, of duration t = 0.1, ..., 2N + L-1 with a time window in which L points of the signal are missing.

The following definitions similar to those given in [12] apply to each individual moment t and are transferred automatically (regardless of the peculiarities of the packet structure); for the initial signal, we have

where the complex random sequence is introduced

having a mean and variance E {ξ _t } = 0, var {ξ _t } = σ ² _z / A ² , respectively. Then (2) can be rewritten identically in the form

As shown in [12], the current phase of signal (4) can be represented as

If we determine the difference-phase measurements for the data (1), then we obtain the following sequence

Let the uncertainty in the desired frequency be small enough in comparison with the duration of the skip between packets

по одиночным пакетам, исходный сигнал переносится на полученную оценку, а на втором этапе уточняется лишь малая (в пределах погрешности начальной оценки) поправка ω₀, для которой условие (7) менее ограничительно.This assumption is justified in most practical cases (synthesized aperture locators, coherent Doppler radars, etc. [3, 15-16, 18]), especially during two-stage processing, when the frequency is estimated at the first stage

for single packets, the initial signal is transferred to the obtained estimate, and at the second stage only the small correction ω ₀ (within the error of the initial estimate) is specified, for which condition (7) is less restrictive.

Taking into account definition (5), the measurement model can be written in vector form

where the measurement vector ψ∈ = R ^2N-1 , and its components are defined by the left ^- hand sides of (6), 1 is the vector of N-1 units, η∈R ^2N-1 is a random vector whose components are given on the right-hand sides of (6) in brackets [•]. For the components η, taking into account (5), we can show

Thus, despite the packet structure of the initial signal, the form of the covariance matrix C = E {ηη ^T } remains the same as that given in [12] and allows the matrix to be inverted explicitly. Given that the order of C is 2N-1, for the elements of the inverse matrix C ^-1 we have

Where

The optimal (in the sense of minimum dispersion) unbiased estimate for model (8) under conditions (9) has the form [9]

where B is a matrix with elements of B _mn normalized to a constant coefficient, the inverse covariance matrix. Expression (13) determines the desired frequency estimate; however, the special form of the matrices included in it makes it possible to explicitly obtain the identity relation

where v ^T (N, L) is the unnormalized weight function, and V (N, L) is the normalizing factor, and both quantities are uniquely determined by the structure parameters of the initial signal (N, L).

If we denote by B _{n the} nth column of the matrix B, then the nth component (1≤n≤2N-1) of the unnormalized weight function can be defined as

The normalizing factor from (11.12) is defined as

The last term in (16) is equal to

then from (16) for the normalizing factor in (14) we finally have

Since the value of estimate (12) identically coincides with estimate (11), using relations (13, 15), it can be shown that the variance of the desired frequency estimate (11) for a signal sample including two packets of a complex sinusoid is

the variance of the desired frequency estimate is equal to (for the sampling frequency equal to unity)

The variance of the frequency estimate for the prototype method, where only one signal packet is used in the measurement, for a sampling frequency equal to unity, as shown in [12-14], is

The ratio K of the variance of the frequency estimate (17) and (18) for the proposed method and the prototype method shows

раз лучше точности измерений способа-прототипа.that the accuracy of measuring the frequency of sinusoidal signals by the proposed method is more than

times better than the accuracy of the measurements of the prototype method.

For sufficiently large N and L = kN

раз лучше точности измерений, получаемой при использовании способа-прототипа.That is, even with k = 1, the accuracy of measuring the frequency of sinusoidal signals by the proposed method more than in

times better than the accuracy of measurements obtained using the prototype method.

за счет учета на интервале измерения когерентности обоих пакетов синусоидальных сигналов.If, when using the prototype method, both signal packets are used independently, then by averaging the two frequency estimates, the variance of the frequency estimate (18) can be halved, while the proposed method will surpass the prototype in accuracy by more than

due to accounting for the coherence measurement interval of both packets of sinusoidal signals.

The proposed measurement method also allows, through direct integration in discrete time, to obtain an estimate of the signal phase increment function in accordance with the following expression

which is used in the second embodiment of the proposed method.

The estimation of the signal phase increment function is needed, in particular, in applications related to the processing of speech signals [18, 19], when it is necessary to restore a unique phase function (in the foreign literature the term “unwrapping” is used to designate such an operation on a signal), which is generally case of a discontinuous 2π-periodic function.

с использованием выражения (21) с исключением влияния дребезга скачка фазы на 2π.A number of methods are known for estimating the function of the phase increment of a signal [20-22], however, they do not provide an unambiguous restoration of the phase function under conditions of bounce of the phase jump by 2π at moderate signal-to-noise ratios [15, 18-22]. Since the proposed method, synthesized for operation at moderate signal-to-noise ratios, provides higher estimation accuracy, it can provide a more accurate restoration of the unique phase function by directly integrating the values in discrete time

using expression (21) with the exception of the effect of bounce of the phase jump by 2π.

According to the second variant of the proposed method for measuring the frequency of sinusoidal signals:

1) accumulate data blocks of in-phase and quadrature samples of a signal of length N digitized, filtered components of the quadrature decomposition of the signal;

2) miss L samples filtered components of the quadrature decomposition of the signal;

3) accumulate the second data blocks in-phase and quadrature samples of a signal of length N;

4) sequentially combine the accumulated data blocks of the in-phase samples into one data block 2N long, and the accumulated data blocks of the quadrature samples into another data block 2N long;

5) for the combined blocks of in-phase and quadrature samples of the signal, each 2N long, for each pair of in-phase and quadrature samples of the signal corresponding to one moment of time, the current phases of the signal are obtained as an argument of a complex number, the corresponding part of which is used as the real part, and as imaginary - the corresponding quadrature samples, converted to digital form, of the filtered components of the quadrature decomposition of the signal;

6) from the obtained current signal phases form (2N-1) the difference ψ _{i of} adjacent current phases of the combined signal, for successive current phase differences of the signal ψ _i , i = 1, 2, ..., (2N-1) obtained from the sample from the combined 2N in-phase and quadrature samples of the signal;

7) form a weight function in accordance with the following expression

8) determine the estimate of the frequency of the signal in accordance with the formula

where f _s is the sampling frequency of the signal samples;

9) determine the function of the phase increment of the signal in accordance with the following expression

A device that implements the proposed method for measuring the frequency of sinusoidal signals (see Fig. 1) contains two parallel chains of a series-connected multiplier (P) 1, a low-pass filter (LPF) 2 and an analog-to-digital converter (ADC) 3 and connected to its input , respectively, from multiplier 4, low-pass filter 5, ADC 6. The second inputs of multipliers 1 and 4 are connected to a common source of a sinusoidal signal (OISS) 7, for multiplier 1 - directly, and for multiplier 4 - through a phase shifter (PV) 8. Constant output mass storage devices a (ROM) 9, which forms the current phase of the signal, is connected to the input of the subtraction device (B) 10. Between the subtracting input of the device 10 and the output of the ROM 9, a memory element (EM) 11 is turned on. 11. The clock inputs of both ADCs 3 and 6 are connected to a common clock (TG) 12, also connected to a counter (MF) 13, the output of which is formed by an index variable i = 1, 2, .... An arithmetic device (AU) 14 is connected to the output of the subtraction device 10, while the output of the device 10 is formed consecutive current differences ψ _i , i = 1, 2, ... of the phase of the signal from a sample of quad mural counts. According to the invention, the second input of the AU 14 is connected to the output of the second ROM 15. The third input of the AU 14 is connected to the control device (CU) 16. The first input of the ROM 9 is connected to the output of the first ADC 3 through a series-connected switch (K) 17 and a buffer random access memory ( BOZU) 18. The second input of the ROM 9 is connected to the output of the second ADC 6 through a series-connected switch 19 and BOZU 20. Both BOSU 18 and 20 contain 2N memory elements. Both BOSU 18 and 20 are connected in the same way as AU 14, with UU 16. The control inputs of the switches 17 and 19 are connected to the counter 13 through UU 16. The reset input of the counter 13 is connected to TG 12 through another control output of UU 16. AU 14, the output signal is proportional to the measured frequency of the signal, operates in accordance with the expression

where f _s is the frequency of TG 12, and the second ROM generates a weight function from the (2N-1) coefficient according to the following formula

In another embodiment of the device described above (see FIG. 2), the second AC 21 is additionally connected to the output of the AC 14. The second input of the AC 21 is connected to the counter 13 through the additional output of the AC 16, while the output of the second AC is proportional to the phase increment function of the signal in accordance with the following expression

The first variant of the proposed device (see figure 1), which implements a method of measuring the frequency of sinusoidal signals, works as follows.

The signal in the mixture with noise from the input of the device is supplied to the first inputs of the multipliers 1, 4. In the products of multiplying the input signal with the signal OISS 7, the frequency of which is equal to the center frequency of the input frequency range of the signal, the output of the multiplier 1 contains the in-phase component of the input signal. In the products of the multiplication of the input signal with the signal OISS 7, phase-shifted 90 ° phase shifter 8, the output of the multiplier 4 contains the quadrature component of the input signal. Each of these components in the low-frequency band equal to the half-width of the frequency range of the input signals is filtered in the low-pass filter 2 and 5, respectively. Next, the ADC 3, clocked by TG 12, converts the in-phase component of the input signal to digital form (into in-phase readings), and the ADC 6, clocked TG 12, converts the quadrature component of the input signal into digital form (into quadrature readings). According to the signals of UU 16, the counter 13 is reset (zeroed) and the switches 17 and 19 are opened to supply in-phase and quadrature samples, respectively, to the BOZU 18 and 20. After accumulating in the BOZU 18 and 20 N N samples of the UU 16 signal determines what is in the counter 13 contains the number N and from the next beat TG 12 closes the switches 17 and 19 and stops the data flow to the BOZU 18 and 20. As soon as the counter status becomes N + L, then from the next beat TG 12 UU 16 reopens the switches 17, 19 and resumes supply of in-phase and quadrature readings, respectively Specifically, in BOZU 18 and 20. After 2N samples of the signal are accumulated in BOZU 18 and 20, UU 16 determines that the state of the counter becomes 2N + L and closes switches 17 and 19 from the next beat of TG 12. The accumulated data from the outputs of BOZU 18 and 20 served in ROM 9. In ROM 9 receive the current phase of the signal as an argument of a complex number, the real part of which is used in-phase samples, and as the imaginary part - quadrature samples. Next, the differences of adjacent current phases of the signal ψ _i are obtained as a result of subtracting the current phase from the output of the ROM 9 in the subtracting device 10 with the previous phase value delayed in the memory element 11.

In the first AC 14 based on the difference ψ _{i of the} adjacent current phases of the signal obtained from the output of the subtraction device 10, based on the index variable i fed to the third input of AC 13 from the AC 16, and also based on the coefficients of the weight functions w (i) obtained , respectively, from the second ROM 15 and formed according to the expression

determine the estimate of the frequency of the signal in accordance with the formula

из АУ 14 подается на выход предлагаемого устройства.The accumulation of the amount in AC 14 according to the last formula is carried out as the index variable i grows, and after the difference ψ _{i of} adjacent current signal phases arrives in AC 14 2N-1, the signal frequency is estimated

from AU 14 is fed to the output of the proposed device.

, полученной из АУ 14, и на основе индексной переменной i, подаваемой на второй вход АУ 21 через УУ 16, определяют функцию приращения фазы сигнала в соответствии со следующим выражениемThe work of the second variant of the proposed device (figure 2) differs from the work of the first in that in the second AU 21 based on an estimate of the signal frequency

obtained from AU 14, and on the basis of the index variable i supplied to the second input of AU 21 through AU 16, the function of the signal phase increment is determined in accordance with the following expression

The results can be used, in particular, to improve the accuracy of measuring the frequency and other parameters of coherent packets of sinusoidal signals (locators with a synthesized aperture, coherent Doppler radars, etc. [3, 15-16, 18]).

Claims (4)

1. A method of measuring the frequency of sinusoidal signals, including obtaining the current phase of the signal as an argument of a complex number, the real part of which is used in-phase samples, and as the imaginary part is quadrature samples converted to digital form, filtered components of the quadrature signal decomposition, obtaining differences раз _i adjacent current phases of the signal, where i = 1, 2, ... is an index variable meaning the sequence number of the current phase difference Ψ _i , characterized in that the filter data blocks are accumulated of in-phase and quadrature samples of a signal of length N, L samples are passed, the filtered components of the quadrature decomposition of a signal are passed, the second data blocks of filtered in-phase and quadrature samples of a signal of length N are accumulated, the accumulated data blocks of in-phase samples are combined into a single data block of 2N length, and the accumulated data blocks quadrature samples into another data block 2N long, for combined blocks of common-mode and quadrature samples of the signal, each 2N long, for each pair of common-mode and quad Atrial samples of the signal corresponding to one moment of time receive the current phases of the signal, from which form the difference (2N-1) Ψ _i , adjacent current phases of the combined signal, for successive current phase differences of the signal Ψ _i , i = 1, 2, ... , (2N-1) obtained from a sample of combined 2N in-phase and quadrature samples of the signal form a weight function in accordance with the following expression

and the estimate of the signal frequency is determined in accordance with the formula

where f _s is the sampling frequency of the signal samples. 2. The method of measuring the frequency of sinusoidal signals according to claim 1, characterized in that it further determines the function of the phase increment of the signal in accordance with the following expression φ (i) = 2πf ₀ i. 3. The device for implementing the method of measuring the frequency of sinusoidal signals according to claim 1, the input of which is connected to two parallel chains of series-connected multipliers, a low-pass filter (LPF) and an analog-to-digital converter (ADC), while the second inputs of the multipliers are connected to a common source sinusoidal signal: for the first multiplier - directly, and for the second multiplier - through the phase shifter, the output of the read-only memory (ROM) forming the current phase of the signal is connected to the input of the device of subtraction, between the subtracting input of which and the output of the ROM a memory element is switched on, and the clock inputs of both ADCs are connected to a common clock generator (TG), also connected to a counter that generates an index variable i = 1, 2, ... at the output an arithmetic device (AU) is connected to the subtraction device, and at the output of the subtraction device, successive current differences Ψ _i , i = 1, 2, ... of the signal phase are drawn from a sample of quadrature samples, characterized in that the second input of the AU is connected to the output of the second ROM, third the AC input is connected to the control device, to the outputs of the first and second ADCs, respectively, the first and second inputs of the first ROM are connected through the corresponding series-connected switch and buffer random access memory (BOS), each of which contains 2N memory elements and is connected, just like and AU, with the control device, the control inputs of the switches are connected to the counter via the control device, the counter reset input is connected to the TG through another control output of the control device, while the AU, the output whose signal is proportional to the measured frequency of the signal, operates in accordance with the expression

where f _s is the TG frequency, and the second ROM generates a weight function from the (2N-1) coefficient according to the following formula

where N is the number of in-phase and quadrature samples (their length), L is the number of skipped samples. 4. A device for implementing the method of measuring the frequency of sinusoidal signals according to claim 3, characterized in that the second output is additionally connected to the output of the AC, the second input is connected to the counter through an additional output of the control device, while the output of the second AC is proportional to the function of the signal phase increment in according to the following expression

Images