JP2021136493A - Signal specification determination device and signal specification determination method - Google Patents

Abstract

To provide a signal specification determination device and a signal specification determination method, capable of identifying a modulation system without regarding a difference of a signal characteristic.SOLUTION: A signal specification determination device comprises: a signal characteristic analysis part 100 that analyzes a signal characteristic of an input signal; a frequency data generation part 200 that generates frequency data from the input signal on the basis of the analyzed signal characteristic; and a signal specification determination part 300 that classifies the frequency data into a class associated with a predetermined signal specification and determines the signal specification of a signal in accordance with the classified class. The frequency data generation part 200 generates the frequency data so as to match the signal characteristic of the input signal with a reference value on the basis of the analyzed signal characteristic.SELECTED DRAWING: Figure 1

Description

The present embodiment relates to a signal specification determination device and a signal specification determination method.

A signal specification determination device for determining the signal specifications of the received radio signal is used. A conventional signal specification determination device determines signal specifications by using a deterministic method for identifying a signal from features such as the shape of a spectrum, the presence or absence of a peak, the variance of an amplitude, and the signal bandwidth. Further, in recent years, a signal specification determination device for determining signal specifications using a machine learning method such as a support vector machine and a neural network has also been proposed.

Japanese Unexamined Patent Publication No. 2001-86171 Japanese Unexamined Patent Publication No. 2019-165320

In a signal specification determination device using a deterministic method, when identifying a signal from the shape of a spectrum, a phase-modulated signal having a different number of modulation multi-values, a frequency-modulated signal with a small number of modulation multi-values, and a frequency shift width are close to each other. It is difficult to identify a frequency modulated signal.

On the other hand, in the signal specification determination device using the machine learning method, the signal characteristics such as the type and bandwidth of the modulated signal can be specified from the spectrum. However, there are a wide variety of combinations of signal modulation methods and signal characteristics, including modulation types and multiple modulation values. Therefore, the difficulty of the problem becomes high, and the amount of data required for learning becomes enormous. Also, it is difficult to deal with unknown bandwidths that have not been learned.

An object to be solved by the embodiment is to provide a signal specification determination device and a signal specification determination method capable of identifying a modulation method regardless of a difference in signal characteristics.

The signal specification determination device according to the embodiment includes a signal characteristic analysis unit that analyzes the signal characteristics of the input signal, and a frequency data generation unit that generates frequency data from the input signal based on the analyzed signal characteristics. , The frequency data is classified into a class associated with a predetermined signal specification, and the signal specification determination unit for determining the signal specification of the signal is provided according to the classified class. The frequency data generation unit generates frequency data based on the analyzed signal characteristics so as to match the signal characteristics of the input signal with the reference value.

FIG. 1 is a conceptual diagram showing a configuration example of a signal specification determination device to which the signal specification determination method of one embodiment is applied. FIG. 2 is a flowchart showing an operation example of the signal specification determination device to which the signal specification determination method of one embodiment is applied. FIG. 3A is a diagram showing an example of a spectrogram of an FSK signal (signal characteristic 1). FIG. 3B is a diagram showing an example of a spectrogram of an FSK signal (signal characteristic 2). FIG. 4A is a diagram showing an example of a spectrogram of a PSK signal (signal characteristic 1). FIG. 4B is a diagram showing an example of a spectrogram of a PSK signal (signal characteristic 2). FIG. 5 is a diagram showing an example of a spectrogram in which an FSK signal (signal characteristic 2) is converted according to the signal characteristic. FIG. 6 is a diagram showing an example of a spectrogram in which a PSK signal (signal characteristic 2) is converted according to the signal characteristic. FIG. 7 is a conceptual diagram showing a modification 1 of the signal specification determination device to which the signal specification determination method of one embodiment is applied. FIG. 8 is a flowchart showing an operation example in the first modification of the signal specification determination device to which the signal specification determination method of one embodiment is applied. FIG. 9A is a diagram showing an example of a spectrogram (without conversion) of the BPSK signal. FIG. 9B is a diagram showing an example of a spectrogram (without conversion) of the QPSK signal. FIG. 10A is a diagram showing an example of a spectrogram (double multiplication) of the BPSK signal. FIG. 10B is a diagram showing an example of a spectrogram (double multiplication) of the QPSK signal. FIG. 11A is a diagram showing an example of a spectrogram (4 multiplication) of the BPSK signal. FIG. 11B is a diagram showing an example of a spectrogram (4 multiplication) of the QPSK signal. FIG. 12 is a conceptual diagram showing a modification 2 of the signal specification determination device to which the signal specification determination method of one embodiment is applied.

(Device description)
Hereinafter, the signal specification determination device to which the signal specification determination method of one embodiment is applied will be described with reference to the drawings. FIG. 1 is a conceptual diagram showing a configuration example of a signal specification determination device to which the signal specification determination method of one embodiment is applied. The signal specification determination device 10 includes a signal characteristic analysis unit 100, a frequency data generation unit 200, and a signal specification determination unit 300. The signal specification determination device 10 is a computer such as a personal computer. The signal characteristic analysis unit 100, the frequency data generation unit 200, and the signal specification determination unit 300 may be realized by software executed by a processor such as a CPU, or may be realized by dedicated hardware.

The signal characteristic analysis unit 100 receives, for example, a digital radio signal output from a receiver (not shown) and analyzes the signal characteristics of the received digital radio signal. Signal characteristics include, for example, bandwidth. Then, the signal characteristic analysis unit 100 outputs the signal characteristics obtained as a result of the analysis to the frequency data generation unit 200 as signal characteristic information.

The frequency data generation unit 200 generates frequency data from the digital radio signal based on the signal characteristic information output from the signal characteristic analysis unit 100. Then, the frequency data generation unit 200 outputs the generated frequency data to the signal specification determination unit 300.

The signal specification determination unit 300 determines the signal specifications from the frequency data output from the frequency data generation unit 200. Specifically, the signal specification determination unit 300 classifies the frequency data into any of a plurality of predetermined classes. Then, the signal specification determination unit 300 determines the signal specifications from the classified classes. Each class is associated with signal specifications in advance, and the signal specifications can be determined from the classified classes.

(Operation explanation)
Next, the operation of the signal specification determination device 10 of FIG. 1 will be described with reference to the flowchart shown in FIG. A digital radio signal from a receiver (not shown) is provided to the signal characteristic analysis unit 100 and the frequency data generation unit 200 of the signal specification determination device 10 (S1).

The signal characteristic analysis unit 100 analyzes the signal characteristics of the digital radio signal. Then, the signal characteristic analysis unit 100 outputs the analysis result as signal characteristic information to the frequency data generation unit 200 (S2).

Here, the signal characteristic information will be further described. In the embodiment, the signal specification determination unit 300 determines the signal specification from the spectrogram. The spectrogram is a two-dimensional image showing the time change of the frequency spectrum obtained as a result of applying the Fourier transform a plurality of times while shifting the input signal in the time direction. For example, the horizontal direction of this image represents a change in frequency, and the vertical direction represents a change in time. Then, the value of each pixel of this image represents the amplitude of the corresponding frequency component at each time.

3A and 3B show an example of a spectrogram generated by applying the same processing to two FSK (frequency modulation) signals having different signal characteristics. In the FSK signal, since the frequency is switched according to the symbol of the signal before modulation, a plurality of power peaks are generated on the frequency axis. Two peaks P11 and P12 are shown in FIG. 3A, and two peaks P21 and P22 are shown in FIG. 3B. The bandwidth as a signal characteristic corresponds to the shift width of the frequency to be switched. In FIG. 3A, the bandwidth is w1 (signal characteristic 1), and in FIG. 3B, the bandwidth is w2 (signal characteristic 2).

The bandwidth w2 of FIG. 3B is twice the bandwidth w1 of FIG. 3A. The signal characteristic analysis unit 100 estimates these bandwidths and outputs them as signal characteristic information. For example, the signal characteristic analysis unit 100 Fourier transforms the input digital radio signal to generate frequency data, detects a plurality of power peaks from the frequency data, and calculates the interval between those power peaks as a bandwidth. The power peak is a frequency at which the power difference from the surrounding frequency is equal to or greater than a predetermined value. The signal characteristic analysis unit 100 may perform a Fourier transform on a digital radio signal having an arbitrary signal length at an arbitrary timing, and obtain a bandwidth based on the obtained spectrum. Of course, the signal characteristic analysis unit 100 may generate a spectrogram to obtain the bandwidth.

4A and 4B show an example of a spectrogram generated by applying the same processing to two BPSK (two-phase shift keying) signals having different signal characteristics. In the PSK signal, since the phase is switched according to the symbol of the signal before modulation, signals having different bandwidths are generated according to the symbol rate. As for the bandwidth as a signal characteristic, the bandwidth is w3 (signal characteristic 1) in FIG. 4A, and the bandwidth is w4 (signal characteristic 2) in FIG. 4B.

The bandwidth w4 of FIG. 4B is twice the bandwidth w3 of FIG. 4A. The signal characteristic analysis unit 100 estimates these bandwidths and outputs them as signal characteristic information. In the case of PSK, for example, the signal characteristic analysis unit 100 Fourier transforms the input digital radio signal to generate frequency data, detects frequencies at a plurality of boundaries having a large power difference from the surroundings from the frequency data, and they are detected. Calculate the frequency interval between the boundaries as the bandwidth. The boundary frequency is a frequency at which the power difference from the surrounding frequency is equal to or greater than a predetermined value.

Here, the case where the bandwidth is obtained by general signal processing has been described, but any method including a machine learning method such as a neural network may be used as long as the bandwidth can be estimated.

Here, the explanation returns to FIG. The frequency data generation unit 200 generates a spectrogram as frequency data based on the signal characteristic information output from the signal characteristic analysis unit 100 (S3). In the embodiment, the frequency data generation unit 200 generates frequency data for signals having the same modulation method but different bandwidths so as to have the same bandwidth. In other words, the frequency data generation unit 200 generates frequency data based on the actual bandwidth provided as the signal characteristic information so that the apparent bandwidth in the spectrogram becomes a predetermined reference value. For example, when the actual bandwidth is n times the reference value, the generation of frequency data can be realized by multiplying the number of points of the Fourier transform by 1 / n times. In this case, the number of points in the frequency direction of the spectrogram differs depending on the bandwidth. On the other hand, the size of the spectrogram can be kept constant by clipping the frequency data according to a predetermined size or padding the value of the point where the value does not exist with a predetermined value. It should be noted that the same thing here does not mean an exact match, and some deviation may be allowed.

Here, the generation of frequency data will be further described. FIG. 5 is an example of generating a spectrogram of the signal shown in FIG. 3B with reference to the bandwidth of FIG. 3A. FIG. 6 is an example of generating a spectrogram of the signal shown in FIG. 4B with reference to the bandwidth of FIG. 4A. The spectrograms are generated so that the bandwidths are the same in both the examples of FIGS. 5 and 6. When simply generating a spectrogram, the spectrogram of FIG. 3A and the spectrogram of FIG. 3B, and the spectrogram of FIG. 4A and the spectrogram of FIG. 4B need to be treated as different classes. In the embodiment, the spectrograms are generated so that the bandwidths are the same, so that the signals of the same modulation method can be treated as the same class. This facilitates the classification in the signal specification determination unit 300. Here, the reference bandwidth does not have to be common to all modulation schemes. For example, a different reference value may be used for each rough modulation method that can be easily determined at the stage of bandwidth estimation such as FSK system and PSK system.

The generation of frequency data is not limited to the generation of the Fourier transform by changing the number of points. For example, after simply generating a spectrogram without changing the number of points of Fourier transform, if the actual bandwidth is wider than the reference value, the average value, median value, maximum value, etc. with the neighbors will be used. A method such as performing reduction processing or thinning processing based on the basis, and performing upsampling processing when the value is narrower than the reference value may be used. Further, the spectrogram may be generated after resampling the digital radio signal itself. Further, these processes may be switched according to the relationship between the actual bandwidth and the reference value. For example, when the actual bandwidth is narrower than the reference value, resampling of the radio signal may be used, and when it is wide, the spectrogram may be reduced or thinned out. Moreover, a combination of several treatments may be used. For example, a spectrogram may be calculated from a digital radio signal resampled so that the bandwidth is an integral multiple of a reference value, and the calculated spectrogram may be thinned out. Further, when the possible values of the bandwidth of the target digital radio signal are diverse, the frequency data may be generated for each group after being roughly divided into several bandwidth groups. .. In this case as well, the problem can be expected to be simplified compared to the case where the spectrogram is simply generated.

Further, in the Fourier transform for a digital radio signal, in general, the number of samples is set to a power of 2, and FFT (Fast Fourier Transform) is performed. Not limited to this, the number of samples other than the power of 2 may be filled with 0, and the FFT may be performed by matching the number of samples with the power of 2, or the DFT (Discrete Fourier Transform) is directly performed. May be good. Furthermore, the probability density function of the spectrum may be estimated based on an autoregressive model or the like.

In addition, a window function may be applied to the input signal when performing the Fourier transform. The window function may also be applied in the embodiment. In this case, in the embodiment, the type of the window function is not limited, and any window function may be applied.

Also, when generating the spectrogram, the Fourier transform may be applied so that the times of the digital radio signals overlap. In the embodiment, this overlap amount is not limited, and any overlap amount may be set. Furthermore, the amount of overlap may be adaptively controlled according to the duration of the signal.

Further, in the above description, it is assumed that the duration of the signal is constant. In reality, it is possible that the signal is present only at some time in the spectrogram. In such a case, the signal characteristic analysis unit 100 may obtain the duration of the signal as the signal characteristic information instead of the bandwidth. At this time, the frequency data generation unit 200 generates a spectrogram whose duration is adjusted to a certain reference value by controlling the amount of overlap in the time direction. Of course, the frequency data generation unit 200 may generate a spectrogram in which both the bandwidth and the duration are adjusted to a certain reference value.

Here, the explanation returns to FIG. The signal specification determination unit 300 performs classification processing on the frequency data output from the frequency data generation unit 200 (S4). Then, the signal specification determination unit 300 determines the signal specifications from the classified class (S5). As described above, since the class is associated with the signal specifications, the signal specifications can be determined from the class.

Here, an example using a neural network for determining the specifications will be described. CNN (Convolutional Neural Network) can be used as the specification determination using the neural network. That is, the classification process can be realized by CNN that inputs the spectrogram and outputs the vector data according to the predetermined number of classes. However, the signal specification determination unit 300 may classify the frequency data by various machine learning methods such as a support vector machine, not limited to CNN, as long as it is a method capable of appropriately classifying the frequency data.

As described above, according to the signal specification determination device 10 to which the signal specification determination method of the embodiment is applied, the signal characteristics are adjusted to a certain reference value according to the signal characteristics analyzed from the digital radio signal. Frequency data is generated. This makes it possible to realize signal specification determination that does not depend on signal characteristics. Further, when a neural network or the like is used for discriminating signal specifications, it is not necessary to prepare a large amount of learning data for each signal characteristic and solve a difficult classification problem.

Here, in the embodiment, the frequency data generated by the frequency data generation unit 200 is assumed to be a spectrogram. Frequency data is not limited to spectrograms. For example, the frequency data may be data having only the amplitude of the spectrogram, or may be data obtained by logarithmic transformation of the amplitude of the spectrogram. Frequency may be used instead of amplitude. Further, instead of generating the two-dimensional data by the Fourier transform, the two-dimensional data may be generated by applying the wavelet transform a plurality of times while shifting the time of the input signal. In addition, one-dimensional spectrum may be used as data instead of two-dimensional data.

(Modification example 1)
Hereinafter, a modification 1 of the embodiment will be described with reference to the drawings. FIG. 7 is a conceptual diagram showing a modified example of the signal specification determination device to which the signal specification determination method of the embodiment is applied. That is, the signal specification determination device 20 includes a non-linear processing unit 400 in addition to the signal characteristic analysis unit 100, the frequency data generation unit 200, and the signal specification determination unit 300. The same reference numerals are given to the processing units equivalent to the signal specification determination device 10, and the description thereof will be omitted as appropriate. That is, the nonlinear processing unit 400 will be mainly described below.

(Device description)
The non-linear processing unit 400 has a signal distribution unit 410 and a conversion unit 420. The signal distribution unit 410 receives, for example, a digital radio signal r (t) output from a receiver (not shown), and provides the received digital radio signal r (t) to a plurality of conversion units 420. In FIG. 1, four conversion units 420 (# 1 to # 4) are shown as an example. The number of conversion units 420 is not limited to four. The number of conversion units 420 may be appropriately determined according to the number and type of signal specifications determined by the signal specification determination device 20.

Functions for converting the digital radio signal r (t) are set in the conversion units 420 (# 1 to # 4), respectively, and a plurality of the conversion units 420 (# 1 to # 4) have a function for converting the digital radio signal r (t). A non-linear function is set.

The conversion units 420 (# 1 to # 4) generate the digital radio signal r (t) using the set functions (f1 (x), f2 (x), f3 (x), f4 (x)), respectively. The digital radio signals (f1 (r (t)), f2 (r (t)), f3 (r (t)), f4 (r (t))) are converted and output to the frequency data generation unit 200.

In this way, the non-linear processing unit 400 emphasizes the different characteristics of the digital radio signal by converting the digital radio signal r (t) with a function including a plurality of different non-linear functions.

The frequency data generation unit 200 uses a frequency conversion unit 210 provided in the conversion unit 420 on a one-to-one basis in order to convert each digital radio signal output from the conversion unit 420 (# 1 to # 4) into frequency data. I have. That is, in the example of FIG. 7, since there are four conversion units 420 (# 1 to # 4), the frequency data generation unit 200 corresponds to the four frequency conversion units 210 (# 1 to # 4). I have. The frequency conversion units 210 (# 1 to # 4) generate frequency data 1, 2, 3, and 4 from the digital radio signals provided by the corresponding conversion units based on the signal characteristic information, respectively.

The signal specification determination unit 300 classifies each of the plurality of frequency data 1 to 4 output from each of the frequency conversion units 210 (# 1 to # 4) into classes.

(Operation explanation)
Next, the operation of the signal specification determination device 20 to which the signal specification determination method of the modified example configured as described above is applied will be described with reference to a specific example according to the flowchart shown in FIG. The digital radio signal r (t) from a receiver (not shown) is provided to the signal characteristic analysis unit 100 of the signal specification determination device 20, and is also provided to the signal distribution unit 410 of the nonlinear processing unit 400. The non-linear processing unit 400 provides the provided digital radio signal r (t) to each conversion unit 420 (S11).

The signal characteristic analysis unit 100 analyzes signal characteristics such as bandwidth and signal duration in the same manner as the signal specification determination device 10. Then, the signal characteristic analysis unit 100 outputs the analysis result as signal characteristic information to the frequency data generation unit 200 (S12).

The conversion units 420 (# 1 to # 4) use the set functions (f1 (x), f2 (x), f3 (x), f4 (x)), respectively, to generate the digital radio signal r (t). To convert. Then, the conversion units 420 (# 1 to # 4) are the conversion result digital radio signals (f1 (r (t)), f2 (r (t)), f3 (r (t)), f4 (, respectively. r (t))) is output to the frequency data generation unit 200 (S13).

Here, as an example of the function applied in the conversion unit 420 (# 1 to # 4), a case where a function having a different multiplication factor is used will be described as an example. Assuming that the multiplication numbers are 1, 2, 4, and 8, respectively, the functions (f1 (x), f2 (x), f3 (x), f4 (x) set in each conversion unit 420 (# 1 to # 4)). ) Are as follows.
f1 (x) = x
f2 (x) = x ²
f3 (x) = x ⁴
f4 (x) = x ⁸
Of these, f1 (x) is a linear function. On the other hand, f2 (x), f3 (x) and f4 (x) are non-linear functions.

Here, the PSK signal has a characteristic that a peak occurs in the spectrum at the center frequency and the frequency of the bandwidth when the multiplication number is an integral multiple of the modulation multi-value number. Therefore, the digital radio signal output from each conversion unit 420 (# 1 to # 4) has different characteristics depending on the number of modulation multi-values.

In general, from the spectrum of the digital radio signal r (t), it is difficult to determine the modulation method of the FSK signal having a small number of modulation values and to classify the FSK signal having a small difference in frequency shift width. In the modified example, the conversion unit 420 (# 1 to # 4) converts the digital radio signal r (t) using a function having a different multiplication factor. As a result, the frequency shift amount of the digital radio signal output from each conversion unit is different, so that different characteristics are emphasized. Therefore, the classification of frequency modulated signals becomes easier.

The multiplication factor of the power function is not limited to 1, 2, 4, and 8. The multiplication number may be a value other than 1, 2, 4, and 8.

Here, the description returns to FIG. The conversion units 420 (# 1 to # 4) are the digital radio signal output from the corresponding conversion unit of the conversion units 420 (# 1 to # 4) and the signal output from the signal characteristic analysis unit 100, respectively. A spectrogram as frequency data is generated based on the characteristic information (S14).

FIG. 9A shows an example of a spectrogram of the BPSK signal generated by the frequency converter 210 (# 1). FIG. 9B shows an example of a spectrogram of the QPSK signal generated by the frequency converter 210 (# 1). In the frequency conversion unit 210 (# 1), neither the BPSK signal nor the QPSK signal reaches a match between the modulation multi-value number and the multiplication number. Therefore, as shown in FIGS. 9A and 9B, neither the BPSK signal nor the QPSK signal has a peak.

FIG. 10A shows an example of a spectrogram of the BPSK signal generated by the frequency converter 210 (# 2). FIG. 10B shows an example of a spectrogram of the QPSK signal generated by the frequency converter 210 (# 2). In the frequency conversion unit 210 (# 2), the modulation multi-value number and the multiplication number reach the same for the BPSK signal, and the modulation multi-value number and the multiplication number do not reach the same for the QPSK signal. Therefore, as shown in FIGS. 10A and 10B, a peak P is generated for the BPSK signal, whereas a peak is not generated for the QPSK signal.

FIG. 11A shows an example of a spectrogram of the BPSK signal generated by the frequency converter 210 (# 3). FIG. 11B shows an example of a spectrogram of the QPSK signal generated by the frequency converter 210 (# 3). In the frequency conversion unit 210 (# 3), the modulation multi-value number and the multiplication number reach the same for both the BPSK signal and the QPSK signal. Therefore, as shown in FIGS. 11A and 11B, a peak occurs in both the BPSK signal and the QPSK signal. Although the frequency conversion unit 210 (# 4) is not shown, the spectrogram is equivalent to that in FIGS. 11A and 11B.

Therefore, for example, if the spectrograms generated by the frequency converters 210 (# 1) to (# 3) are a combination of FIGS. 9A, 10A, and 11A, the input digital radio signal is a BPSK signal and has a frequency. If the spectrograms generated by the conversion units 210 (# 1) to (# 3) are a combination of FIGS. 9B, 10B, and 11B, the input digital radio signal can be determined to be a QPSK signal.

As described above, in the modified example 1, by converting the digital radio signal r (t) by a function including a plurality of different nonlinear functions, it is possible to emphasize different features in each converted digital radio signal. ..

In the first modification, the frequency conversion unit 210 (# 1 to # 4) generates frequency data based on the signal characteristics so that the spectrograms are the same regardless of the signal characteristics in the same modulation method. The frequency conversion units 210 (# 1 to # 4) may generate frequency data based on reference values of the same signal characteristics, or may generate frequency data based on different reference values.

Here, the description returns to FIG. The signal specification determination unit 300 performs classification processing on the frequency data whose features are emphasized, which is output from the frequency data generation unit 200 (S15). Then, the signal specification determination unit 300 determines the signal specifications from the classified class (S16). As described above, since the class is associated with the signal specifications, the signal specifications can be determined from the class. Here, the method for determining the specifications of the signal is not limited as long as it is a method capable of classifying a plurality of frequency data. For example, when a neural network is used, specification determination may be performed by a multi-channel neural network, or a plurality of frequency data are combined in a cascade in an arbitrary dimension, and a neural network is used for one combined data. Specifications may be determined by the network.

As described above, according to the first modification, the digital radio signal is transformed by a plurality of different non-linear functions. Then, from each converted digital data, frequency data whose signal characteristics are matched to a certain reference value is generated. Each converted digital data has different characteristics depending on the modulation method. Thereby, for example, it is possible to identify a phase-modulated signal having a different number of multi-values, a frequency-modulated signal with a small number of modulation multi-values, and a frequency-modulated signal having a similar frequency shift amount, which are difficult to identify in the past.

Here, in the first modification, the conversion unit 420 converts the digital radio signal using a power function. However, the function applied to the conversion unit 420 is not limited to the power function. As the function applied to the conversion unit 420, any non-linear function such as an exponential function, a sigmoid function, a logarithmic function, and an absolute value function can be used depending on the type and number of the identified signals.

(Modification 2)
Modification 2 will be described. In the above-described embodiment and the first modification, a plurality of methods are described for the method of generating frequency data according to the signal characteristics. On the other hand, the methods of the embodiment and the first modification can be applied even when the signal itself is shaped like resampling when generating frequency data. For example, the signal shaping unit 500 may be provided as in the signal specification determination device 30 shown in FIG. In this case, the signal shaping unit 500 performs shaping processing so as to match the signal characteristics of the input digital radio signal with a certain reference value. It is expected that the amount of processing will be reduced by performing common processing in advance in each frequency conversion unit.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, as well as in the scope of the invention described in the claims and the equivalent scope thereof.

10 ... Signal specification determination device, 100 ... Signal characteristic analysis unit, 200 ... Frequency data generation unit, 210 ... Frequency conversion unit, 300 ... Signal specification determination unit, 400 ... Nonlinear processing unit, 410 ... -Signal distribution unit, 420 ... Conversion unit, 500 ... Signal shaping unit.

Claims (5)

A signal characteristic analysis unit that analyzes the signal characteristics of the input signal,
A frequency data generator that generates frequency data from the input signal based on the analyzed signal characteristics, and
The frequency data is classified into a class associated with a predetermined signal specification, and a signal specification determination unit that determines the signal specification of the signal according to the classified class, and a signal specification determination unit.
With
The frequency data generation unit is a signal specification determination device that generates the frequency data so as to match the signal characteristics of the input signal with a reference value based on the analyzed signal characteristics. The signal specification determination device according to claim 1, wherein the signal characteristic includes a bandwidth of the signal. The signal specification determination device according to claim 1 or 2, wherein the signal characteristic includes the duration of the signal. The signal specification determination device according to any one of claims 1 to 3, wherein the signal specification determination unit classifies the frequency data into the class using a neural network. Analyzing the signal characteristics of the input signal and
Based on the analyzed signal characteristics, the frequency data is generated so as to match the signal characteristics of the input signal with the reference value.
By classifying the frequency data into classes associated with predetermined signal specifications,
Determining the signal specifications of the signal according to the classified class
A signal specification determination method comprising.

Images