JPH07113830A - Apparatus and method for analyzing frequency - Google Patents

Abstract

PURPOSE:To obtain a circuit and a method for analyzing a frequency whereby a continuing time of data and a time point when the data are generated are correctly measured even when the to-be-measured time series data are smaller in size than a capacity of a memory used for spectral analyses. CONSTITUTION:A data-extracting part 1 is constituted of two adjacent memory parts A11 and B12 each formed of a series of shift registers. Data stored in the memory parts A11 and B12 divided in N sections per predetermined memory capacity are subjected to frequency analysis at two spectral analyzing parts A21 and B22, whereby frequency spectra SA(f) and SB(f) are output. Only when time series data S(t) are stored in both memory parts A11 and B12 and two frequency spectra SA(f) and SB(f) are output, a mutual spectrum C(f) is output from a multiplier 23. If either spectrum is output, the mutual spectrum C(f) is not output.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a frequency analysis device and a frequency analysis method, and more particularly to a frequency analysis device and a frequency analysis method for analyzing a frequency spectrum of time series data.

[0002]

2. Description of the Related Art As shown in FIG. 2, a conventional frequency analyzer and frequency analysis method uses a memory section 3 for temporarily storing time series data S (t) and a predetermined number N of memory sections. It has the spectrum analysis unit 4 that reads out the series data, performs spectrum analysis, and outputs the frequency spectrum S (f).

[0003]

However, this conventional frequency analysis apparatus and frequency analysis method outputs the frequency spectrum of a signal when there is a signal in the N time series data to be analyzed by the spectrum analysis unit. To be done. When the duration of the signal is shorter than the time corresponding to N, the same frequency spectrum S (f) is maintained during the period in which all the signals are in the memory unit and is the analysis target of the spectrum analysis unit, that is, the time corresponding to N. Since it is output, it is difficult to measure the duration of the signal and the time of occurrence of the signal.

The present invention solves the problem of difficulty in measuring the duration and generation time of a signal that occurs when the duration of the signal is shorter than the time length of the time series data used for spectrum analysis. The purpose is to provide an analytical method.

[0005]

In order to achieve such an object, the frequency analysis apparatus of the present invention has an N storing one data.
Memory circuits connected in series to each other, two spectrum analysis means connected to each of the N storage circuits, and two frequency spectra analyzed by the two spectrum analysis means. And memory means for taking a product of these inputs and outputting a cross spectrum, and the two memory means are connected in series to form a series of shift registers. The spectrum analysis means performs a frequency analysis of the time series data based on the time series data stored in the storage circuit and sequentially stored in the storage circuit, and the frequency spectrum subjected to the frequency analysis is output to the multiplication means and two spectrum analyzes are performed. It is characterized in that the cross spectrum is outputted from the multiplication means only when the frequency spectrum is outputted from both means.

Further, the frequency analysis method for analyzing the frequency spectrum of the time-series data of the present invention includes a sequential storage step of sequentially storing the time-series data by the two memory means of the front stage and the rear stage, and the front stage and the rear stage. Frequency analysis step for performing frequency analysis for each time series data based on the time series data stored in the memory means of the section, and outputting the frequency spectrum subjected to the frequency analysis for each of the front section and the rear section, and the front section and the rear section. And a multiplication step of multiplying two frequency spectra for each part and outputting as one cross spectrum, when the two frequency spectra are output together, the cross spectrum is output.

[0007]

According to the frequency analyzing apparatus of the present invention, the time series data to be analyzed is sequentially inputted and stored in the storage circuit at the tip of the shift memory which is constituted by connecting two memory means in series, and stored in the storage circuit. The spectrum analyzing means performs frequency analysis of the time series data based on the time series data, the two frequency spectra analyzed by the two spectrum analyzing means are input, and the multiplying means for taking the product of these inputs and outputting the cross spectrum is The cross spectrum is output when the frequency spectrum is output from both of the two spectrum analyzing means, and the cross spectrum is not output when the frequency spectrum is not output from one or both of the two spectrum analyzing means. Therefore, regardless of the data length of the time series data, the cross spectrum is output only during the time when the time series data shifts the serial connection point of the two memory means.

According to the frequency analysis method of the present invention,
Frequency analysis is performed on the basis of time-series data sequentially stored in a shift memory composed of two memory units, a front stage part and a rear stage part, and a frequency spectrum subjected to frequency analysis is output for each of the front stage part and the rear stage part. When the two frequency spectra for each memory means in the latter part are multiplied and the two frequency spectra are output together, the mutual spectrum is output. When only one frequency spectrum is output, the cross spectrum is not output. Therefore, regardless of the data length of the time series data, the cross spectrum is output only during the time when the time series data shifts the series connection point of the two memory means.

[0009]

Embodiments of the frequency analyzing apparatus and the frequency analyzing method according to the present invention will be described in detail with reference to the accompanying drawings. Referring to FIG. 1, there is shown an embodiment of the frequency analyzer of the present invention. Hereinafter, the configuration of the present invention will be described in detail based on the embodiments shown in the drawings.

The frequency analyzing apparatus shown in FIG. 1 comprises a data extracting section 1 and a cross spectrum analyzing section 2. The data extraction unit 1 has a memory unit A11 and a memory unit B12, and the mutual spectrum analysis unit 2 is composed of a spectrum analysis unit A21, a spectrum analysis unit B22, and a multiplier 23.

Memory unit A11 constituting the data extraction unit 1
The memory unit B12 is a temporary storage unit including a so-called shift register. Each of the memory unit A11 and the memory unit B12 has N storage circuits each having a predetermined storage capacity.
12 are connected to each other in series and are formed by a series of shift registers.

The spectrum analysis unit A21 and the spectrum analysis unit B22 that form the mutual spectrum analysis unit 2 are connected to the individual storage circuits that form the memory unit A11 and the memory unit B12, respectively. Spectrum analysis unit A
21 and the spectrum analysis unit B22 determine the frequency spectrum SA of the data stored in the connected storage circuit.
(F) and SB (f) are analyzed, the frequency spectra analyzed for each memory circuit are added, and the result is output as one frequency spectrum. Therefore, if there is data in at least one of the memory circuits constituting the memory unit A11 or the memory unit B12, the spectrum analysis unit A2
1 or spectrum analysis unit B22 outputs frequency spectrum data, and if there is no storage data in all N storage circuits, no frequency spectrum is output.

The multiplier 23 takes the product of the two frequency spectrum signals SA (f) and SB (f) output from the spectrum analysis unit A21 and the spectrum analysis unit B22, and outputs the product as a cross spectrum C (f). It is a department. Therefore, when the spectrum signals SA (f) and SB (f) are output from both the spectrum analysis unit A21 and the spectrum analysis unit B22, the cross spectrum C
When (f) is output and neither one or both outputs the spectrum signal, the cross spectrum C (f) is not output.

When the time-series data S (t) is input to the data extraction unit having the above structure, the time-series data S (t)
Is input to the data extraction unit 1. In the data extraction unit 1, the memory unit A11 sequentially stores the latest N time-series data S (t), and the memory unit B12 sequentially stores the previous N time-series data. A series of time series data is divided into two adjacent time series data.

The time series data S (t) of the memory unit A11 and the memory unit B12 are subjected to frequency analysis by the spectrum analysis unit A21 and the spectrum analysis unit B22 of the mutual spectrum analysis unit 2, respectively. Spectrum analysis unit A2
1 and the frequency spectrum SA (f) and the frequency spectrum SB (f) output from the spectrum analysis unit B22.
Is further multiplied by the multiplier 23 in the cross spectrum analysis unit 2 and output as a cross spectrum C (f).

When the duration of the signal included in the time-series data S (t) is shorter than the time corresponding to the number N of data, the signal is initially included only in the memory unit A11, and therefore the frequency spectrum of the signal is the spectrum. It appears only in the frequency spectrum SA (f) that is the output of the analysis unit A21, and the frequency spectrum SB (f) is not output from the spectrum analysis unit B22. Therefore, the frequency spectrum S from the multiplier 23
Cross spectrum C, which is the signal of the product of A (f) and SB (f)
(F) is not output.

However, as time passes, the signal remains in the memory section A1.
1 and the memory unit B12, the frequency spectra SA (f) and SB (f) are output, and the multiplier 23 outputs the cross spectrum C (f).

When the time further passes, the signal is sent to the memory section B12.
In the state of being stored only in the frequency spectrum SB
Since only (f) is output and the frequency spectrum SA (f) is not output, the output signal of the cross spectrum C (f) does not appear.

Therefore, the cross spectrum C (f) is output at the time when the time series data S (t) is stored in both the memory unit A11 and the memory unit B12, and this time is two memories. The connection point of the time series data S (t)
Is almost the same as the shift time.

As described above, the present invention comprises data extraction for dividing a series of time series data into two adjacent time series data and cross spectrum analysis for calculating a cross spectrum of the two time series data. Therefore, even if the duration of the signal is shorter than the time length of the time-series data used for spectrum analysis, the frequency spectrum of the signal appears at the output approximately equal to the duration of the signal, so the measurement of the duration and generation time of the signal is possible. It has the effect of making it easier.

Although the above-described embodiment is a preferred example of the present invention, the present invention is not limited to this and various modifications can be made without departing from the gist of the present invention.

[0022]

As is apparent from the above description, in the frequency analysis apparatus of the present invention, the two spectrum analysis means perform the frequency analysis of the time series data based on the time series data stored in the storage circuits of the two memory means. Then, the two analyzed frequency spectra are input, the product is obtained, and the cross spectrum is output. This cross spectrum is output only when there is a frequency spectrum from both of the two spectrum analysis means, that is, only when the time series data shifts the series connection point of the two memory means.
Therefore, regardless of the data length of the time series data, it becomes easy to detect the start point and the end point of the time series data and measure the duration time.

According to the frequency analysis method of the present invention,
Frequency analysis is performed on the basis of time-series data sequentially stored in a shift memory composed of two memory units of a front stage and a rear stage, and a frequency spectrum subjected to the frequency analysis is output for each of the memory units of the front stage and the rear stage, When the two frequency spectra for each of the front stage and the rear stage are multiplied and two frequency spectra are output together, the cross spectrum is output, and when one is output, the cross spectrum is not output. Therefore, regardless of the data length of the time-series data, the cross spectrum is output only at the time when the time-series data shifts the series connection point of the two memory means. Therefore, the start point and the end point of the time-series data S are detected and continued. The time can be easily measured.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a frequency analysis circuit of the present invention.

FIG. 2 is a block diagram showing an example of a conventional frequency analysis circuit.

[Explanation of symbols]

 1 Data Extraction Section 2 Mutual Spectrum Analysis Section 11 Memory Section A 12 Memory Section B 21 Spectrum Analysis Section A 22 Spectrum Analysis Section B 23 Multiplier

Claims (2)

[Claims] 1. Two memory means configured by connecting N storage circuits for storing one data in series, and two spectrum analysis means connected to each of the N storage circuits, And two multiplication means for inputting the two frequency spectrums analyzed by the two spectrum analysis means, taking the product of these and outputting the cross spectrum, the two memory means being connected in series to form a series of shift memories. The analyzed time-series data is input to and sequentially stored in the storage circuit at the tip of the shift memory, and the spectrum analysis means performs frequency analysis of the time-series data based on the time-series data stored in the storage circuit. When the frequency spectrum subjected to frequency analysis is output to the multiplication means, and the frequency spectrum is output from both of the two spectrum analysis means. Frequency analyzing apparatus characterized by mutual spectrum is outputted from the look said multiplying means. 2. A frequency analysis method for analyzing a frequency spectrum of time-series data, comprising a sequential storage step of sequentially storing the time-series data by two memory means of a front stage and a rear stage, the front stage and the rear stage. A frequency analysis step of performing frequency analysis for each time series data based on the time series data stored in the memory means of the unit, and outputting the frequency spectrum subjected to the frequency analysis to each of the memory means of the front stage and the rear stage. And a multiplication step of multiplying two frequency spectra for each of the front-stage part and the rear-stage part and outputting as one cross spectrum, wherein the cross spectrum is output when both of the two frequency spectra are output. A frequency analysis method characterized by the above.