JP2007248348A - Signal component calculation apparatus and measuring apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately compute signal data by simple computations on the basis of observation data in which a signal component and a noise component are mixed and appropriately compute signal data, even if noise sources of different frequency bands are present. <P>SOLUTION: Fourier transformation is applied to observation data a(t), to determine the amplitude A(ω) and the phase ϕ(ω) of a component of a frequency ω. Fourier transformation is applied to noise data n(t), to determine the amplitude N(ω) and the phase component Ψ(ω) of the frequency ω. The amplitude S(ω) of the signal vector of the frequency ω is determined, on the basis of the amplitude A(ω), the phase ϕ(ω), and the phase Ψ(ω). The phase θ(ω) of a signal vector is determined on the basis of the phase Ψ(ω). Accurately computed signal data can be obtained by simple computations, on the basis of observation data in which the signal component and the noise components are mixed. Appropriately computed signal data can be obtained, even if the noise sources of different frequency bands are present. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a signal component calculation device and a measurement device. More specifically, the present invention can accurately calculate a signal component from observation data in which a signal component and a noise component are mixed with a simple calculation, and a noise source having a different frequency band. The present invention relates to a signal component calculation device and a measurement device that can appropriately calculate a signal component even if there is an error.

Conventionally, the observation data a in which the signal S and the noise N are mixed is collected and the noise data n is collected to obtain a weight wo that minimizes ∫ (a−w · n) ₂ , and s = a−wo · An apparatus for calculating signal data s from n is known (see Patent Document 1).
JP-A-9-243722

The conventional apparatus has a problem in that the calculation for obtaining the weight wo that minimizes ∫ (a−w · n) ₂ becomes complicated.
Further, the conventional apparatus has a problem that it is difficult to obtain the weight wo appropriately when there are a plurality of noise sources having different frequency bands. For example, when only low-frequency noise of about 0.1 Hz to 10 Hz is mixed, the weight wo can be obtained appropriately, but when the power supply noise per 50 Hz is added to it, the weight wo should be obtained appropriately. There was a problem that became difficult.
Therefore, an object of the present invention is to calculate a signal component with high accuracy by simple calculation from observation data in which a signal component and a noise component are mixed. An object of the present invention is to provide a signal component calculation device and a measurement device that can be calculated.

In the first aspect, the present invention obtains the amplitude A (ω) and the phase φ (ω) of the component of the frequency ω included in the observation data a including the signal S and the noise N, and the noise data n including only the noise N Frequency component acquisition means for obtaining the amplitude N (ω) and phase ψ (ω) of the component of frequency ω included in the signal, and based on the amplitude A (ω), the phase φ (ω), and the phase ψ (ω) Signal amplitude calculation means for obtaining the amplitude S (ω) of the component of frequency ω of the signal S, and signal phase calculation for obtaining the phase θ (ω) of the component of frequency ω of the signal S based on the phase ψ (ω) And a signal component calculating device.
In the signal component calculation apparatus according to the first aspect, the amplitude A (ω) and the phase φ (ω) of the component of the frequency ω are obtained by performing, for example, Fourier transform on the observation data a (t). Further, the amplitude N (ω) and the phase ψ (ω) of the component of the frequency ω are obtained by performing Fourier transform on the noise data n (t).

As shown in FIG. 4, an observation vector [A (ω), φ (ω)] and a noise vector [N (ω), ψ (ω)] at the same frequency ω are considered. The weighted noise vector [W (ω) · N (ω), ψ (ω)] obtained by multiplying the noise vector [N (ω), ψ (ω)] by the weight W (ω) is used as the observation vector [A (ω ), Φ (ω)] subtracted from [A (ω), φ (ω)] − [W (ω) · N (ω), ψ (ω)] represents an error vector [E (ω) , Θ (ω)]. When a weight W (ω) that minimizes this error is given, the error vector [E (ω), θ (ω)] is a signal vector representing a signal.
However, as can be seen from FIG. 3, the error vector [E (ω), θ (ω)] is minimized because the error vector [E (ω), θ (ω)] is a weighted noise vector [W ( ω) · N (ω), ψ (ω)].

As can be seen from the geometrical relationship in FIG. 3, the amplitude S (ω) of the signal vector [S (ω), θ (ω)] is, for example, S (ω) = A (ω) · cos (φ ( (ω) −ψ (ω)). Further, the phase θ (ω) can be obtained by, for example, θ (ω) = ψ (ω) + π / 2. In this way, the signal amplitude S (ω) and phase θ (ω) can be obtained for each frequency ω.
The signal S (t) in the time domain is
S (t) = sum_ω {S (ω) · sin (ωt + θ (ω))}
= Sum_ω {A (ω) · cos (φ (ω) −ψ (ω)) · sin (ωt + ψ (ω) + π / 2)}
It becomes. Here, sum_ω {} means that a sum is taken for ω.
Alternatively, the signal S (t) can be obtained by performing an inverse Fourier transform on S (ω) and θ (ω).
In addition, when the value of N (ω) is very smaller than the value of A (ω) at a certain frequency ω (for example, when the ratio A (ω) / N (ω) is 4 or more), the calculation is forcibly performed. May be considered S (ω) = A (ω), θ (ω) = φ (ω), assuming that the influence of noise can be ignored because the S / N ratio is good.

In a second aspect, the present invention relates to an observation sensor and an observation data collection device that collect observation data a including a signal S and noise N, and a noise sensor and noise that collect noise data n including only the noise N. A data collection device calculates the amplitude A (ω) and phase φ (ω) of the component of frequency ω included in the observation data a, and the amplitude N (ω) and phase of the component of frequency ω included in the noise data n Frequency component acquisition means for obtaining ψ (ω), and the amplitude S (ω) of the component of frequency ω of the signal S based on the amplitude A (ω), the phase φ (ω), and the phase ψ (ω) There is provided a measuring apparatus comprising: a signal amplitude calculation unit to be obtained; and a signal phase calculation unit to obtain a phase θ (ω) of a component of the frequency ω of the signal S based on the phase ψ (ω). .
In the measurement apparatus according to the second aspect, the signal component calculation apparatus according to the first aspect is used, so that the calculation is simpler than before.

  According to the signal component calculation device and the measurement device of the present invention, there are noise sources having different frequency bands and capable of calculating the signal component with high accuracy from simple observation data in which the signal component and the noise component are mixed. Can properly calculate the signal component.

  Hereinafter, the present invention will be described in more detail with reference to embodiments shown in the drawings. Note that the present invention is not limited thereby.

FIG. 1 is an explanatory diagram of a measurement apparatus 100 according to the first embodiment.
The measuring apparatus 100 includes an observation sensor 1a and an observation data collection apparatus 2a that collect observation data a (t) including a signal S and noise N, and noise that collects noise data n (t) including only noise N. The sensor 1n, the noise data collecting device 2n, the amplitude A (ω) and the phase φ (ω) of the component of the frequency ω included in the observation data a, and the amplitude N (ω of the component of the frequency ω included in the noise data n are obtained. ) And the phase ψ (ω), the amplitude S (ω) of the component of the frequency ω of the signal S is obtained based on the amplitude A (ω), the phase φ (ω), and the phase ψ (ω), and the phase ψ (ω ) To obtain the phase θ (ω) of the frequency ω component of the signal S.

  The observation sensor 1a and the noise sensor 1n are sensors such as radio waves, magnetism, light, and vibration.

FIG. 2 is a flowchart showing processing in the signal component calculation apparatus 10.
In step R1, the observation data a (t) is read from the observation data collection device 2a.
In step R2, the observed data a (t) is subjected to Fourier transform to obtain the amplitude A (ω) and phase φ (ω) of the component of the frequency ω included in the observed data a (t). The phase angle is an angle viewed counterclockwise from an appropriate reference, and is 0 or more and less than 2π.
In step R3, noise data n (t) is read from the noise data collecting device 2n.
In step R4, Fourier transform is performed on the noise data n (t) to obtain the amplitude N (ω) and phase ψ (ω) of the component of the frequency ω included in the noise data n (t).

  In step R5, it is determined whether or not the ratio A (ω) / N (ω) is 4 or more for a component of a certain frequency ω. If it is 4 or more, the process proceeds to step R6, and if not, the process proceeds to step R7.

  In step R6, S (ω) = A (ω) · cos (φ (ω) −ψ (ω)), θ (ω) = ψ (ω) + π / 2, and the amplitude of the component of the frequency ω of the signal S. S (ω) and phase θ (ω) are obtained.

  In step R7, the amplitude S (ω) and phase θ (ω) of the component of the frequency ω of the signal S are obtained from S (ω) = A (ω) and θ (ω) = φ (ω). Then, the process proceeds to Step R8.

In step R8, steps R5 to R8 are repeated for the necessary frequency ω. For example, steps R5 to R8 are repeated for ω corresponding to 20 Hz, 21 Hz, 22 Hz,..., 199 Hz, 200 Hz, which are main frequencies of the brain magnetic field.
In step R9, signal data s (t) is calculated from S (t) = sum_ω {S (ω) · sin (ωt + θ (ω))}. Here, sum_ω {} means that a sum is taken for ω.
Alternatively, signal data s (t) is calculated by performing inverse Fourier transform on S (ω) and θ (ω).

  According to the measurement apparatus 100 of the first embodiment, signal data can be accurately calculated from observation data in which a signal component and a noise component are mixed with simple calculation, and even if there are noise sources having different frequency bands, the measurement apparatus 100 can appropriately calculate the signal data. Signal data can be calculated.

  The signal component calculation device and the measurement device of the present invention can be used to extract a signal component from observation data in which a signal component and a noise component are mixed.

1 is a configuration explanatory diagram of a measuring apparatus according to Example 1. FIG. FIG. 3 is a flowchart illustrating the operation of the signal component calculation apparatus according to the first embodiment. It is a vector diagram which shows the relationship between an observation vector, a noise vector, a weighted noise vector, an error vector, and a signal vector. It is a vector diagram which shows the relationship between an observation vector, a noise vector, a weighted noise vector, and an error vector.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1a Observation sensor 1n Noise sensor 2a Observation data collection device 2n Noise data collection device 10 Signal component calculation device 100 Measurement device

Claims (2)

  The amplitude A (ω) and the phase φ (ω) of the component of the frequency ω included in the observation data a including the signal S and the noise N are obtained, and the amplitude N of the component of the frequency ω included in the noise data n including only the noise N is obtained. frequency component acquisition means for obtaining (ω) and phase ψ (ω), and amplitude of the component of frequency ω of the signal S based on the amplitude A (ω), phase φ (ω), and phase ψ (ω) Signal amplitude calculating means for obtaining S (ω) and signal phase calculating means for obtaining a phase θ (ω) of a component of frequency ω of the signal S based on the phase ψ (ω) Signal component calculation device. An observation sensor and an observation data collection device for collecting observation data a including a signal S and noise N, a noise sensor and a noise data collection device for collecting noise data n including only the noise N, and the observation data a. Frequency component acquisition means for determining the amplitude A (ω) and phase φ (ω) of the included component of frequency ω and determining the amplitude N (ω) and phase ψ (ω) of the component of frequency ω included in the noise data n A signal amplitude calculating means for obtaining an amplitude S (ω) of a component of frequency ω of the signal S based on the amplitude A (ω), the phase φ (ω), and the phase ψ (ω), and the phase ψ A measuring device comprising: a signal phase calculating means for obtaining a phase θ (ω) of a component of frequency ω of the signal S based on (ω).

Images