Vibration oscillator and signal detection device

Abstract

A fluctuation oscillator (1) is provided with: an adder (11) that has an input terminal (111) into which inputted is an input signal (S1) including a main signal and an uncorrelated signal which is uncorrelated to the main signal and has a higher frequency than the main signal, and that adds a feedback signal (S5) to the input signal (S1); a threshold value discrimination unit (12) for generating a pulse signal (S3) by comparing, against a threshold value, an addition signal (S2) obtained by addition by the adder (11); a transient response unit (13) for causing transient response of the generated pulse signal (S3) and generating an output signal (S4); and a feedback loop (16) for feeding back the output signal (S4) to the adder (11) as the feedback signal (S5).

Description

Technical Field

[0001] The present invention relates to a fluctuation oscillator and a signal detection device including the fluctuation oscillator.

Background Art

[0002] There is known a fluctuation oscillator which is an oscillator that utilizes a stochastic resonance phenomenon in which a signal is enhanced and a reaction is improved under a certain probability of adding noise to the signal. Since the fluctuation oscillator oscillates with fluctuations such as 1 / f fluctuations, it is possible to realize a comfortable control for humans such as blinking an illumination device like a firefly or moving a robot biologically. Further, the fluctuation oscillator can also be used as a sensor for detecting a weak signal buried in noise. Since conventional fluctuation oscillators were composed of analog circuits, it has been required to flexibly adjust circuit parameters such as resistors and capacitors in order to oscillate the fluctuation oscillator at a desired frequency.

[0003] Therefore, Patent Document 1 discloses a fluctuation oscillator in which circuit parameters can be flexibly adjusted. Specifically, Patent Document 1 discloses a noise generator, an adder that adds a noise signal generated by the noise generator and a transient response signal fed back to an input signal, a threshold discriminator that generates a pulse signal by comparing an addition signal output from the adder with a threshold value, a transient response unit that transiently responds to the pulse signal, a feedback loop that feeds back a transient response signal from the transient response unit, and an intensity adjuster that adjusts the intensity of the transient response signal flowing through the feedback loop.

[0004] However, since the oscillator in Patent Document 1 includes a noise generator, there are problems that the circuit scale increases accordingly and the cost rises. Furthermore, since it includes a noise generator, there is also a problem that the power consumption increases. Further, when adjusting the natural frequency of the oscillator, an operation of adjusting circuit parameters such as the threshold value of the threshold discrimination unit and the time constant of the transient response unit is required. However, when there is a noise generator, it is also required to adjust the level of the noise signal generated by the noise generator, and there is a problem that it takes time to adjust the circuit parameters.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

[0006] An object of the present invention is to provide an oscillator that does not require a noise generator.

[0007] An oscillator according to an aspect of the present invention has an input terminal to which an input signal including a main signal and an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal is input, an adder that adds a feedback signal to the input signal, a threshold discrimination unit that generates a pulse signal by comparing the added signal added by the adder with a threshold value, a transient response unit that transiently responds to the generated pulse signal to generate an output signal, and a feedback loop that feeds back the output signal to the adder as the feedback signal.

[0008] According to the present invention, a noise generator can be omitted from the oscillator. Thereby, the circuit scale of the oscillator can be reduced and the cost can be reduced. Furthermore, since there is no noise generator, when adjusting the circuit parameters, the trouble of adjusting the level of the noise signal can be saved.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Embodiments for Carrying Out the Invention

[0010] A signal such as a brain wave signal includes a main signal indicating the brain wave itself and a signal uncorrelated with the main signal. When such a signal is input to the fluctuation oscillator, if the frequency of the uncorrelated signal is higher than the frequency of the main signal, it has been found that the uncorrelated signal acts as a noise signal and stochastic resonance occurs, enabling detection of a weak main signal. In this case, it becomes possible to omit the noise generator from the fluctuation oscillator.

[0011] The present invention has been made based on such findings. Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. Note that the following embodiments are an example of embodying the present invention and do not have the character of limiting the technical scope of the present invention.

[0012] (Embodiment 1) FIG. 1 is a diagram showing an example of the configuration of the fluctuation oscillator 1 in Embodiment 1 of the present invention. As shown in FIG. 1, the fluctuation oscillator 1 includes an adder 11, a threshold discrimination section 12, a transient response section 13, a monitor section 14, an intensity adjustment section 15, and a feedback loop 16.

[0013] The adder 11 has an input terminal 111 to which the input signal S1 is input and a feedback terminal 112 to which the feedback signal S5 is input, and adds the input signal S1 and the feedback signal S5 output from the intensity adjustment unit 15. The adder 11 is constituted by, for example, an analog addition circuit.

[0014] The input signal S1 is a signal including a main signal and an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal. In the present embodiment, an electroencephalogram signal detected by the sensor 20 is adopted as the input signal S1. The electroencephalogram signal is a weak signal including a main signal indicating the electroencephalogram itself and an uncorrelated signal having a higher frequency than the main signal. Therefore, the electroencephalogram signal is a signal suitable for signal detection of the fluctuation oscillator 1. Note that this uncorrelated signal is a noise signal mixed into the electroencephalogram signal due to some factor in the process of measuring the electroencephalogram signal.

[0015] The sensor 20 is an electroencephalogram sensor including a probe that contacts a person's head. Here, an electroencephalogram signal is used as the input signal S1, but this is an example, and any signal may be adopted as long as it is a signal including a main signal and an uncorrelated signal having a higher frequency than the main signal. For example, a biological signal can be adopted as the input signal S1.

[0016] In the example of FIG. 1, only one input signal S1 is input to the adder 11, but the present invention is not limited to this, and a plurality of input signals may be input. In this case, the adder 11 may have a plurality of input terminals corresponding to the plurality of input signals.

[0017] The threshold discrimination unit 12 compares the addition signal S2 output from the adder 11 with a predetermined threshold value, and when this addition signal S2 is equal to or greater than the threshold value, outputs a pulse signal S3 that becomes a high level, and when this addition signal S2 is less than the threshold value, outputs a low level. When the pulse signal S3 is at a high level, the fluctuation oscillator 1 becomes a high state, and when the pulse signal S3 is at a low level, the fluctuation oscillator 1 becomes a low state.

[0018] The threshold value may include a first threshold value and a second threshold value greater than the first threshold value. In this case, in the high state, the threshold discriminator 12 may output a low-level pulse signal S3 when the addition signal S2 falls below the first threshold value. As a result, the fluctuation oscillator 1 switches from the high state to the low state. Also, in the low state, the threshold discriminator 12 may output a high-level pulse signal S3 when the addition signal S2 exceeds the second threshold value. As a result, the fluctuation oscillator 1 switches from the low state to the high state. By configuring the threshold value with the first threshold value and the second threshold value, it becomes possible to give the fluctuation oscillator 1 hysteresis. Thus, the configuration for giving the fluctuation oscillator 1 hysteresis can be easily realized by configuring the threshold discriminator 12 with a Schmitt trigger circuit.

[0019] The transient response unit 13 transiently responds to the pulse signal S3 output from the threshold discriminator 12 and generates a transient response signal of the pulse signal S3 as the output signal S4. The transient response unit 13 is composed of, for example, an integrator or a differentiator. Alternatively, the transient response unit 13 may be composed of an LCR circuit. The LCR circuit is a circuit in which a coil (L), a capacitor (C), and a resistor (R) are connected in series or in parallel. In this case, an output signal S4 having an LCR transient response waveform that changes logarithmically is output from the transient response unit 13.

[0020] The monitor unit 14 monitors the output signal S4. In the present embodiment, the monitor unit 14 is composed of, for example, an information processing device such as a computer, and displays the waveform of the output signal S4 on a display, detects the frequency of the output signal S4, and displays it on the display. As a result, the operator can grasp the input signal from the information displayed on the display.

[0021] The intensity adjustment unit 15 is composed of a variable resistor provided on the feedback loop 16, adjusts the intensity of the feedback signal S5, and inputs it to the adder 11.

[0022] The feedback loop 16 is composed of a circuit provided between the output terminal 151 and the feedback terminal 112, and feeds back the output signal S4 to the adder 11 as the feedback signal S5.

[0023] The output terminal 151 is provided on the output side of the monitor unit 14 and outputs the output signal S4 to the outside.

[0024] In the present embodiment, the fluctuation oscillator 1 shown on the left side of FIG. 1 is represented using the symbols on the right side of FIG. 1. Also, in FIG. 1, the fluctuation oscillator 1 includes the monitor unit 14, but the monitor unit 14 may be omitted. In this case, the feedback loop 16 is connected between the output terminal of the transient response unit 13 and the feedback terminal 112. The intensity adjustment unit 15 may be omitted.

[0025] The fluctuation oscillator 1 shown in FIG. 1 operates as follows. An input signal S1 including an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal is input to the adder 11. This input signal S1 is added to the feedback signal S5 by the adder 11 and input to the threshold discrimination unit 12. In the threshold discrimination unit 12, the added signal S2 output from the adder 11 is compared with the threshold value, and a pulse signal S3 indicating the comparison result is generated. Since the added signal S2 input to the threshold discrimination unit 12 includes an uncorrelated signal, this uncorrelated signal acts as a noise signal, and even if the main signal included in the input signal S1 is lower than the threshold value, the added signal S2 can probabilistically exceed the threshold value by stochastic resonance. The pulse signal S3 output from the threshold discrimination unit 12 is shaped into an output signal S4 having a transient response waveform according to the time constant of the transient response unit 13. This output signal S4 is fed back to the adder 11 by the feedback loop 16. Due to this feedback, eventually, the fluctuation oscillator 1 oscillates in synchronization with the main signal. Thereby, a weak main signal is accurately detected.

[0026] Figure 2 is a waveform diagram of the output signal S4 when the transient response section 13 is configured by a differentiator. In Figure 2, the vertical axis represents voltage and the horizontal axis represents time. At time T1, the addition signal S2 exceeds the threshold value and the pulse signal S3 becomes high level. Therefore, the output signal S4 immediately rises to the positive power supply voltage VDD and then decays toward the ground level GND according to the time constant of the differentiator.

[0027] At time T2, the addition signal S2 falls below the threshold value and the pulse signal S3 becomes low level. Therefore, the output signal S4 immediately drops to the negative power supply voltage VSS and then increases according to the time constant of the differentiator. Thereafter, the fluctuation oscillator 1 repeats this behavior to oscillate.

[0028] Figure 3 is a waveform diagram of the output signal S4 when the transient response section 13 is configured by an integrator. In Figure 3, the vertical axis represents voltage and the horizontal axis represents time. At time T1, the addition signal S2 exceeds the threshold value and the pulse signal S3 becomes high level. Therefore, the output signal S4 increases according to the time constant of the integrator.

[0029] At time T2, the addition signal S2 falls below the threshold value and the pulse signal S3 becomes low level. Therefore, the output signal S4 decays according to the time constant of the integrator. Thereafter, the fluctuation oscillator 1 repeats this behavior to oscillate.

[0030] As shown in Figure 3, compared with the differentiator, the integrator can change the output signal S4 smoothly. Therefore, when it is desired to make the light-emitting element emit light in a smooth light-emitting pattern or to make the controlled object operate in a smooth operation pattern, an integrator may be adopted as the transient response section 13. On the other hand, when it is desired to make the light-emitting element emit light in a light-emitting pattern with a rapid change in dimming, or to make the controlled object operate in an operation pattern with a rapid change in operation, a differentiator may be adopted as the transient response section 13.

[0031] Figure 4 is a graph showing the frequency spectrum when the brain wave signal of a person is subjected to fast Fourier transform. In Figure 4, (a) shows the frequency spectrum 401 when the method of the comparative example is used, and (b) shows the frequency spectrum 402 when the method of the present case is used.

[0032] The method of the comparative example is a method of observing a peak from the frequency spectrum 401 obtained by subjecting the brain wave signal detected by the sensor 20 to fast Fourier transform. The method of the present case is a method of subjecting the output signal S4 obtained by inputting the brain wave signal detected by the sensor 20 to the fluctuation oscillator 1 to fast Fourier transform and observing a peak from the obtained frequency spectrum 402. Note that each of the frequency spectra 401 and 402 is a double logarithmic graph, the vertical axis indicates the intensity of the brain wave signal, and the horizontal axis indicates the frequency.

[0033] Peaks are observed in both the frequency spectra 401 and 402, but it can be seen that the peak in the frequency spectrum 402 is steeper than that in the frequency spectrum 401. Therefore, it can be seen that the method of the present case can observe the peak more accurately than the method of the comparative example.

[0034] In the method of the comparative example, since the peak does not appear prominently in the frequency spectrum, it is required to perform signal processing such as filtering processing on the fast Fourier transformed signal to make the peak appear. On the other hand, in the method of the present case, since the peak appears prominently, signal processing such as filtering processing is not required, and the simplification of the processing can be achieved.

[0035] Note that in both the frequency spectra 401 and 402, the high-frequency side from the peak decreases linearly overall, and it is observed that the brain wave signal has the characteristics of 1 / f fluctuation.

[0036] Figure 5 is a graph comparing the peak frequency of the brain wave signal detected by the method of the present case with the peak frequency of the brain wave signal detected by the method of the comparative example. Note that the peak frequency refers to the frequency corresponding to the peak in the frequency spectra 401 and 402 shown in Figure 4.

[0037] In FIG. 5, the vertical axis represents the peak frequency observed by the method of the present case, and the horizontal axis represents the peak frequency observed by the method of the comparative example. In the method of the present case, the natural frequency of the oscillator 1 is set to 1 Hz, which is lower than the frequency band of the electroencephalogram.

[0038] For example, for the observation point P1, the value on the vertical axis is approximately 19 Hz, while the value on the horizontal axis is approximately 10 Hz. Since there is a deviation in the peak frequencies observed by the method of the present case and the method of the comparative example, it is impossible to determine which method is accurate. On the other hand, for an observation point located on the line L50 at a 45-degree angle, such as the observation point P2, the peak frequencies observed by the method of the present case and the method of the comparative example are approximately the same, and it can be determined that the peak frequencies can be detected by both methods.

[0039] The points to note in this graph are the observation points belonging to the regions A1 to A3 surrounded by the dotted circles. For example, the observation points belonging to the region A1 have values on the vertical axis in the range of approximately 3.5 Hz to 9 Hz, and can be inferred to be observation points of theta waves. However, the value on the horizontal axis is about 1 Hz, and it is difficult to say that the peak frequency of the electroencephalogram can be observed. Similarly, the observation points belonging to the region A2 have values on the vertical axis in the range of approximately 8 Hz to 14 Hz, and can be inferred to be a group of observation points of alpha waves. However, the value on the horizontal axis is 1 Hz, and it is difficult to say that the peak frequency of the electroencephalogram can be observed. The same is true for the observation points of beta waves whose values on the vertical axis belonging to the region A3 are in the range of approximately 13 Hz to 28 Hz.

[0040] From the above, it can be seen from the graph shown in FIG. 5 that the peak frequencies of theta waves, alpha waves, and beta waves that could not be detected in the comparative example can be detected by the method of the present case. Therefore, even if the oscillator 1 does not have a noise generator, it is possible to detect weak electroencephalograms that could not be detected by the method of the comparative example.

[0041] The fluctuation oscillator 1 has a natural frequency. The natural frequency is the oscillation frequency of the fluctuation oscillator 1 when only a noise signal is input to the fluctuation oscillator 1. In order to detect a weak signal, it is necessary to set the natural frequency of the fluctuation oscillator 1 lower than the frequency of the input signal. Circuit parameters for adjusting the natural frequency include the threshold value of the threshold discrimination unit, the time constant of the transient response unit, the gain of the intensity adjustment unit, and the level of the noise signal, etc. Among these circuit parameters, the main parameter is the threshold value of the threshold discrimination unit. Hereinafter, taking the case where the electroencephalogram signal is the input signal and the circuit parameter to be adjusted is the threshold value of the threshold discrimination unit as an example, the adjustment operation of the circuit parameters of the conventional fluctuation oscillator will be described. First, an electroencephalogram signal is input to the fluctuation oscillator, and the threshold value of the threshold discrimination unit is adjusted so that the natural frequency is lower than the frequency of the electroencephalogram signal and becomes a frequency suitable for detecting the electroencephalogram signal. Next, the fluctuation oscillator is connected to a noise generator, and the level of the noise signal is adjusted so that the noise signal is input to the fluctuation oscillator and the natural frequency is lower than the frequency of the electroencephalogram and becomes a frequency suitable for detecting the electroencephalogram. Here, the reason why the adjustment of the noise signal level is necessary is that the noise signal generated by the noise generator is further superimposed on the uncorrelated signal acting as the noise signal, so that the threshold value of the threshold discrimination unit deviates from the appropriate value. Thus, in the conventional fluctuation oscillator, the natural frequency has been adjusted mainly through two-stage adjustment.

[0042] On the other hand, in the fluctuation oscillator 1, the uncorrelated signal included in the electroencephalogram signal acts as a noise signal and probabilistic resonance occurs. Therefore, in the fluctuation oscillator 1, mainly the operation of inputting the electroencephalogram signal to the fluctuation oscillator 1 and adjusting the threshold value of the threshold discrimination unit 12 so that the natural frequency is lower than the frequency of the electroencephalogram and becomes a frequency suitable for detecting the electroencephalogram needs to be performed, and then the operation of inputting the noise signal and adjusting the level of the noise signal does not need to be performed. Therefore, in the fluctuation oscillator 1, the labor cost for adjusting the circuit parameters is reduced.

[0043] As described above, according to the present embodiment, since the uncorrelated signal included in the input signal acts as a noise signal and stochastic resonance occurs, there is no need to use a noise generator for generating a noise signal as in the conventional oscillator with fluctuations. Therefore, the noise generator can be omitted from the oscillator with fluctuations. As a result, the circuit scale of the oscillator with fluctuations can be reduced, and cost reduction can be achieved. Furthermore, since there is no noise generator, when adjusting the circuit parameters, the trouble of adjusting the level of the noise signal can be saved.

[0044] (Embodiment 2) Embodiment 2 is characterized in that a signal detection device for detecting an unknown input signal S1 is configured using a plurality of oscillators with fluctuations 1. FIG. 6 is a diagram showing the overall configuration of the signal detection device 2. In the present embodiment, the same components as those in Embodiment 1 are denoted by the same reference numerals, and the description thereof is omitted.

[0045] The signal detection device 2 includes a plurality of oscillators with fluctuations 1 and a detection unit 2000. In the example of FIG. 6, the oscillators with fluctuations 1 are composed of four oscillators with fluctuations 1_1, 1_2, 1_3, and 1_4. However, this is merely an example, and the signal detection device 2 may be configured with a plurality of oscillators with fluctuations 1 other than four.

[0046] The input terminals of the oscillators with fluctuations 1_1 to 1_4 are connected to a common input terminal 2001, and a common input signal S1 is input. The oscillators with fluctuations 1_1 to 1_4 are adjusted in circuit parameters so that their respective natural frequencies are different.

[0047] The detection unit 2000 is connected to the output terminals of the oscillators with fluctuations 1_1 to 1_4 and is composed of an information processing device such as a computer. Then, the detection unit 2000 receives the respective output signals S4 of the oscillators with fluctuations 1_1 to 1_4.

[0048] As shown in Patent Document 1, the synchronization state of the oscillator 1 has incomplete synchronization and complete synchronization. Incomplete synchronization refers to a synchronization state in which the oscillator 1 oscillates with a certain degree of frequency variation. Complete synchronization refers to a synchronization state in which the oscillator 1 oscillates with little frequency variation. Note that for incomplete synchronization and complete synchronization to occur, the oscillator 1 is required to have a natural frequency lower than the frequency of the input signal.

[0049] In incomplete synchronization, when an unknown input signal S1 is input to the oscillator 1, the oscillator 1 oscillates at a frequency shifted from the natural frequency by the frequency shift amount. This frequency shift amount increases as the frequency of the input signal approaches the natural frequency of the oscillator 1.

[0050] On the other hand, in complete synchronization, the frequency (oscillation frequency) of the output signal S4 of the oscillator 1 becomes the same as the frequency of the main signal included in the input signal S1. Therefore, by detecting the frequency of the output signal S4, the frequency of the input signal S1 can be detected.

[0051] For this reason, in the case of incomplete synchronization, when the frequency shift amount is observed, the oscillator 1 can detect that a main signal with a frequency equal to or higher than the natural frequency is input, and when the frequency shift amount is not observed, the oscillator 1 can detect that an input signal S1 with a frequency equal to or higher than the natural frequency is not input.

[0052] Therefore, in the signal detection device 2 of FIG. 6, if all the synchronization states of the oscillators 1_1 to 1_4 are in incomplete synchronization, the frequency of the input signal S1 can be estimated to have the value closest to the natural frequency of the oscillator 1 with the largest frequency shift amount among the oscillators 1_1 to 1_4.

[0053] For example, assume that the natural frequencies of the vibration oscillators 1_1 to 1_4 are "5 Hz", "10 Hz", "15 Hz", and "20 Hz" respectively, and an input signal S1 including a main signal with a frequency of "17 Hz" is input. In this case, for the vibration oscillators 1_1 to 1_3 whose natural frequencies are lower than the frequency of the main signal, a frequency shift amount is observed. Further, in this case, among the vibration oscillators 1_1 to 1_3, the frequency shift amount of the vibration oscillator 1 with a natural frequency of "15 Hz", which is closest to the frequency of the input signal S1, is the largest. Therefore, in the signal detection device 2, when all the synchronization states of the vibration oscillators 1_1 to 1_4 are in incomplete synchronization, it can be estimated that the main signal included in the input signal has a frequency close to the natural frequency of the vibration oscillator 1 with the largest frequency shift amount.

[0054] If the synchronization state of any one of the vibration oscillators 1_1 to 1_4, i.e., the vibration oscillator 1, is in complete synchronization, it can be estimated that the frequency of the main signal is the same as the frequency of the output signal S4 of the completely synchronized vibration oscillator 1.

[0055] As described above, according to the signal detection device 2 of the second embodiment, since it includes a plurality of vibration oscillators 1 having different natural frequencies respectively, even when all the vibration oscillators 1 cannot be completely synchronized with the input signal S1, the frequency of the input signal S1 can be estimated from the natural frequency of the vibration oscillator 1 with the largest frequency shift amount. Also, according to the signal detection device 2 of the second embodiment, when any one of the vibration oscillators 1 is completely synchronized with the input signal S1, the frequency of the input signal S1 can be estimated from the frequency of the output signal S4 of that vibration oscillator 1. (Summary of the embodiment)

[0056] The fluctuation oscillator according to one aspect of the present invention has an input terminal to which an input signal including a main signal and an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal is input, an adder that adds a feedback signal to the input signal, a threshold discrimination unit that generates a pulse signal by comparing the added signal added by the adder with a threshold value, a transient response unit that generates an output signal by subjecting the generated pulse signal to a transient response, and a feedback loop that feeds back the output signal to the adder as the feedback signal.

[0057] According to this configuration, an input signal including a main signal and an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal is input to the adder. This input signal is added to the feedback signal by the adder and input to the threshold discrimination unit. In the threshold discrimination unit, the added signal output from the adder is compared with the threshold value, and a pulse signal indicating the comparison result is generated. Since the added signal input to the threshold discrimination unit includes an uncorrelated signal, this uncorrelated signal acts as a noise signal, and the added signal can probabilistically exceed the threshold value due to stochastic resonance. The pulse signal output from the threshold discrimination unit is shaped into an output signal having a waveform in a transient response state by the transient response unit. This output signal is fed back to the adder by the feedback loop. Due to this feedback, eventually, the fluctuation oscillator oscillates in synchronization with the main signal. As a result, a weak main signal is accurately detected.

[0058] As described above, in this configuration, since stochastic resonance occurs based on the uncorrelated signal included in the input signal, it is not necessary to use a noise generator for generating a noise signal as in a conventional fluctuation oscillator. Therefore, the noise generator can be omitted from the fluctuation oscillator. As a result, the circuit scale of the fluctuation oscillator can be reduced, and cost reduction can be achieved. Furthermore, since there is no noise generator, when adjusting circuit parameters, the trouble of adjusting the level of the noise signal can be saved.

[0059] In the above-described fluctuation oscillator, it is preferable that the input signal is an electroencephalogram signal.

[0060] The electroencephalogram signal is a signal that includes a main signal indicating the electroencephalogram itself and an uncorrelated signal having a higher frequency than the main signal. Therefore, this configuration can accurately detect weak electroencephalograms without using a noise generator.

[0061] In the above-mentioned oscillation generator, it is preferable to have a natural frequency lower than the frequency of the main signal.

[0062] When an input signal including a main signal having a frequency higher than the natural frequency is input to the oscillation generator that feeds back the output signal to the input signal, the frequency shift amount (the frequency shift amount of the output signal with respect to the natural frequency) can be observed. However, when a main signal having a frequency lower than the natural frequency is input, it has the characteristic that the frequency shift amount cannot be observed. Based on this frequency shift amount, it is possible to detect a weak signal. In this configuration, since the natural frequency of the oscillation generator is set to be lower than the frequency of the main signal to be detected, the oscillation generator can detect a weak main signal.

[0063] In this configuration, since the circuit parameters are adjusted so that the natural frequency of the oscillation generator is lower than the frequency of the main signal to be detected, the oscillation generator can detect the main signal. Also, when a main signal having a frequency lower than the natural frequency is input, since the oscillation generator cannot detect the main signal, the oscillation generator can be given the function of a high-pass filter.

[0064] The signal detection device according to another aspect of the present invention includes a detection unit that detects the frequency of the input signal based on the output signals from the respective oscillation generators, where each oscillation generator has a different natural frequency and a common input signal is input.

[0065] When the frequency of the main signal included in the input signal is higher than the natural frequency, the fluctuation oscillator has a characteristic that the frequency shift amount increases as the frequency of the main signal approaches the natural frequency. Therefore, the frequency of the main signal can be estimated to be close to the natural frequency of the fluctuation oscillator with the maximum frequency shift amount. In this configuration, since a plurality of fluctuation oscillators having different natural frequencies are provided, the frequency of the unknown main signal can be detected using the natural frequency of the fluctuation oscillator in which the maximum frequency shift amount is observed.

Claims

1. It has an input terminal to which an input signal including a main signal and an uncorrelated signal that is uncorrelated with the main signal and has a higher frequency than the main signal is input, an adder that adds a feedback signal to the input signal, a threshold discriminator that generates a pulse signal by comparing the added signal added by the adder with a threshold value, a transient response unit that generates an output signal by subjecting the generated pulse signal to a transient response, and a feedback loop that feeds back the output signal to the adder as the feedback signal, wherein the input signal is an electroencephalogram signal, the threshold value is adjusted so that the natural frequency of the fluctuation oscillator is lower than the main signal of the electroencephalogram signal, a fluctuation oscillator.

2. A signal detection device comprising a plurality of the fluctuation oscillators according to Claim 1, each of the fluctuation oscillators has a different natural frequency, and a common input signal is input, A signal detection device comprising a detection unit that detects the frequency of the input signal based on the output signals from the respective fluctuation oscillators.

Images