JPS624971B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Background of the Invention] Technical Field The present invention relates to a period measuring device that measures the period of an electrical signal obtained by converting a fetal heartbeat signal using a transducer at the time of a biological signal.

Prior Art Conventionally, as a method for measuring the period of a biological signal, a period measurement method using a correlation method is known in which an autocorrelation function of the biological signal is determined and the period is measured from this autocorrelation function.

The period measurement method using the correlation method is to sample the electrical signal representing the biological signal at an appropriate sampling period, calculate the autocorrelation function of the biological signal from the sampled data, and calculate the peak of the biological signal from the calculated autocorrelation function. It detects and calculates the period.

Incidentally, period measurement using the correlation method is based on the fact that if the autocorrelation function of the biological signal is determined, a peak of the biological signal can be obtained near the period of the heartbeat.
However, in general, there are often small peaks due to vibration components etc. in the vicinity of the true peak corresponding to the period. Additionally, noise components may be added from the outside when detecting biological signals.

In order to measure the correct period of the biological signal, it is necessary to detect the true peak corresponding to the period of the biological signal among the plurality of peaks due to these components.

In addition, when measuring the period using the conventional correlation function method, the autocorrelation function over the entire period range of the biological signal is calculated first, and the peak indicating periodicity is detected from this autocorrelation function to measure the period of the biological signal. method is used.

Problems with the Prior Art However, such a method requires a large capacity memory for storing the autocorrelation function. In addition, since the autocorrelation function is calculated over the entire period range of the biological signal, if the period of the biological signal is small, there is a possibility that a large amount of meaningless calculation work will be performed, which is difficult from a real-time processing perspective. It was not desirable.

[Purpose of the invention]

This invention has been made in view of the above-mentioned circumstances, and its purpose is to reduce the required storage capacity of a memory for storing the autocorrelation function of biological signals, and to It is an object of the present invention to provide a period measuring device that can detect the true peak corresponding to the period of a biological signal and measure the correct period.

According to the present invention, there is provided a sampling means for sampling an electrical signal representing a biological signal at a predetermined sampling period, and an autocorrelation function for calculating an autocorrelation function of the electrical signal using data obtained by the sampling means. function calculation means; peak detection means for detecting peaks of the calculated autocorrelation function and storing the detected peaks; and comparing peaks occurring after the peaks stored in the peak storage means with the stored peaks. and a peak comparing means for generating an output signal when a later generated peak is large; and an updating means for updating the peak stored in the peak storage means by the output signal of the comparing means. a peak detection means; and a timekeeping means that restarts timekeeping every time it receives an output signal from the comparison means of the peak detection means, and outputs a signal when a time corresponding to a substantially minimum value of the measurement cycle is counted; It is characterized by comprising period calculation means for receiving the output signal from the time measurement means and calculating the time corresponding to the value of the phase difference variable determined by the position of the peak at that time on the time axis. A period measuring device is provided.

It is preferable that the peak detection means includes level setting means for setting a peak detection level based on the peak value when the timer outputs the signal, and detects a peak exceeding this level.

[Specific description of the invention]

Hereinafter, the case where the apparatus according to the embodiment of the present invention is used for measuring the cycle of a fetal heartbeat signal will be explained with reference to FIGS. 1 to 5.

By the way, biological signals are sampled at a constant sampling period T.
If the data at each sampling time obtained by sampling at _s is f(k) (k = 1, 2, 3...n), the autocorrelation function A ₍ 〓 ₎ of this biological signal is (1)
It is expressed by the formula.

In equation (1), τ is a phase difference variable that gives a phase difference to an electrical signal representing a biological signal, and k is a sampling ordinal number. n is the number of times of sampling in one sampling cycle, and one sampling cycle refers to the process of calculating one autocorrelation function value at a certain value of the phase difference variable τ by sampling n times.

When equation (1) is expanded, it can be expressed as equation (2).

A ₍ 〓 ₎ = 1/n {f ₍₁₎ f (1 + τ) + f ₍₂₎ f (2 + τ) + f ₍₃₎ f (3 + τ) +...+f _(o) f (n + τ)} ...(2) (2 ) formula is the sampling data f(k) and f at two times that are shifted on the time axis by the phase difference variable τ.
This means that the autocorrelation function of the biological signal can be obtained by adding the product f(k)f(k+τ) with (k+τ) n times. That is, to explain with reference to Fig. 1, the autocorrelation function A (n) when the phase difference variable τ is set to the value m is the autocorrelation function A _(n) of data sampled at intervals of sampling period T _s on the time axis. Among them, two sampling data that are shifted by the time corresponding to the phase difference variable τ = m on the time axis, for example, f ₍₁₎ and f (m + 1), f ₍₂₎ and f (m + 2), ...
The product of f(n+m) and f _(o) ...etc., f ₍₁₎ f(m+
1), f ₍₂₎ f (m+2)...f _(o) f (m+n)..., etc., are added n times.

Note that here, f ₍₁₎ is the latest data.

A case in which such an autocorrelation function calculation form is applied to a fetal heartbeat signal will be described with reference to FIG. 2.

The period of the fetal heartbeat signal is approximately 300ms.
It is in the range of 1500ms. Therefore, in order to calculate the autocorrelation function over the entire period range of the electrical signal representing the heartbeat signal, it is necessary to set the minimum measurement period to 300ms and the phase difference to a maximum of 1500ms to calculate the autocorrelation function. . That is, the phase difference variable τ
It is necessary to vary the value between 300/ _Ts and 1500/ _Ts . When the phase difference variable τ is set to the period T of the heartbeat signal within this range, that is, m=
Since the autocorrelation function has the maximum peak when T is determined, the period of the electrical signal representing the heartbeat signal can be determined by detecting this maximum peak.

As mentioned above, since the minimum period of a fetal heartbeat signal is usually approximately 300 ms, the calculation of the autocorrelation function starts from the practical minimum value of the measurement period, ie, 300 ms. That is, an autocorrelation function for a period of 300 ms is determined in the first sampling cycle. In this case, the phase difference variable τ is determined by τ=300/T _s , and if the sampling period T _s is set to 5 ms, the phase difference variable τ becomes 60. In calculating the autocorrelation function, first, the heartbeat signal is sampled with a sampling period T _s = 5ms, and the data obtained by each sampling is
Store f ₍₁₎ , f ₍₂₎ , f ₍₃₎ , ... f _(o) in memory. Using these data, the phase difference variable τ=
Find the autocorrelation function A(60) for 60. The autocorrelation function A(60) is determined by the method described with reference to FIG. That is, the phase difference variable τ=60
Two data f(k) and f(k+60) at two sampling times shifted by 1, for example, f ₍₁₎ and f(1
+60), f ₍₂₎ and f (2 + 60), ..., and calculate the sum of these products. Autocorrelation function A(60) when the phase difference variable τ is set to 60 in this way
seek. The value of A(60) is stored in memory for comparison until the autocorrelation function for the next sampling cycle is obtained.

Next, in the second sampling cycle, the value of the phase difference variable τ is advanced by "1" and the autocorrelation function calculation for A(61) is performed. The autocorrelation function calculation for A(61) is substantially the same as that for A(60), so it will be omitted. The autocorrelation function A(61) is compared with the previously calculated and stored autocorrelation function A(60). The state of change between two consecutive sampling cycles is investigated from the comparison results. If there is a state change from increase to decrease between two sampling cycles, then there is a peak in the sampling cycle.

In the above-mentioned method, the autocorrelation function for a certain value of the phase difference variable τ is calculated during one sampling cycle, the calculation result is stored in memory, and "1" is calculated in the next sampling cycle.
Since the peak is detected by comparing the calculation result of the autocorrelation function with the phase difference variable τ advanced by can be reduced.

By the way, as mentioned above, among the peaks obtained from the autocorrelation function, in addition to the original peak corresponding to the period of the signal, peaks that tend to exist in the vicinity are also detected. In order to measure the correct period, it is necessary to detect the original peak corresponding to the period, that is, the true peak, from among these peaks.

The conditions for detecting a peak as a potential true peak are to set a minimum level as a threshold and perform a level check based on this minimum level, and if a peak is detected, the peak detection After that, the autocorrelation function continues to be calculated for the time corresponding to the minimum measurement period, and the calculated autocorrelation function value is compared with the stored autocorrelation function value, and the calculation is performed for the time corresponding to the minimum measurement period. When no peak is detected, that peak is output as a true peak detection signal.

In the level check, a certain threshold value is set for a signal to be detected as a peak, and whether or not the peak is to be detected is determined depending on whether the peak exceeds this threshold value.

The optimal value for this threshold value is selected depending on the signal state at that time. Generally, the true peak value indicating the period is affected by the signal strength and waveform, but noise is a particular problem. In other words, the less noise there is, the more clearly the true peak appears, but
When there is a lot of noise, it becomes difficult to identify which is the true peak.

For this reason, when detecting a peak, it is necessary to set a threshold value according to the state of the signal at that time.

As an example of this, in this embodiment, the peak adopted in the previous measurement, that is, the latest true peak is set as a reference, and a value that is, for example, 1/2 of that is set as the threshold value.

To explain with reference to FIG. 3, assume that peak P ₁ is detected at current time t ₁ . When the peak P ₁ is detected, first, a level L ₁ which is 1/2 of the peak P ₁ is set as a threshold value. Once level L ₁ is set, only peaks that exceed level L ₁ will be detected thereafter. In the example shown in the figure, peak P ₂ exists at time t ₂ , but since this peak P ₂ is below the level L ₁ set as the threshold at that time, it is not considered a peak that could be a true peak. Not detected. Next, a peak _P3 exists at time _t3 . Since this peak P ₃ exceeds the level L ₁ and is larger than the previous peak P ₁ , it is detected as a peak that has the possibility of becoming a true peak. When the peak P ₃ is detected, the threshold value is changed, and a level L ₃ that is 1/2 of the level of the peak P ₃ is set as a new threshold value. Level L ₃
After is set, peaks detected thereafter are limited to those exceeding level _L3 . Thereafter, a peak exceeding the new threshold level is detected in the same manner, and each time the peak exceeds the previous threshold level, the threshold value is changed, and only peaks exceeding the changed threshold value are detected.

In this way, detection of false peaks can be prevented. Note that instead of setting a new threshold value each time a peak exceeds the threshold value, it is also possible to use the threshold value determined by the peak at the time of the previous cycle detection as it is at the time of the next cycle detection.

By performing such a level check,
After the peak is detected, the calculation of the autocorrelation function is continued for a time corresponding to substantially the minimum value of the measurement period, and the detected peak is compared with the peak detected after it, In addition to checking whether or not a true peak exists, the accuracy of detecting a true peak is further improved.

As described above, the peak obtained from the autocorrelation function includes not only the true peak corresponding to the period but also several peaks that exist near this true peak. In order to measure the correct period, the true peak among these peaks must be detected. Since peaks that exist nearby are generally located quite close to the true peak, after a certain peak is detected, the calculation of the autocorrelation function is continued for a certain period of time, and the peaks detected within that certain period of time are calculated. By checking whether or not a peak of the above magnitude exists, it is possible to prevent a peak that also exists nearby from being detected as a true peak. It is sufficient to set this constant time to the minimum value of the time corresponding to the measurement period, and by setting it to the minimum value, measurement results can be obtained at intervals equal to the true period of the biological signal. Therefore, in the embodiment, the time to continue calculating the autocorrelation function after peak detection is set to 300 ms, which corresponds to substantially the minimum value of the measurement period.

To explain with reference to FIG. 4, as shown in the figure, if the peak P ₁ is detected at the current time t ₁₁ , the autocorrelation function calculation is continued for 300 ms from the detection time t ₁₁ , that is, until the time t ₁₂ . . Then, as shown in Fig. 4, there is a peak at time _t21 .
Suppose a peak P ₂ larger than P ₁ is detected. In this case, peak P ₁ is cleared and discarded, and peak P ₂
As a new peak, autocorrelation function calculation is continued for 300ms, that is, until time t ₂₂ ,
If there is no peak larger than peak P ₂ between them, peak P ₂ is detected as a true peak. In FIG. 4 _, a peak with smaller amplitude than peak P ₂ occurs within 300 ms after time t ₂₁ when peak P 2 is detected, that is, at a certain time t ₃₁ up to time t ₂₂ .
Suppose we get P ₃ . However, the peak _P3 , which has a smaller amplitude than the peak _P2 , is not adopted as a potential peak that can be a true peak. In this way, when 300 ms passes after time t ₂₁ , that is, when time t ₂₂ is reached, the peak P ₂ obtained at time t ₂₁ is detected as a true peak indicating the period.
At this point, the correlation function calculation is finished, and the cycle calculation is performed. The value of the phase difference variable τ when the true peak P ₂ obtained in this way is obtained, that is, τ ₂ , corresponds to the period T, and the period T is T = t ₂ ×
It is given by the formula T _s (T _s is the sampling period). In other words, the position of the true peak P ₂ on the time axis is found by going back 300 ms from time t ₂₂ on the time axis, and the phase difference is determined by the position of the found true peak on the time axis. The time corresponding to the value τ ₂ of the variable τ becomes the period T of the electrical signal representing the biological signal, and the period T is given by T=τ ₂ ×T _s (T _s is the sampling period). The next period measurement starts again from τ=60 (corresponding to a period of 300 ms) and is performed in the same way.

In the manner described above, the correct period of the biological signal is measured. In FIG. 4, τ ₁ and τ ₃ indicate the values of phase difference variables determined by the positions of the peaks P ₁ and P ₃ on the time axis, respectively.

Here, the correlation function calculation starts from 300ms and ends at (biological signal period T + 300ms).
This has important implications in terms of detecting true peaks and outputting measurement results.

First, regarding the detection of true peaks, it is impossible for a true peak to exist below the minimum cycle that the biological signal to be measured can take, and it is also impossible for a true peak to exist at an interval within the minimum cycle. do not have.
Therefore, it is possible to completely eliminate the possibility that the peak confirmed in this way shows a peak with a period twice the true period.

Regarding the output timing of the measurement results, there is an effect that the measurement results can be output in synchronization with the true cycle of the biological signal. That is, the period measurement is started from the minimum period of 300 ms, but on the other hand, the peak confirmation period is set to 300 ms, which is equal to the minimum period, so that the measurement results can be output at time intervals equal to the true period. For example, if the true cycle is 500ms, the measurement results will also be output every 500ms. When the period changes, the output interval also changes accordingly. This means that if the correlation function calculation interval matches the data sampling period, the correlation function calculation will proceed in real time on the time axis, that is, from the minimum period of the signal to the sum of this minimum period and the true period. This is because the correlation calculation for the time up to the time represented by is performed within a time equal to the minimum period for checking whether or not it is a true peak.

FIG. 5 schematically shows the configuration of one embodiment of the period measuring device of the present invention for performing the period measurement described with reference to FIGS. 1 to 4.

In the embodiment shown in FIG. 5, the transducer 22 is placed on the abdomen W of a woman and detects the heart rate signal of the fetus. The electrical signal representing the heartbeat signal output from the transducer 22 is given an appropriate waveform shape by the preprocessing circuit 23 coupled thereto, and then processed into a predetermined waveform by the sampling circuit 24, as shown in FIG. A digital signal sampled with a sampling period
f ₍₁₎ , f ₍₂₎ ,...f (m+1), f (m+2)..., f
Analog-to-digital conversion (A
-D conversion).

Data f ₍₁₎ , f ₍₂₎ , ... f (m
+1), f(m+2), . . . are stored in a data memory 26 connected thereto. The data memory 26 is composed of a plurality of shift registers, and each time new data is input, the previously stored data is moved one bit at a time, and the oldest data is pushed out and disappears. data memory 2
6 is connected to a multiplier 28, and the multiplier 28
An adder 30 is connected to. Multiplier 28
is f ₍₁₎・f when the value of the phase difference variable τ is m
(m+1), f ₍₂₎・f(m+2), ... is performed to obtain multiple products. The products obtained by the multiplier 28 are f ₍₁₎・f(m+1), f ₍₂₎・f(m+2),
… is f ₍₁₎・f(m+1)+f ₍₂₎・f by the adder
(m+2)... is added, and the sum is obtained. This sum represents the autocorrelation function A _(n) when the phase difference variable τ is m. That is, the multiplier 28 and the adder 30 constitute a calculation circuit for calculating the autocorrelation function, and as described above, using the data stored in the data memory 26, substantially (2) Calculate the formula and calculate the autocorrelation function. That is, an autocorrelation function is calculated for the phase difference variable τ set to a certain value for each sampling cycle. A peak detector 32 is connected to the adder 30.
2 has a small capacity storage means and a comparison means (see FIG. 6 for the detailed structure). The storage means of the peak detector 32 stores the autocorrelation function calculated value before one sampling cycle, for example, A(m-
1), is stored, and the comparison means compares this stored calculated value with the newly input calculated value A _{(n) of the next sampling cycle, and also compares the newly input calculated value A (n).} If the value A _(n) is larger than the previously stored value A (m-1), the newly input calculated value A _(n) is stored. Since it is sufficient for the peak detector 32 to store only the maximum value among the autocorrelation function calculation values, its storage means may have a small capacity.
Compare the calculated values of the autocorrelation function in two consecutive sampling cycles and check the increase/decrease state. If a trend from increasing to decreasing is observed from the comparison of the two consecutive calculated values, a peak exists in the first sampling cycle. I will do it. The peak detector 32 includes a peak level check circuit 34.
is connected. Peak level check circuit 3
Reference numeral 4 is configured with a comparator as a main element, and a peak detection signal from the peak detector 32 and a reference signal indicating a reference level are inputted thereto. The peak level check circuit 34 then compares the level of the peak detected by the peak detector 32 with a reference level. As the reference level, for example, as described above, a level that is 1/2 of the true peak measured last time is set. If the detected peak exceeds the reference level, the peak level check circuit 34 outputs a signal. Each time the timing means receives both the output signal from the peak detector 32 and the output signal from the peak level check circuit 34, it is reset and restarts timing, and measures the time corresponding to substantially the minimum value of the measurement cycle. Outputs a signal when
The peak confirmation circuit 36 mainly includes a counter which is a timekeeping means for performing such timekeeping, and each time it receives both the output signal from the peak detector 32 and the output signal from the peak level check circuit 34. When the timer is reset and the timer restarts, and the time corresponding to the minimum value of the measurement period (300ms in this example) is measured, the detected peak is determined to be the true peak and the true peak detection signal is output. Output. The reason for setting the counting operation time to 300 ms is that, as mentioned earlier, the minimum heartbeat signal period of a fetus is usually approximately 300ms, and it is impossible for a true peak to exist within the minimum heartbeat signal period, so the detected peak To check whether is a true peak or not, use
This is because 300ms is sufficient.

The peak detector 32, peak level check circuit 34, and peak confirmation circuit 36 are constructed of, for example, a circuit as shown in FIG. The memory 52 is connected to the adder 30 under the control of the write signal from the comparator 54.
The autocorrelation function calculated value is read and stored. Comparator 54 compares the new autocorrelation function calculation value A _(n) from adder 30 with the autocorrelation function calculation value previously stored in memory 52, for example A(m-1). If the new autocorrelation function calculation value A _(n) is larger, the contents of the memory 52 are updated to this new autocorrelation function calculation value A _(n) by the output signal from the comparator 54. The output signal of comparator 54 is also AND
It is input to a counter 62 via a gate 60.
The calculated autocorrelation function value A _(n) stored in the memory 52 is input to a comparator 56 for checking the peak level, and the comparator 56 compares it with the reference level input from the reference level generator 58. be compared. The reference level generator 58 obtains, for example, a half level of the true peak in the previous measurement, and outputs this level as a reference level. As a result of the comparison, if the level of the calculated autocorrelation function value A _(n) stored in the memory 52 exceeds the reference level, the comparator 56 outputs a signal. The output signals of comparators 54 and 56 are
It is input to AND gate 60. Comparator 54,5
AND gate 6 when both output signals of 6 are present
0 outputs a signal. The output signal of the AND gate 60 is input as a reset signal to a counter 62 constituting a peak confirmation circuit. The counter 62 is
If the clock is restarted every time the output signal of the AND gate 60 is received, and the time corresponding to the minimum value of the measurement period (300 ms in this embodiment) is counted, specifically, the number of clock pulses corresponding to 300 ms is counted. When counted, it outputs a signal. The output signal of the counter 62 is input as a true peak detection signal to the period calculation circuit 38 as a period calculation means, and is also input to the memory 52 as a memory clear signal.

With the configuration and operation described above, peak detection and peak confirmation are performed, and the true peak is detected.

When the period calculation circuit 38 connected to the peak confirmation circuit 36 receives the output signal from the peak confirmation circuit 36, that is, the true peak detection signal, it calculates a phase difference variable τ determined by the position of the peak on the time axis at that time. Calculate the time corresponding to the value of. This time corresponds to the period T of the electrical signal representing the heartbeat signal. In other words, the value of the phase difference variable τ when the true peak is obtained corresponds to the period T, and therefore the period T is the phase difference determined by the position of the true peak on the time axis. It is expressed as a time corresponding to the value of the variable τ. A heart rate calculation circuit 40 connected to the period calculation circuit 38 calculates the heart rate based on the calculated period. Heart rate calculation circuit 40
is connected to the control circuit 42. Control circuit 42
A display device 44 including, for example, a light emitting diode (LED) is connected to the display device 44 . The display 44 displays the heart rate of the heartbeat signal based on the signal output from the heart rate calculation circuit 40 via the control circuit 42.
At this time, the control circuit 42 includes a heart rate calculation circuit 40.
If the signal from the heart rate calculation circuit 40 contains a noise component or if the probe is dislodged, the signal from the heart rate calculation circuit 40 is controlled not to be input to the display 44 to prevent false heart rate display. It is a good idea to have auxiliary means to do so. However, since this is not directly related to this invention, detailed explanation will be omitted.

The control circuit 42 also outputs a clock pulse to the sampling circuit 24 to
The sampling timing in 4 is controlled. The control circuit 42 further outputs a signal instructing the value of the phase difference variable τ to the multiplier 28 for each sampling. The value of the phase difference variable τ starts from a value corresponding to a time substantially corresponding to the minimum value of the period of the heartbeat signal, e.g. Accordingly, τ=m, τ=(m+1), τ=
(m+2), . . . and so on. The multiplier 28 calculates two data that are separated by the value of the phase difference variable τ indicated by the signal from the control circuit 42, for example, f ₍₁₎ and f (m+1) when the value of the phase difference variable τ is m. Read from the data memory 26,
Find the product f ₍₁₎・f(m+1). Control circuit 4
2 outputs a timing signal to the adder 30, and the adder 30 outputs the calculation results sequentially calculated by the multiplier 28 based on the timing signal from the control circuit 42.
Add f ₍₁₎・f(m+1), f ₍₂₎・f(m+2), etc. That is, multiplier 28 and adder 30
Under the control of the control circuit 42, the data is read from the data memory and the autocorrelation function A _(n) shown in equation (2) is substantially calculated.

The control circuit 42 further includes a reference level detector 46.
is connected. The reference level detector 46 detects the optimum reference level (zero level) for coding the sampled data according to the timing signal output from the control circuit 42 at appropriate time intervals, and detects the optimum reference level (zero level) for coding the sampled data. 24
Outputs an optimum reference level indication signal to When coding data, the better the balance between positive and negative data is maintained, the more reliable the obtained autocorrelation function represents periodicity. During sampling, the maximum value, minimum value, or average value of data is detected to determine the optimal value of the reference level.

As described above, according to the present invention, the calculation of the autocorrelation function is started from the substantially minimum value of the measurement period of the biological signal on the time axis, and even after the peak is detected, the calculation of the autocorrelation function is started from the substantially minimum value of the measurement period of the biological signal. The calculation of the autocorrelation function is continued for a period of time, and when there is no peak larger than the first peak within the time period corresponding to the minimum value after the detection of the peak, the detected peak is detected as a true peak. By this,
Not only is it possible to reliably detect only the true peak indicating the period of the original signal, thereby making it possible to measure the correct period, but also the autocorrelation function calculation range is from the above-mentioned practical minimum value to this minimum value. It is limited to the maximum value to be measured, and calculations below the minimum value are unnecessary, meaningless calculations are eliminated and it is effective for real-time processing. A period measuring device capable of outputting measurement results can be provided.

[Brief explanation of the drawing]

FIG. 1 is a diagram used to explain the calculation form of the autocorrelation function in the period measuring device according to the present invention, and FIG. 2 shows the case where the period measuring device according to the present invention is applied to periodic measurement of fetal heart rate. FIG. 3 is a diagram showing a fetal heartbeat signal for explanation, and FIG. Figure 5 is a diagram for explaining the operation of continuing to calculate an autocorrelation function for a certain period of time after detecting a peak in order to confirm whether the peak detected by the apparatus of the present invention is a true peak. The figure is a block diagram schematically showing the configuration of one embodiment of the period measuring device of the present invention, and FIG. 6 is a peak detector, peak level check circuit, and peak confirmation circuit in the period measuring device shown in FIG. FIG. 23... Preprocessing circuit, 24... Sampling circuit, 26... Data memory, 28... Multiplier, 3
0... Adder, 32... Peak detector, 34...
Peak level check circuit, 36... Peak confirmation circuit, 38... Cycle calculation circuit, 40... Heart rate calculation circuit, 42... Control circuit, 44... Display, 4
6...Reference level detector.

Claims (1)

[Scope of Claims] 1. Sampling means for sampling an electrical signal representing a biological signal at a predetermined sampling period; and a self-correlation function for calculating an autocorrelation function of the electrical signal using data obtained by the sampling means. correlation function calculation means; peak detection means for detecting peaks of the calculated autocorrelation function and storing the detected peaks; and peaks occurring after the peaks stored in the peak storage means as stored peaks. The peak comparing means compares and generates an output signal when a later generated peak is larger, and the updating means updates the peak stored in the peak storage means by the output signal of the comparing means. and a clocking means that restarts timekeeping every time it receives an output signal from the comparison means of the peak detection means and outputs a signal when a time corresponding to a substantially minimum value of the measurement cycle is counted. , period calculation means for receiving the output signal from the time measurement means and calculating the time corresponding to the value of the phase difference variable determined by the position of the peak at that time on the time axis; Characteristic period measuring device. 2. The peak detection means is characterized in that it includes a level setting means for setting a peak detection level based on the peak value when the timer outputs the signal, and detects a peak exceeding this level. , a period measuring device according to claim 1.