JP2000023932A - Pattern signal interval detecting device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect the accurate interval between pattern signals, using a pattern signal interval detecting device for detecting the interval between pattern signals having peaks, by making the device environmental resistance to environmental noises, etc., and enhancing the accuracy of determining pattern signals. SOLUTION: When the peak width D (m) between the poles P131 and P132 of the peak P13 of each input signal is within a permissible range, the signals are determined to be predetermined pattern signals, or they are determined as the pattern signals by anticipating the respective permissible values of the slice level Ss and the pattern (peak) interval I (m) of the peak on the basis of the previous values, or they are determined as the pattern signals through pattern matching by preparing a reference pattern. The peak interval between the pattern signals determined is computed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pattern signal interval detecting device, and more particularly to a pattern signal interval detecting device for detecting an interval between pattern signals having a peak.

[0002]

2. Description of the Related Art FIG. 9 shows a configuration example of a general pattern signal interval detecting device. In this example, a device for detecting an interval of a heartbeat fluctuation waveform signal of a human body is particularly shown. The living body electrodes 21 to 23 are attached to the human body 20, and the potential of the skin surface corresponding to the heart activity is amplified by the AC amplifier unit 24 by obtaining the difference voltage between the electrodes 21 and 23 and the electrodes 22 and 23, and the differential output is obtained. The voltages are sent to the A / D conversion unit 25 as the voltages A and B, where they are converted into digital signals and then given to the calculation unit 26. The calculation unit 26 performs a predetermined calculation process and outputs an electrocardiogram signal and its interval signal.

In such a pattern signal interval detecting device, a known pattern signal such as an electrocardiogram, a pulse wave and respiration is measured by an electrophysiological index converted into a voltage signal, and the voltage signal is determined by a predetermined reference voltage. Find the peak time beyond. Then, a method of detecting an interval between pattern signals using this time as an interval delimiter is generally used.

A problem in detecting an interval is a slow and large voltage fluctuation in an electrophysiological index, that is, a baseline change. For a method to eliminate the effect of this baseline change, use a filter,
A number of methods have been proposed, such as Japanese Patent Laid-Open No. 5-56942 and Japanese Patent Laid-Open No. 5-220121, in which a gain is automatically adjusted to eliminate it.

In addition, the value of the interval between signal peaks and the slice level, which is the reference voltage for detecting the peak voltage used as the reference for detecting the break, are defined by the ratio to the immediately preceding data.
There are many conventional examples of methods for defining the upper and lower limits of the peak interval value.

[0006]

FIGS. 10 and 11 show examples of signal waveforms that may be erroneously detected by the conventional pattern signal interval detecting device. FIG. 10 shows a case where the fluctuation of the peak interval is large, and the peak P10 is m
This is the peak of the pattern signal of the pattern range PD detected the second time, and the time at this time is the past peak interval I1 to Im
This is the time Isum (1 (m) at which (time) is added (integrated). The peak P11 is a time Isum (1 ‥ (m + 1)) (= Isum (1 ‥ m) + (Im +) in which the peak interval from the peak P10 is within the upper limit Iu of the allowable range.
This is the peak detected as the (m + 1) th pattern signal generated in 1)).

The peaks P12 and P13 indicate peaks detected within the peak interval upper limit value Iu and in the range exceeding the upper limit value Iu, respectively. The waveform near the peak P13 indicates a pattern signal within the pattern range PD. Therefore, with the conventional detection device, it is impossible to determine the peak to be detected as the (m + 2) th pattern signal from the peaks P12 and P13, which are peak candidates, and the time Isum (1 ‥ ( m + 2)) could not be determined.

FIG. 11 shows a case where environmental noise is mixed in the peak P15 of the pattern signal between the peaks P14 and P16 of the two detected pattern signals.
The interval I with 16 exceeds the peak interval upper limit value Iu, during which a peak P15 of a noise-like pattern signal is detected. When the lower limit Il of the allowable range is widened, the peak P15 is detected as a pattern signal because the peak interval I ′ with the peak P14 falls within the allowable range, but the width of the peak waveform at this time is Since it is smaller than the range, it is incorrectly detected as noise.

In such a conventional pattern signal interval detection apparatus, when an interval abnormality such as an arrhythmia seen in an electrocardiogram occurs, the peak detected as described above is a false detection peak due to the interval abnormality and noise contamination. It was not possible to judge which one, and as a result, it was impossible to make an automatic judgment.

As a method of solving this problem, there is a conventional example in which an electrophysiological index is measured at a plurality of measurement points, and when all or a majority of the intervals have an interval abnormality, it is determined to be abnormal. However, to automate this method,
In order to improve the reliability of the determination, at least three or more measurement points are required, and deriving a signal from the measurement points is troublesome. For the above reasons, it has been difficult to mount a pattern signal interval detecting device capable of easily measuring a physiological phenomenon on a vehicle or the like which originally has large environmental noise.

Accordingly, the present invention provides a pattern signal interval detecting apparatus for detecting an interval between pattern signals having a peak, by increasing the pattern signal determination accuracy (S / N ratio), thereby enabling accurate pattern signal interval detection. The task is to realize

[0012]

In order to solve the above-mentioned problems, a pattern signal interval detecting apparatus according to the present invention will be described with reference to the example of FIG.
After obtaining P13, the minimum points P131, P13 on both sides of the peak P13
2 is obtained, a peak width D (m) is obtained, and the peak is determined to be a pattern signal peak only when the width D (m) is within a predetermined allowable range that appears in a normal pattern signal. Simulate environmental noise. Then, the interval calculation unit can calculate an accurate interval (peak interval) between the determined pattern signals.

According to the present invention, in the above-mentioned present invention, the pattern determination section may calculate the allowable range based on a past peak width. That is, by dynamically changing the allowable range based on past data, noise removal with high determination accuracy is performed, and pattern signal interval detection becomes possible.

Further, in order to solve the above-mentioned problems, a pattern signal interval detecting device according to the present invention has a structure as shown in FIG.
The calculation unit calculates the past slice level Ss (= peak
Current peak P13 based on the value of P12) and the interval I (m-1).
Of the slice level Ss (for example, Ss-SDs, Ss + SD
s), and the pattern interval I which is the current pattern signal interval
The allowable range of (m) (for example, Id-IDd, Id + IDd) is calculated.

[0015] Then, the pattern judging section detects the pattern signal P13 based on both the allowable range of the slice level and the interval.
(The code P13 also indicates a pattern signal including the peak P13. Hereinafter, the same code is used for the peak and the pattern signal including the peak.) Is determined to be a predetermined pattern signal. The interval calculation unit can calculate the peak interval I (m) between the determined pattern signals.

In the present invention, when the first or second peak is detected, the current allowable range of the slice level is calculated based on the average value or standard deviation value of past input signal data. You can also. That is, when the peak detection is started, there is no slice level of the past peak. Therefore, from the start of detection to the time of detection of the second peak at which the peak interval is determined, the allowable range of the slice level may be calculated based on the average value or standard deviation value of the input signal data.

Further, in order to solve the above-mentioned problem, a pattern signal interval detecting apparatus according to the present invention will be described with reference to the example of FIG. 2. When a pattern storage section stores a reference pattern of a pattern signal shown in FIG. Hold, and the pattern judgment unit
It is determined whether or not a pattern signal by performing pattern matching in a state where the peak P1 or P2 of the input signal shown in the above is matched with the peak P0 of the reference pattern, and the interval calculation unit determines whether the pattern signal is The peak interval can be calculated.

Further, in the present invention, as shown in FIG. 3 (3), the pattern judging section detects, for example, the peak of the input signal within a predetermined pattern range PD of the reference pattern P0.
The peak width D of the peak P2 adjacent to P1 and the adjacent peak P
After normalizing the reference pattern P0 and the input signal P1 in a range (shaded portion) excluding the range on the opposite side of the peak P1 of the input signal from 2, the difference Dit ( When the integrated value DIF (P1) of n) is equal to or smaller than a predetermined value, it can be determined that the input signal is the pattern signal.

The pattern matching between the standard pattern P0 and the peak P2 of the input signal can be performed in the same manner (see the lower diagram in FIG. 2B). Further, in the present invention described above, the peak width D of the adjacent peak P2 is as shown in FIG.
As shown by the peak P13, the width D (m) between the extreme points on both sides of the peak P2 may be used.

As a result, even when the fluctuation of the peak interval of the pattern signal is large and when the peak-like noise is mixed between the peaks of the pattern signal, the signal pattern determination accuracy is high and the accurate interval between the pattern signals is high. Detection can be realized.

[0021]

FIG. 3 shows an embodiment of a pattern signal interval detecting apparatus according to the present invention. In this embodiment, FIG.
The pattern signal interval detection device indicated by the symbol is shown in more detail. That is, the AC amplifier section 24 is composed of a differential input amplifier 1 for inputting and differentially amplifying a signal from the heart rate sensor and a bandpass filter 2 for cutting and outputting low-frequency and high-frequency components of the output signal of the amplifier 1 It is configured.

The A / D converter 25 amplifies the output signal of the band-pass filter 2 and samples the output signal of the amplifier 3 in synchronization with the timing signal 30 from the arithmetic unit 26 to temporarily sample the output signal. Sample and hold circuit to hold
4, an A / D converter 5 for A / D converting the held signal in synchronization with the timing signal 30 and outputting it as digital data, and the digital data at the timing of the control signal 31 from the arithmetic unit 26. The buffer memory 6 for storing the data and the timing signal 37 obtained by dividing the synchronization signal 35 from the arithmetic unit 26
The frequency divider 13 is provided to the D conversion unit 25.

The operation section 26 includes a pattern storage section, a pattern determination section and an interval operation section. When the A / D converter 5 has a circuit configuration including a sample-hold circuit or a circuit configuration not requiring the sample-hold circuit,
4 may be omitted.

The arithmetic unit 26 further includes a RAM 8 and a CPU 10 connected to the buffer memory 6 via an address / data bus 7.
A control signal 33 is supplied to the ROM 9 for the operation algorithm and the RAM 8, a control signal 34 is transmitted / received to / from the CPU 10, a timing signal 30 and a control signal 31 are supplied to the A / D converter 25, and an address is supplied to the address / data bus 7. It comprises a controller 11 and a crystal oscillation circuit 12 that supplies a timing signal 36 to the CPU 10 and a synchronization signal 35 to the DMA controller 11 and the A / D converter 25. Note that the RAM 8 may include the buffer memory 6.

FIG. 4 shows the algorithm of the A / D converter 25. In the following description of the operation, the sampling order n (hereinafter, referred to as time n) is used as a time reference without using the sampling time ΔT of the A / D converter 5. And 1
The number of data read by the buffer memory 6 per time is N, the start time of sampling is N1, and the end time is N2. Therefore, a relationship of N2-N1 + 1 = N is established. Also, A / D at time n
The output data of converter 5 is represented by V (n).

The A / D converter 5 receiving the analog signal (FIG. 3
) Executes the following steps T11 to T14. Step T11 : A / D conversion is performed and data V (n) (time n = N1
‥ N2, “‥” is a code that designates all times of sampling points in the target range). Step T12 : The buffer memory 6 reads this data V (n). Step T13 : When the control signal 31 from the arithmetic unit 26 does not indicate an update request, the process returns to Step T11, and when it indicates an update request, the process proceeds to Step T14. Step T14 : Clear data in the range N1 ‥ N2 of the buffer memory 6, secure a memory area for storing data from the next A / D converter 5, and return to step T11.

FIGS. 5 to 8 show a main routine in the calculation unit 26 and a subroutine T2 for calculating and setting the allowable range of the peak slice level, peak width and peak interval.
2 shows an algorithm of a subroutine T23 for detecting a peak and a subrune T28 for performing pattern matching. The operation of the arithmetic unit 26 will be described below with reference to FIGS.
This will be described with reference to FIGS.

The meanings of the symbols used below are as follows. V '(n): Indicates a moving average value of V (n). V'd (n): Indicates the differential value of V '(n). V'P (m): indicates the m-th peak candidate time and the voltage value at the peak time. I (m): indicates an m-th peak interval candidate and a peak interval. D (m): Indicates the width of the m-th peak for which I (m) was calculated.

The suffix "av" added to the above symbol
“g” indicates the average value, the suffix “std” indicates the standard deviation, the suffix “sum” indicates the integral value, and A, B, and C are arbitrary constants used in the equations. And a,
b, c, d indicate one or more arbitrary constants, and e, f, g, h indicate one or less arbitrary constants.

A flag Fl indicating the control state of the operation unit 26
g1 to 5 are used, and the meaning of each is as follows. Flag Flg1 : Indicates the following peak order state. 1: The first peak 2: The second peak 3: The third and subsequent peaks However, the interval between the first peaks may be short. Flag Flg2 : indicates the state of the following peak interval, and is used for redetection and storage branch processing. 1: shorter than the allowable range 2: longer than the allowable range 3: flag within the allowable range Flg3 : indicates the following peak determination state, and is used to prevent an infinite loop of redetection and storage. 1: First peak judgment 2: Peak judgment when the allowable range is widened 3: Pattern signal judgment by pattern matching

Flag Flg4 : Indicates the following peak interval abnormality determination state, and is used during branch processing and storage. 1: Peak interval is more than the allowable range in the first judgment 2: Peak interval is more than the allowable range in the second judgment 3: Peak interval is less than the allowable range in the first judgment 4: Peak interval is less than the allowable range in the second judgment 5: The peak interval is equal to or more than the allowable range when the pattern signal is determined by the pattern matching. 6: The peak interval is equal to or less than the allowable range when the pattern signal is determined by the pattern matching. 8: There is no problem.

Flag Flg5 : Indicates the processing state of the following A / D conversion data V (n), and is used to select whether or not to execute the moving average processing in step T23. 1: First processing after data reading 2: Second and subsequent processing after data reading In the RAM 8, the RAM areas I '(m), Isum (1 ‥ m), V'P (m), D
(m) is secured.

First, the arithmetic processing of the main routine of FIG. 5 is as follows. Step T20 : First, the arithmetic unit 26 sets initial values, and
The flags Flg1 = 1 and Flg3 = 1 are set to indicate that the second peak is searched in the first allowable range, and the flag Flg5 = 1 is set to indicate the first process of reading data. Furthermore, N1 = 1, N
Set 2 = N and clear all data arrays set in the buffer memory 6. As a result, the flags Flg1 = 1, Flg2 = *, Fl
g3 = 1, Flg4 = *, Flg5 = 1, and m = 1. In addition, * means undecided.

Step T21 : Voltage signal V (n) from A / D conversion 5
Is read into the buffer memory 6. Step T22 : This is a subroutine for calculating a permissible determination range, and proceeds to Step T40 in FIG. Step T40 : Since the flag Flg1 ≦ 2 and the flag Flg3 = 1, the process proceeds to Step T41 for setting an initial allowable range.

Step T41 : Set the slice level upper limit value Su to e.
・ V'avg (N1 ‥ N2) + f ・ V'std (N1 ‥ N2), and slice level lower limit Sl as e , The peak candidate width upper limit Du is F · Du, and the peak candidate width lower limit Dl
Is set to F · Dl, the peak interval upper limit value Iu is set to F · Iu, and the peak interval lower limit value Il is set to F · Il, and initial setting of an allowable range when no peak is detected is performed. The process returns to the main routine of FIG. 5 and proceeds to step T23.

Step T23 : This is a subroutine for detecting a peak, and proceeds to step T50 in FIG. Step T50 : Since the flag Flg5 = 1 indicates that the process is to be performed for the first time after data reading, the process proceeds to step T51. Step T51 : Moving average value V '(n) of V (n) and its differential value V'
d (n) is calculated in the range of N1 ‥ N2, and the process proceeds to Step T52. Step T52 : Sl <V '(nn) <Su and V'd (nn) = 0 without Condition in Time n Range to Detect First Peak
And in the data V '(n) before and after V' (nn), the differential value V'd
One point at which the polarity of (n) changes is calculated, and two points before and after this point are calculated.
Calculate the peak width D, which is the time difference between the points, and calculate Dl <D (m) <Du
Time information nn and peak candidate voltage V '(nn)
Are stored in the RAM areas I ′ (m) and V′P (m), respectively.

For example, when the voltage V '(nn) and the peak width D are out of the allowable range, V'P (m = 1) = 0 is set, indicating that the first peak was not detected.

Step T53 : Set RAM area Isum (1 ‥ m) = I ′ (m) and clear the value of RAM area I ′ (m). Step T54 : Thereafter, the flag Flg5 is set to 2 to indicate that the processing of this subroutine is the second or subsequent processing state after reading the A / D conversion data. As a result, the flag Flg1 = 1, F
lg2 = *, Flg3 = 1, Flg4 = *, Flg5 = 2, and m = 1. The process returns to the main routine of FIG. 5 and proceeds to step T24.

Step T24 : Since no peak has been detected and the flag Flg1 ≦ 2 and V′P (m = 1) = 0, the process proceeds to step T36. Step T36 : The flag Flg3 is set to 2 to widen the allowable range of the peak judgment. As a result, the flags Flg1 = 1, Flg2 = *, Flg3
= 2, Flg4 = *, Flg5 = 2, and m = 1. It returns to step T22. Step T22 : Go to step T40 in FIG. Step T40 : Since the flag Flg1 ≦ 2 and the flag Flg3 = 2, the process proceeds to Step T42 for resetting the peak width, the initial allowable range of the slice level and the peak interval described later, and expanding the allowable range.

Step T42 : The slice level upper limit value Su is set to h.
・ E ・ V'avg (N1 ‥ N2) + d ・ f ・ V'std (N1 ‥ N2), and slice level lower limit Sl is h ・ e ・ V'avg (N1 ‥ N2) -d ・ g ・ V 'std (N1 ‥ N2)
And the peak candidate width upper limit Du is dF / Du, the peak candidate width lower limit Dl is hF / Dl, and the peak interval upper limit Iu is dF / Iu.
The peak interval lower limit Il is set to h · F · Il, and the process returns to the main routine of FIG. 5 and proceeds to step T23.

Step T23 : In steps T50 to T54 of FIG. 7, the first peak is detected, and the peak value, which is the slice level, is stored in the RAM area V'P (m = 1) (> 0). The width is stored in the RAM area D (m). The process returns to the main routine and proceeds to Step T24.

Step T24 : The flag Flg1 ≦ 2 and V′P (x =
1) Since it is> 0, it is determined that the first peak has been detected,
Proceed to step T25. In step T23, the first peak is not detected and the RAM region V′P (m = 1) = 0 remains,
When returning to step T24, the flag Flg1 = 2
And V′P (m = 1) = 0, a detection error occurs. This error is an error indicating that no peak has been detected in spite of the fact that the allowable range for determining the peak has been expanded in step T42.

Step T25 : This is a step of the interval calculating section constituting the calculating section 26. If a peak is detected first, the peak interval I (m) cannot be calculated.
When the peak interval I (m) ≦ I1, the process proceeds to step T26 with the flag Flg2 = 1. As a result, the flags Flg1 = 1, Flg2 = 1, Flg
3 = 2, Flg4 = *, Flg5 = 2, and m = 1.

Step T26 : Since the flags Flg1 = 1 and Flg2 = 1, the process proceeds to step T35. Step T35 : The flag Flg3 = 1, m = m + 1, and the process returns to step T22 to start the operation of detecting the second peak. As a result, the flags Flg1 = 1, Flg2 = 1, Flg3 = 1, Flg4 = *, Flg5 = 2, and
m = 2. If I (m)> Iu in step T25, the process returns to step T22 via steps T26 and T36, and peak detection is executed again.

Hereinafter, detection of the second peak will be described. The difference between the detection of the second peak and the detection of the first peak is that when a peak with a peak interval shorter than the allowable range is detected, whether or not a pattern including the peak is a normal pattern signal is determined by pattern matching. That is, for example, the following step T2 of whether or not it is an arrhythmia
The determination in steps T61 to T68 executed in step 8 is performed.

Step T22 : Since the flags Flg1 = 1 and Flg3 = 1, the initial value allowable range in step T41 is set. Steps T23 to T25 : When the flag Flg1 = 2 (= m) is set and a peak generated in a short peak is detected, it is determined that I (m) <Il, and the flag Flg2 = 1 is set. As a result, the flag Fl
g1 = 2, Flg2 = 1, Flg3 = 1, Flg4 = *, Flg5 = 2, and m = 2.

Step T26 : Flag Flg1 = 2 and Flg2 = 1
Therefore, the process proceeds to Step T27. Step T27 : Since the flags Flg3 = 1 and Flg2 = 1, the process proceeds to step T28. Step T28 : This is a subroutine for performing pattern matching. This subroutine is executed by a pattern storage unit and a pattern determination unit (not shown) constituting the operation unit 26.

The meaning of symbols used in this subroutine will be described below. Fig. 2 (1) held in the pattern storage unit
Is sampled at a time ΔT, and the sampling time is represented by a sampling order r and is assumed to be a time r. Then, the reference pattern signal is represented by P0 (r). The time at which the peak P0 of the reference pattern signal P0 (r) occurs is denoted by r. The pattern signals corresponding to the peaks P1 and P2 to be determined are V (P1) and V (P2), respectively, and the RAM areas PRP (s) and P
OP (s), DIFsum (P1) and DIFsum (P2) are secured.

First, the process proceeds to step T61 in FIG. Step T61 : Time r, pattern range PD, and matching determination reference value E are defined. The data string of the reference pattern signal is read as the RAM area PRP (s) = RP (s + R-PD / 2) in the pattern range PD. Step T62 : The pattern range data of the judgment target V (P1) is converted to V
When (os ‥ oe), V (P1) is read into dataset POP (s) within the following os ‥ oe range. V (P1) start point os: Isum (1 ‥ m-1) -PD / 2 V (P1) end point oe: Isum (1 ‥ m-1) + PD / 2 excluded section od: (Isum (1 ‥ m-1) + PD / 2) − (Isum (1 ‥ m))-D2 /
2) However, if od ≦ 0, od = 0. The data of the RAM areas POP (s) and PRP (s) are normalized to an average of 0 and a standard deviation of 1, respectively, within the above range.

Step T63 : DIFsum (P1) = Σ | PRP (s) -POP
(s) | is calculated and stored in the RAM area DIFsum (P1). However, 0 ≦ s ≦ PD-od Step T64 : RAM area POP (s), os and oe are cleared once. Step T65 : The peripheral data of the judgment target V (P2) is V (os ‥ oe)
Then, read V (P2) into the data set POP (s) within the os ‥ oe range specified below. V (P2) start point os: Isum (1 ‥ m) -PD / 2 V (P2) end point oe: Isum (1 ‥ m) + PD / 2 excluded section od: Isum (1 ‥ m-1) + D1 / 2) − (Isum (1 ‥ m) -PD / 2) However, if od ≦ 0, then od = 0. The average of the data of the RAM areas POP (s) and PRP (s) is 0 in the above range, and the standard. Normalize to deviation 1.

Step T66: DIFsum (P2) = Σ | PRP (s) -POP
(s) | is calculated and stored in the RAM area DIFsum (P2). However, 0 ≦ s ≦ PD step T67 : DIFsum (P1)> DIFsum (P2) and | DIFsum (P
1) If -DIFsum (P2) |> E, determine V (P1) to be determined as noise and replace I (P1) with I (P2) or peak interval I
It is only necessary to attach additional information (flag) indicating an abnormality to (P1). DIFsum (P2)> DIFsum (P1) and | DIFsum (P2) -DIFs
In the case of um (P1) |> E, the determination target VP (P2) is determined as noise, and I (P2) is ignored, or additional information (flag) indicating an abnormality is simply added to I (P2). good.

In cases other than the above, it is determined that V (P2) is a normal pattern signal but the peak interval is abnormal, and I (P2)
Add additional information to 2) that the interval is abnormal. That is,
By performing pattern matching between the standard pattern and the pattern signals V (P1) and V (P2), it is determined whether or not V (P1) and V (P2) are noise, and the position where a normal pattern signal is generated although it is normal (Peak interval from the previous peak) is determined to be abnormal.

Note that r is Iavg (1 ‥ m-1), Davg (1 ‥ m-1)
And Vavg (1 ‥ m−1) may be calculated as a minimum.
In addition, E obtained by correcting the average value of the past DIFsum (P1) or DIFsum (P2) with a constant may be used.

Step T68 : Assuming that the flag Flg3 = 3, FIG.
It returns to step T27 of the main routine. As a result, the flags Flg1 = 2, Flg2 = 1, Flg3 = 3, Flg4 = *, Flg5 = 2, and m = 2. Step T27 : Since the flags Flg2 = 1 and Flg3 = 3,
The process proceeds to Step T30 with the flag Flg4 = 6. Step T30 : The peak interval I (m) and the flag Flg4 = 6 indicating abnormal peak interval are stored in the RAM area. If the processing up to the data of V (N2) has not been completed yet, steps T31 and T3
Proceed to step T22 via 5. As a result, the flag Flg1 =
3, Flg2 = 1, Flg3 = 1, Flg4 = 6, Flg5 = 2, and m = 3.

Hereinafter, the third and subsequent peak detection operations will be described. In this operation, m indicating the order of peaks is incremented by 1 each time a peak is detected.
Since Flg1 satisfies m ≧ 3, it is fixedly set to “3” in step T24.

Step T22 : The process proceeds to step T43 in FIG. Step T40 : Since the flag Flg1 = 3 and Flg3 = 1 indicate that this is the first determination of the third peak, the process proceeds to step T43. Step T43 : Set the slice level upper limit value Su to V'P (m-1) + A · V '
P (m-1), and slice level lower limit Sl is V'P (m-1) -AV'P
(m-1), the peak candidate width upper limit Du is D (m-1) + BD (m-1), and the peak candidate width lower limit Dl is D (m-1) -BD (m -1), the peak interval upper limit Iu is I (m-1) + C · I (m-1), and the peak interval lower limit Il is I (m-1) -C · I (m-1) The third and subsequent allowable ranges are changed in the same manner as in the first and second embodiments. Note that the constants A, B, and C are weighting coefficients. Thereafter, the process returns to the main routine of FIG. 5 and proceeds to step T23.

Steps T23 to T25 : A peak is detected. When the peak interval I (m) satisfies Il <I (m) <Iu, a flag Flg2 = 8 indicating that the peak interval is within an allowable range. age,
Proceed to step T26. Step T26 : Since the flag Flg2 = 8, the process proceeds to step T27. Step T27 : When the flag Flg2 = 8, the flag Flg4 = 8
To step 29. Step T29 : The peak interval is stored in the RAM area I (m), and the process proceeds to step T31.

Step T31 : It is determined whether or not the processing up to the final data V (N2) stored in the buffer memory 6 has been completed. If not, the process returns to step T22 via step T35, and returns to step T22. Is repeated. When the processing is completed, the process proceeds to step T33 via step T32. Step T33 : Set N1 = N1 + N, N2 = N2 + N, return the flag Flg5 = 1, and proceed to step T34. Step T34 : The A / D drive program sends a control signal 31 indicating an update request to the A / D converter 5, and returns to step T21 to start processing new data.

In the above steps T23 to T25, if the peak interval I (m) <Il, the flag Flg2 is set to 1 and 2
As in the case of the second peak detection, a pattern matching step T28 is executed via steps T26 and T27.

On the other hand, if the peak interval I (m)> lu, the flag Flg2 is set to 2 at step T25, and the routine proceeds to step T26. Step T26 : Since the flag Flg1 = 3, step T27
Proceed to. Step T27 : Since the flags Flg3 = 1 and Flg2 = 2,
Assuming that the flag Flg4 = 1, the process proceeds to step T36 through step T36.
Proceed to 22. As a result, the flags Flg1 = 3, Flg2 = 2, Flg3 = 2, Flg
4 = 1, Flg5 = 2, and m = 3. Step T22 : Go to step T40 in FIG. Step T40 : Since the flags Flg1 = 3 and Flg3 = 2, the process proceeds to step T44. Step T44 : Set the slice level upper limit value Su to V'P (m-1) + aA.
V'P (m-1), and slice level lower limit Sl is V'P (m-1) -aA
・ V'P (m-1), and the peak candidate width upper limit Du is D (m-1) + b ・ B ・ D
(m-1), and the peak candidate width lower limit Dl is D (m-1) -bBBD (m-1)
And the peak interval upper limit Iu is I (m-1) + c · C · I (m-1),
When the peak interval lower limit Il is defined as I (m-1) -c.C.I (m-1), if the third and subsequent allowable ranges are not within the allowable range even in the prediction of the allowable range (step T43), the allowable range is again set. Is changed to expand. Note that the constants a, b, and c are weighting coefficients.
Thereafter, the process proceeds to step T23 in FIG.

Steps T23 to T26 : Peaks are detected. Then, for example, when the peak interval I (m)> Iu, the flag Flg
It returns to step T27 as 3 = 2 and Flg2 = 2. Step T27 : Since the flags Flg3 = 2 and Flg2 = 2,
Flag Flg4 = 2 indicating that the second peak interval is abnormal
To step T28. Steps T28 and T61 above
After T68 is executed and pattern matching is performed,
The flag Flg3 is set to 3 and the process returns to step T27.

Step T27 : Flag Flg3 = 3 and Flg2 = 2
Therefore, the flag Flg4 is set to 5 and the process proceeds to step T30. Step T30 : The peak interval with the abnormality flag Flg4 = 5 is stored in the RAM area I (m), and the process proceeds to step T31. If the processing up to the final data V (N2) has not been completed, Returning to step T23 via steps T31 and T35, the above operation is repeated.
The process returns to step T21 via 34 to start processing new data.

[0063]

As described above, according to the pattern signal interval detecting apparatus according to the present invention, when the peak width between the poles of the peak P of the input signal is within the allowable range, the predetermined pattern signal is Judgment or prediction of each allowable value of peak slice level and pattern interval based on past values and judge as a pattern signal, or prepare a reference pattern and judge as a pattern signal by pattern matching and judge Since it is configured to calculate the peak interval between pattern signals, it has high resistance to environmental noise and the like, has high signal pattern determination accuracy, and can realize accurate pattern signal interval detection. Was.

[0064]

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an operation principle (1) of a pattern signal interval detection device according to the present invention.

FIG. 2 is a diagram illustrating an operation principle (2) of the pattern signal interval detecting device according to the present invention.

FIG. 3 is a block diagram showing a configuration example of a pattern signal interval detection device according to the present invention.

FIG. 4 is a flowchart showing an operation example (1) of the pattern signal interval detecting device according to the present invention.

FIG. 5 is a flowchart showing an operation example (2) of the pattern signal interval detection device according to the present invention.

FIG. 6 is a flowchart of a subroutine T22 in the pattern signal interval detection device according to the present invention.

FIG. 7 is a flowchart of a subroutine T23 in the pattern signal interval detection device according to the present invention.

FIG. 8 is a flowchart of a subroutine T28 in the pattern signal interval detection device according to the present invention.

FIG. 9 is a block diagram showing a configuration example of a general pattern signal interval detection device.

FIG. 10 is a diagram illustrating an example (1) of a pattern signal detected by a general pattern signal interval detection device.

FIG. 11 is a diagram illustrating an example (2) of a pattern signal detected by a general pattern signal interval detection device.

[Explanation of symbols]

1 Differential input amplifier 2 Band pass filter 3 Amplifier 4 Sample hold circuit 5 A / D converter 6 Buffer memory 7 Address / data bus 8 RAM 9 Operation algorithm ROM 10 CPU 11 D
MA controller 12 Crystal oscillation circuit 13 Divider 20
Human body 21-23 Biological electrode 24 AC amplifier 25
A / D conversion unit 26 Operation unit Flg1 to 5 Flags P0 to P3, P11 to P1
6 Peak Ss Slice level Su Slice level upper limit Sl Slice level lower limit I (m) Peak interval Iu Peak interval upper limit Il Peak interval lower limit D (m), D, D1, D2 Peak width, peak candidate width Du Peak candidate Width upper limit value Dl Peak width lower limit value PD pattern range In the figures, the same symbols indicate the same or corresponding parts.

Claims (7)

[Claims] 1. A pattern determining unit for determining a peak width between poles on both sides of a peak of an input signal, and determining the peak as a pattern signal only when the width is within a predetermined allowable range; A pattern signal interval detecting device, comprising: an interval calculating unit that calculates a peak interval between the two. 2. The method according to claim 1, wherein the pattern determination unit comprises:
A pattern signal interval detecting device, wherein the allowable range is calculated based on a past peak width. 3. A calculating unit for calculating a current allowable range of a slice level of a peak of an input signal and a current allowable range of an interval between pattern signals based on past slice levels and intervals, respectively. A pattern signal interval detection device, comprising: a pattern determination unit that determines the pattern signal based on both allowable ranges of the interval; and an interval calculation unit that calculates a peak interval between the determined pattern signals. 4. The method according to claim 3, wherein the current allowable range of the slice level is calculated based on an average value or a standard deviation value of past input signal data at the time of detecting the first or second peak. A pattern signal interval detection device characterized by the above. 5. A pattern storage unit for holding a reference pattern, a pattern determination unit for determining whether or not the input signal is a pattern signal by performing pattern matching in a state where a peak of the input signal matches the peak of the reference pattern; An interval calculating unit for calculating a peak interval between the determined pattern signals; and a pattern signal interval detecting device. 6. The method according to claim 5, wherein the pattern determination unit determines a width of a peak adjacent to the peak of the input signal within a predetermined pattern range of the reference pattern, and a width of the peak of the input signal opposite to the width of the adjacent peak. After normalizing the reference pattern and the input signal in a range excluding the range on the side, when the integrated value of the difference between the two is smaller than a predetermined value, it is determined that the input signal is the pattern signal. A pattern signal interval detection device characterized by the above. 7. The pattern signal interval detecting apparatus according to claim 6, wherein the width of the adjacent peak is a width between poles on both sides of the adjacent peak.