KR100961993B1 - Method and Apparatus of signal detection using long term observation window and Recording medium using by the same - Google Patents

Abstract

Disclosed are an activation signal detection method using a wide-area observation window, an apparatus thereof, and a recording medium recording the same.

An activation signal detection method using a wide area observation window according to the present invention includes: outputting a signal waveform measured from an object to be examined, determining starting and ending time points of the output signal waveform, and counting the number of signals within the determined time point; Setting a size of a wide area observation window including a local observation window and the local observation window using the object to be examined and the number of counted signals, and determining a first threshold according to an amplitude of the measured signal waveform; Setting a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and counting the number of local observation windows having a signal number greater than or equal to the set secondary threshold in the global observation window; Determining a third threshold according to the counted number of local observation windows, and detecting an activation signal by detecting a wide area observation window exceeding the third threshold; And activating signal information of the object under test including a start time and an end time of the counted wide area observation window, a start time and an end time of the output signal waveform, a count of the counted wide area observation window, and an identifier of the object to be examined. Storing in the.

According to the present invention, it is possible to accurately determine the start time and the end time of the event to be measured from the object to be tested, making it easy to observe whether or not the event occurs and the daily change of the object to be tested, without having to configure a separate filter Not only can the error probability of signal detection be greatly reduced from the measured signal waveforms, but also the manufacturing cost of the measuring device can be reduced and the measurement accuracy can be improved.

Description

Activation signal detection method using wide area observation window, apparatus and recording medium recording the same {Method and Apparatus of signal detection using long term observation window and Recording medium using by the same}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to signal detection. In particular, the present invention relates to a method for detecting an activation signal at a low signal-to-noise ratio, which can reduce the influence of noise in an environment having a low signal-to-noise ratio, and to observe a start point and an end point of a specific signal according to daily changes. To the device.

EMG, electrocardiogram, electroencephalogram or microphone output as sinusoidal signals that are generally measured contain some noise.

1 illustrates an example of a typical sinusoidal signal.

As shown in FIG. 1, a typical sinusoidal signal is a voltage signal including an original signal and a noise. In this signal, it is important to distinguish noise and an activation time of the original signal, that is, to determine when the activation starts and ends.

As described above, since the sinusoidal signal is mixed with noise, it is difficult to determine the presence or absence of a signal when the signal to noise ratio (SNR) is low, and even when the signal to noise ratio is high, the activation of the original signal starts due to the noise. It is difficult to determine the time point and the end time point.

As such, quantitative evaluation of the degree of activation in the original signal is of course important, but it is also important to distinguish between the noise and the activation time of the original signal, that is, the method of determining when the activation starts and ends.

The measurement of such a signal may be performed for a short time for the purpose of diagnosis or for a long time for the purpose of treatment.

Electrocardiogram is a signal that occurs periodically and extracts information related to heart disease by observing the period and waveform of the signal regardless of the subject's will, so the measurement time is considerably longer than the activation time for continuous activation signal measurement.

In the case of EMG, a signal may be activated once during the measurement time because aperiodic measurement may be performed for a short time to evaluate the condition or ability of the subject.

When measuring snoring during sleep with a microphone, the output signal is almost similar to the non-periodic EMG output.However, once snoring progresses, it occurs periodically in the same way as the breathing cycle and snoring changes in diagnosis and treatment. This position requires observation.

In particular, the ECG signal-to-noise ratio is very large, so the probability of making an error in the measurement result analysis is low.

However, undesired noise may occur in EMG measurements, snoring measurements, or vibrations or noises caused by the operation of machines or equipment.

Such noise causes a decrease in validity and a high probability of error in the method of determining the characteristics of the original signal.

Various methods have been proposed to remove noise from the original signal. For example, there are filtering methods. However, the filtering method removes signals except for frequency components inherent to the original signal. Sometimes used in combination.

However, the signal detection device which performs the signal detection by removing the noise from the original original signal can obtain the desired measurement result by using the filter, but it is uneconomical and perfect due to the increase in the cost incurred due to the filter configuration. There is a problem that can not remove the noise.

The first problem to be solved by the present invention is to reduce the effect of noise in the environment having a low signal-to-noise ratio, the activation signal detection method using a wide-area observation window that can observe the start time and end time of a specific signal according to the change of day To provide.

A second object of the present invention is to provide an activation signal detection apparatus using a wide area observation window to which the activation signal detection method using the wide area observation window is applied.

A third problem to be solved by the present invention is to provide a recording medium recorded by a program that can be performed by a computer for the activation signal detection method using the wide area observation window.

The present invention to solve the first problem,

Outputting a signal waveform measured from an object to be tested, determining start and end time points of the output signal waveform, and counting the number of signals within the determined time point; Setting a size of a wide area observation window including a local observation window and the local observation window using the object to be examined and the number of counted signals, and determining a first threshold according to an amplitude of the measured signal waveform; Setting a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and counting the number of local observation windows having a signal number exceeding the set secondary threshold in the global observation window; Determining a third threshold according to the counted number of local observation windows, and detecting an activation signal by detecting a wide area observation window exceeding the third threshold; And activating signal information of the object under test including a start time and an end time of the counted wide area observation window, a start time and an end time of the output signal waveform, a count of the counted wide area observation window, and an identifier of the object to be examined. It provides a method for detecting an activation signal using a wide area observation window comprising the step of storing in.

On the other hand, the signal waveform is characterized in that the signal waveform having a low signal-to-noise ratio.

In addition, the start and end time points of the output signal waveform is set based on the identifier of the object to be stored in the log.

The storing of the activation signal information in the log may be included in the first-area counted wide-area observation window among the counted wide-area observation windows, and the fastest time value among the signals exceeding the first threshold of the counted local observation window. Storing a signal having a start time of the activation signal; And a signal which is included in the last counted wide area observation window among the counted wide area observation windows, and stores the signal having the latest time value among the signals exceeding the first threshold of the counted local observation window as an end point of the activation signal. Characterized in that it comprises a step.

In addition, the signal waveform is characterized by consisting of a sinusoidal signal including an electrocardiogram, an electrocardiogram, an electroencephalogram or a microphone output.

On the other hand, in the activation signal detection method using the wide area observation window to calculate the reception operation characteristics by comparing the activation signal information of the test target and the actual event, the number of local observation windows of the third threshold according to the reception operation characteristics And resetting and storing the same in the log.

라고 할 때, 상기 국부 관찰창을 상기 광역 관찰창 내에

가 존재하도록 설정할 경우, 단위 광역 관찰창에서 상기 설정된 2차 문턱치를 초과하는 신호 갯수를 가지는 국부 관찰창의 갯수가

이상인지의 여부를 이용하여 수신 동작 특성의 확률을 연산함으로써 판단되며, 상기 수신 동작 특성의 확률

는

에 의해 연산되는 것을 특징으로 한다.Here, the reception operation characteristic is a probability of detecting an activation signal calculated using only the local observation window.

Said local observation window within said wide area observation window.

Is set to exist, the number of local observation windows having the number of signals exceeding the set secondary threshold in the unit wide observation window.

The probability of the reception operation characteristic is determined by calculating the probability of the reception operation characteristic using whether or not it is abnormal.

Is

It is characterized in that calculated by.

The present invention to solve the second problem,

An output signal counting unit configured to output a signal waveform measured from a test target, determine start and end time points of the output signal waveform, and count the number of signals within the determined time point; An observation window setting unit configured to set a size of a wide area observation window including a local observation window and the local observation window by using the object to be examined and the number of counted signals; Determine a primary threshold according to the amplitude magnitude of the measured signal waveform, set a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and set the secondary threshold in the global observation window. A threshold setting unit for determining a third threshold according to the number of local observation windows having an excessive number of signals; An activation signal detection unit for counting the number of local observation windows having a signal number greater than or equal to the secondary threshold, and detecting an activation signal by detecting a wide area observation window in which the counted local observation windows exceed the third threshold; And a start time and an end time of the counted wide area observation window, a start time and an end time of the output signal waveform, a count of the counted wide area observation window, and an identifier of the test object. It provides an activation signal detection apparatus using a wide area observation window including a log storage unit.

Here, the signal waveform is characterized in that the signal waveform having a low signal-to-noise ratio.

The start and end time points of the output signal waveform may be set based on the activation signal information of the target object stored in the log storage unit.

In addition, the log storage unit is included in the counted wide area observation window among the counted wide area observation window, the signal having the fastest time value among the signals exceeding the first threshold of the counted local observation window the activation signal And a signal having the latest time value among signals which are included in the last counted wide area observation window among the counted wide area observation windows and exceeding the first threshold of the counted local observation window. The apparatus may further include a timing recording module for storing at the end point of the.

The log storage unit may further include a feedback module that compares the activation signal information of the test target with an event that actually occurs, calculates a reception operation characteristic, and resets the number of local observation windows of the third threshold according to the reception operation characteristic. It is characterized by.

라고 할 때, 상기 국부 관찰창을 상기 광역 관찰창 내에

가 존재하도록 설정할 경우, 단위 광역 관찰창에서 상기 설정된 2차 문턱치를 초과하는 신호 갯수를 가지는 국부 관찰창의 갯수가

이상인지의 여부를 이용하여 수신 동작 특성의 확률을 연산함으로써 판단되며, 상기 수신 동작 특성의 확률

는

에 의해 연산되는 것을 특징으로 한다.The reception operation characteristic is based on a probability of detecting an activation signal calculated using only the local observation window.

Said local observation window within said wide area observation window.

Is set to exist, the number of local observation windows having the number of signals exceeding the set secondary threshold in the unit wide observation window.

The probability of the reception operation characteristic is determined by calculating the probability of the reception operation characteristic using whether or not it is abnormal.

Is

It is characterized in that calculated by.

According to the present invention, it is possible to accurately determine the start time and the end time of the event to be measured from the object to be tested, making it easy to observe whether or not the event occurs and the daily change of the object to be tested, without having to configure a separate filter Not only can the error probability of signal detection be greatly reduced from the measured signal waveforms, but also the manufacturing cost of the measuring device can be reduced and measurement accuracy can be improved.

Statistical methods used in signal measurement include single and double threshold methods.

The single threshold method is most commonly used for observing the signal-to-noise ratio and determining when the signal exceeds a certain value as shown in FIG. 2 as the activation of the signal, but the magnitude of the error is significantly higher than the set threshold 210. Depends on

는 하기의 수학식 1에 의해 정의된다.Detection probability of a signal exceeding a single threshold determined by the single threshold method

Is defined by Equation 1 below.

As can be seen in Equation 1, since the single threshold method determines the presence or absence of a signal with a single threshold value according to the magnitude of the amplitude in relation to the magnitude of noise, when the threshold is set incorrectly, the probability of false detection of the signal is increased. It is high, and the probability of false detection of the signal increases further under noise other than white Gaussian noise, that is, in an environment in which the magnitude of the noise is difficult to predict.

FIG. 3 is a graph illustrating a double threshold method used to reduce the probability of signal misdetection according to the single threshold value shown in FIG. 2.

으로 관찰창(320)의 크기를 정하고, 정해진 관찰창(320) 내에서 1차 문턱치

(310) 를 초과한 신호의 갯수를 이차 문턱치로 정하여 이 두 개의 문턱치를 이용하여 신호의 활성화 여부를 판단한다.In the double threshold method, the number of signals generated within the time determined by the user according to the usage environment is first used.

Determining the size of the observation window 320 to the first threshold within the determined observation window 320

The number of signals exceeding 310 is determined as the secondary threshold, and the two thresholds are used to determine whether the signals are activated.

중 일차 문턱치를 초과한 연속적인 신호

개 중에서

개의 신호가 발생할 확률

은 연속적인 베르누이 시도(repetition of Bernoulli trials)의 결과로 이항분포의 형식이 되며 이는 하기의 수학식 2와 같다.That is, the number of signals generated in time

Continuous signal exceeding primary threshold

Out of

Probability of occurrence of signals

Is a form of binomial distribution as a result of successive repetition of Bernoulli trials, which is shown in Equation 2 below.

개 중에서 이차 문턱치인

개 이상 발생할 확률은 하기의 수학식 3과 같이 표현된다.And the continuous signal exceeding the primary threshold 310

Second threshold among dogs

Occurrence probability is expressed by Equation 3 below.

와 상기 검출된 신호 중 오류일 확률

은 각각 수학식 4 및 수학식 5으로 표현된다.Here, the probability that the signal is detected beyond the double threshold

And probability of error among the detected signals

Are represented by equations (4) and (5), respectively.

는 일차 문턱치를 초과한 연속적인 신호

개 중에서 이차 문턱치인

개 이상 발생하며, 일차 문턱치를 초과하여 나타날 신호의 확률은

로 이항분포의 형식이 된다.The probability that a signal will be detected beyond the double threshold

Is a continuous signal that exceeds the primary threshold

Second threshold among dogs

Or more, and the probability of the signal appearing above the primary threshold

This results in the form of a binomial distribution.

은 연속적인 신호

개 중에서 이차 문턱치인

개 이상 발생하며, 잡음의 발생 확률은

로 마찬가지로 이항분포의 형식이 된다.The signal was detected as having exceeded the double threshold, but the probability of the detected signal is an error

Is a continuous signal

Second threshold among dogs

Or more, and the probability of occurrence of noise

Similarly, the form of binomial distribution.

은 20이고,

는 일차 문턱치(310)로 잡음의 최대값이며, 이차 역치

는 1인 경우이다. 이 경우에는 실제 신호가 아닌 잡음에 의한 확률

를 관찰할 수 있고, 이러한 원인은

값이 너무 작게 설정되었기 때문이다.Referring to FIG. 3, the size of the local observation window 320

Is 20,

Is the primary threshold 310, which is the maximum of noise, and the secondary threshold

Is a case of 1. In this case, the probability of noise rather than the actual signal

Can be observed and these causes

This is because the value is set too small.

를 줄일 수 있고, 신호대잡음비가 낮은 경우에도 적용 가능한 신호 검출 기법으로 이중 문턱치를 이용한 신호의 검출에는 상기에서 본 바와 같이 실제 신호가 아닌 잡음에 의한 확률

이 큰 값을 가지게 되므로, 이를 줄이고, 신호 검출의 정확성을 기함과 동시에 신호 검출 시간 및 발생 이벤트의 일변화 관찰을 용이하게 수행할 수 있다.On the other hand, the present invention as described above the probability by noise rather than the actual signal

As a signal detection technique that can be applied even when the signal to noise ratio is low, the detection of a signal using a double threshold is more likely than the actual signal.

Since this value is large, it is possible to reduce this, to ensure the accuracy of signal detection, and to easily observe the daily change of the signal detection time and the occurrence event.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

However, the following embodiments of the present invention may be modified into various other forms, and the scope of the present invention is not limited to the embodiments described below.

4 is a flowchart illustrating a method of detecting an activation signal using a wide area observation window according to the present invention.

Referring to FIG. 4, first, a signal waveform measured from an object to be tested is output, and start and end time points of the output signal waveform are determined to count the number of signals within the determined time point (step 410).

The object to be tested can be configured in various embodiments.

The electrocardiogram and electrocardiogram of the user to be examined may be measured by the vibration of the diaphragm, and snoring using a microphone or vibration generated when the machine or equipment is driven may be the object of the present invention.

Accordingly, the signal waveform may be a sinusoidal signal including an electrocardiogram, an electrocardiogram, an electroencephalogram, or a microphone output, and the signal waveform may be a signal waveform having a low signal to noise ratio.

Based on this, the signal waveform measured from the object to be tested is output using a signal waveform output device and the like and the start and end time points of the output signal waveform are determined.

The start and end time points of the signal waveform may be set based on an identifier of an object to be inspected in a log storing previously measured data when there is a previously measured history. Therefore, an identifier of an object to be tested is assigned when an actual test is performed. The result of the test is stored together with the identifier of the test target.

If there is no history previously measured, the actual measurement start time and the measurement end time may be set as the start and end points of the signal waveform.

Meanwhile, after determining the start and end time points of the output signal waveform, the number of signals within the determined time point is counted.

The number of signals may refer to the number of peaks of a signal whose peak value is recognized by a waveform recorder or the like.

Then, the size of the local observation window and the wide-area observation window including the local observation window is set using the object to be tested and the number of counted signals, and the first threshold value is determined according to the amplitude of the measured signal waveform. 420 courses).

The above-described local observation window is set using the object to be examined and the counted signal number, and a wide-area observation window including the local observation window is set.

Here, the local observation window is calculated based on the number of the counted signals. When the frequency value of the signal is small, the number of signals included in the local observation window is relatively large, and when the frequency value of the signal is large, the local observation window is large. The size of the window can be set to include a relatively small number of signals included in the.

Meanwhile, the local observation window and the wide area observation window are set, and at the same time, the first threshold value according to the amplitude of the waveform of the measurement signal is set.

Since the sinusoidal signal has its own data and is accompanied by data transmission through a noisy channel, the measurement signal under test contains noise.

Therefore, the signal waveform before measuring the signal waveform from the object to be tested may be recorded, and the first threshold of the measurement signal may be set based on the calculated maximum value of noise from the recorded signal waveform.

Then, the secondary threshold of the local observation window is set according to the number of signals exceeding the determined primary threshold, and the number of local observation windows having a signal number exceeding the set secondary threshold is counted in the wide area observation window. (430 course).

으로 관찰창의 크기를 정하고, 정해진 관찰창 내에서 1차 문턱치

를 초과한 신호의 갯수를 이차 문턱치로 정하여 이 두 개의 문턱치를 이용하여 신호의 활성화 여부를 판단한다.Secondary threshold is the number of signals generated within the time set by the user according to the usage environment.

Size of observation window, and the first threshold within the specified observation window.

The number of signals exceeding is determined as the secondary threshold to determine whether the signal is activated using these two thresholds.

If there is a local observation window that exceeds this secondary threshold, the local observation window is counted to determine the following third threshold value.

Next, the third threshold is determined according to the counted number of local observation windows, and the activation signal is detected by detecting the wide observation window exceeding the third threshold.

The third threshold is easily influenced by noise in a low signal-to-noise ratio environment to determine whether the signal is activated only by the local observation window, so that the counted local observation window is more than a predetermined number to further exclude the influence of the noise. It is determined that the signal is activated.

The third threshold is obtained by receiving a receiver operating characteristics (ROC) to verify the validity of the variable selection corresponding to the third threshold according to the comparison and analysis with the actual value.

라고 할 때, 상기 국부 관찰창을 상기 광역 관찰창 내에

가 존재하도록 설정할 경우, 단위 광역 관찰창에서 상기 2차 문턱치를 초과하는 신호 갯수를 가지는 국부 관찰창의 갯수가

이상인지의 여부인 3차 문턱치를 이용하여 수신 동작 특성의 확률을 연산함으로써 판단되며, 상기 3차 문턱치에 의한 수신 동작 특성의 확률

는 하기의 수학식 6에 의해 연산될 수 있다.Here, the reception operation characteristic is a probability of detecting an activation signal calculated by Equation 4 using only the local observation window.

Said local observation window within said wide area observation window.

Is set to exist, the number of local observation windows having the number of signals exceeding the secondary threshold in a unit wide observation window.

It is determined by calculating the probability of the reception operation characteristic using the third threshold, which is whether or not it is abnormal, and the probability of the reception operation characteristic by the third threshold.

May be calculated by Equation 6 below.

Finally, the activation signal information of the object under test, including the start time and the end time of the counted wide area observation window, the start time and the end time of the output signal waveform, the number of times of the counted wide area observation window, and the identifier of the test object. Save to the log (step 450).

If the aperiodic characteristics are performed for a short time to evaluate the condition or ability of the subject such as EMG, the activation of the signal may occur a small number aperiodically during the measurement time. The timing of the time point and the end point is important, and when measuring snoring that occurs during sleep with a microphone, the output signal is periodically generated in the same manner as the breathing cycle, so the daily change of the snoring can be observed. The starting point and end point of the wide area observation window are important.

As described above, in the activation signal measurement based on the third-order threshold according to the present invention, it is easy to recognize not only a simple signal but also to observe the change in the signal.

Meanwhile, in order to store the activation signal information in a log, the fastest time value among signals that are included in the first-area counted wide-area observation window among the counted wide-area observation windows and exceed the first threshold of the counted local observation window. A signal having a start time of the activation signal and included in the last counted wide area observation window among the counted wide area observation windows, and is the latest time of the signal exceeding the first threshold of the counted local observation window; A signal having a value may be stored as an end point of the activation signal.

In addition, in step 450, a reception operation characteristic may be calculated by comparing the activation signal information of the target object with an event that actually occurs, and the number of local observation windows of the third threshold is reset according to the reception operation characteristic and stored in the log. Can be.

When a large number of errors occur between the event actually occurring and the count of the activation signal information of the test object, the activation signal information which further increases the number of the local observation window is stored in the log, so that the count number of the activation signal information and the event actually occurred. This is to add a process of performing feedback in order to reduce the difference in the number of.

Figure 5 shows an example of the activation signal detection method using a wide area observation window according to the present invention.

이 6개가 존재하도록 설정하였고, 단위 광역 관찰창(630)에서 상기 1차 문턱치를 초과하는 신호 갯수로 설정되는 2차 문턱치를 만족하는 국부 관찰창의 갯수를

3개 이상으로 하는 3차 문턱치를 설정하여 신호의 활성화 여부를 판단함을 알 수 있다.Referring to FIG. 5, in the detection of an activation signal using a wide-area observation window according to the present invention, a local observation window 520 is provided in a unit-wide observation window 630.

The number of local observation windows satisfying the secondary threshold set to the number of signals exceeding the primary threshold in the unit wide area observation window 630 is set to exist.

It can be seen that the third threshold is set to three or more to determine whether the signal is activated.

또는 3차 문턱치

을 재설정하여 수신 동작 특성을 확인함으로써 오차 값을 줄일 수 있다.When the error occurs largely in the number of events actually occurring in FIG. 5 and the number of activations of the signal determined according to FIG. 5, the local observation window

Or third threshold

The error value can be reduced by resetting the value and confirming the reception operation characteristic.

6 illustrates an activation signal detection apparatus using a wide area observation window according to the present invention.

Referring to FIG. 6, the activation signal detection apparatus using the wide-area observation window according to the present invention includes an output signal counting unit 610, an observation window setting unit 620, a threshold setting unit 630, an activation signal detection unit 640, and the like. The log storage unit 650 may be configured.

First, the output signal counting unit 610 outputs a signal waveform measured from an object to be tested, determines start and end time points of the output signal waveform, and counts the number of signals within the determined time point.

According to the object to be tested, the test form may be configured in various embodiments. The electrocardiogram and electrocardiogram of the user who wants to perform the test may be measured by vibration of the diaphragm. Vibration or the like generated during driving can be the object of the present invention.

Accordingly, the signal waveform may be a sinusoidal signal including an electrocardiogram, an electrocardiogram, an electroencephalogram, or a microphone output, and the signal waveform may be a signal waveform having a low signal to noise ratio.

Based on this, the output signal counting unit 610 determines start and end time points of the signal waveform by using the signal waveform of the test target output by the signal waveform output unit or the like.

The start and end time points of the signal waveform may be set based on an identifier of an object to be stored in the log storage unit 650 which stores previously measured data when there is a previously measured history.

Therefore, an identifier of the object to be inspected is assigned when the inspection is actually performed, and the result of the inspection is stored in the log storage unit 650 together with the identifier of the object to be inspected.

If there is no history previously measured, the actual measurement start time and the measurement end time may be set as the start and end points of the signal waveform.

Accordingly, the output signal counting unit 610 determines the start and end time points of the output signal waveform and counts the number of signals within the determined time point.

The number of signals may mean the number of peaks of a signal in which the peak value is recognized.

On the other hand, the observation window setting unit 620 sets the size of the regional observation window including the local observation window and the local observation window using the number of signals counted by the object to be examined and the output signal counting unit 610.

Here, the observation window setting unit 620 sets the above-described local observation window by using the object to be examined and the counted signal number, and sets a wide observation window including the local observation window.

Here, the local observation window is calculated based on the number of the counted signals. When the frequency value of the signal is small, the number of signals included in the local observation window is relatively large, and when the frequency value of the signal is large, the local observation window is large. The size of the window may be set such that the number of signals included in the window is relatively small. The above-described local observation window may be set using the target object and the counted number of signals, and a wide observation window including the local observation window may be provided. Set it.

Here, the reason for setting the wide observation window is that if only the local observation window is set, an error occurs in an environment having a low signal-to-noise ratio because only the signal is determined according to the number of signals exceeding the primary threshold value counted on the local observation window. Since the probability is high, the probability of false detection of the signal is significantly reduced by determining whether the signal is activated in the wide-area observation window according to the number of local observation windows determined to be activated.

In addition, the threshold setting unit 630 determines a primary threshold according to the amplitude of the measured signal waveform, sets a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and the wide area. The third threshold is determined according to the number of local observation windows having the number of signals exceeding the set secondary threshold in the observation window.

That is, the threshold setting unit 630 first sets the primary threshold according to the amplitude of the waveform of the measurement signal according to the local observation window and the wide area observation window set by the observation window setting unit 620.

Since the sinusoidal signal has its own data and is accompanied by data transmission through a noisy channel, the measurement signal under test contains noise.

Therefore, the threshold setting unit 630 may calculate the maximum value of the noise from the signal waveform counted by the output signal counting unit 610 and set the primary threshold of the measurement signal based on the calculated value.

The threshold setting unit 630 sets the secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and the activation signal detection unit 640 sets the secondary threshold in the global observation window. Count the number of local observation windows with the number of signals exceeding.

으로 관찰창 설정부(620)에서 관찰창의 크기를 정하고, 정해진 관찰창 내에서 문턱치 설정부(630)는 1차 문턱치

를 초과한 신호의 갯수를 이차 문턱치로 정하며, 이 두 개의 문턱치를 이용하여 활성화 신호 검출부(640)에서 신호의 활성화 여부를 판단한다.The number of signals generated within the time determined by the user according to the usage environment

The observation window setting unit 620 determines the size of the observation window, and the threshold setting unit 630 within the determined observation window is the primary threshold

The number of signals exceeding is determined as a secondary threshold, and the activation signal detector 640 determines whether the signal is activated using the two thresholds.

If it is determined by the activation signal detection unit 640 that there is a local observation window that exceeds this secondary threshold, the local observation window is counted to determine the following third threshold value.

Then, the threshold setting unit 630 determines the third threshold according to the counted number of local observation windows, and the activation signal detector 640 detects an activation signal by detecting a wide area observation window exceeding the third threshold. .

The third threshold used in the activation signal detection unit is susceptible to noise in a low signal-to-noise ratio environment to determine whether the signal is activated only by the local observation window. It is judged that the signal is activated only when the observation window is over a certain number.

Meanwhile, in order to determine whether the set third threshold is valid, a log storage unit stores the third threshold, and the third threshold is used for comparison and analysis with an actual value according to a receiver operating characteristics (ROC). Therefore, the validity of the variable selection corresponding to the third threshold is verified.

라고 할 때, 상기 국부 관찰창을 상기 광역 관찰창 내에

가 존재하도록 설정할 경우, 단위 광역 관찰창에서 상기 2차 문턱치를 초과하는 신호 갯수를 가지는 국부 관찰창의 갯수가 이상인지의 여부인 3차 문턱치를 이용하여 수신 동작 특성의 확률을 연산함으로써 판단되며, 상기 3차 문턱치에 의한 수신 동작 특성의 확률

는 상기의 수학식 6에 의해 연산된다.Here, the reception operation characteristic is a probability of detecting an activation signal calculated by Equation 4 using only the local observation window.

Said local observation window within said wide area observation window.

Is set to exist, the number of local observation windows having the number of signals exceeding the secondary threshold in a unit wide observation window. It is determined by calculating the probability of the reception operation characteristic using the third threshold, which is whether or not it is abnormal, and the probability of the reception operation characteristic by the third threshold.

Is calculated by the above equation (6).

The log storage unit 650 includes the start point and the end point of the counted wide area observation window, the start point and the end time of the output signal waveform, the count of the counted wide area observation window, and the identifier of the object to be examined. The activation signal information may be stored, and the activation signal information of the test target previously stored in the log storage unit 650 may be used when the signal waveform of the same test target is measured.

Therefore, the start and end time points of the signal waveform output from the signal waveform output unit 610 may be set based on the activation signal information of the target object stored in the log storage unit 650.

The log storage unit 650 is included in the first-area counted wide-area observation window among the counted wide-area observation windows, and has a fastest time value among signals exceeding the first threshold of the counted local observation window. Is stored as a start time of the activation signal, is included in the last counted wide area observation window of the counted wide area observation window, the signal having the latest time value of the signal exceeding the first threshold of the counted local observation window The timing recording module may further include a timing recording module configured to store the activation signal as an end point of the activation signal.

The log storage unit 650 calculates a reception operation characteristic by comparing the activation signal information of the test target with an event that actually occurs, and resets the number of local observation windows of the third threshold according to the reception operation characteristic. It may further include a module (not shown).

The feedback module further increases the number of local observation windows and stores the number of local observation windows in the log storage unit 650 when a large number of errors occur in the count of the event and the activation signal information of the test target. It serves as a feedback to reduce the difference between the number and the number of events actually occurring.

FIG. 7 illustrates an activation signal determined according to an activation signal detection method DTD using a double threshold.

을 5로 국부 관찰창의 크기를 정하고, 정해진 관찰창 내에서 1차 문턱치(710)

를 초과한 신호의 갯수로 정해지는 이차 문턱치

을 4로 설정하여 이 두 개의 문턱치를 이용하여 신호의 활성화 여부를 판단하고 있음을 알 수 있다.Referring to FIG. 7, the number of signals generated within a time determined by a user

Set the size of the local observation window to 5, and the primary threshold 710 within the given observation window.

Secondary threshold determined by the number of signals above

It can be seen that it is set to 4 to determine whether the signal is activated using these two thresholds.

As shown in FIG. 7, the activation signal determined according to the double threshold can be easily seen by the naked eye, as shown in FIG. There is an error in the determination of the time and the end time.

8 shows an example of an activation signal detection method (TTD) using a wide area observation window according to the present invention.

가 존재하도록 설정할 경우, 단위 광역 관찰창에서 상기 2차 문턱치를 초과하는 신호 갯수를 가지는 국부 관찰창의 갯수가

이상인지의 여부인 3차 문턱치를 이용하여 신호의 활성화 여부를 판단한다.Referring to FIG. 8, the activation signal detection method using the wide area observation window according to the present invention includes a local observation window within the wide area observation window.

Is set to exist, the number of local observation windows having the number of signals exceeding the secondary threshold in a unit wide observation window.

It is determined whether the signal is activated by using the third threshold, which is whether or not abnormality.

This can be seen that a large number of noises are excluded unlike in FIG. 7, and it is possible to determine not only the presence or absence of the activation signal, but also the accuracy of the determination of the start time and the end time of the signal.

Table 1 below shows the results of the activation signal detection ability according to the double threshold.

This results from detecting the activation signal using the double threshold method by including the Gaussian normal distribution with a white Gaussian noise with a signal-to-noise ratio of 3 dB.

을 5로 국부 관찰창의 크기를 정하고, 정해진 관찰창 내에서 1차 문턱치를 초과한 신호의 갯수로 정해지는 이차 문턱치

을 4로 설정하여 이 두 개의 문턱치를 이용하여 신호의 활성화 여부를 판단하였다.Here, the mean of the original signal is 0.5 and the standard deviation is 0.08, and the number of signals in the signal containing the white Gaussian noise

The size of the local observation window is set to 5, and the secondary threshold determined by the number of signals exceeding the primary threshold within the given observation window.

Set to 4 to determine whether the signal is activated using these two thresholds.

Here, although the actual value of 231, which is actually activated, was correctly determined among the 1001 signals used, it can be seen that the error due to noise is 103, which causes a considerable error.

을 상기 광역 관찰창 내에 5가 존재하도록 설정하고, 3차 문턱치

은 국부 관찰창의 갯수가 3개 이상이 되도록 설정하였다.Table 2 below shows the results of the activation signal detection capability for the activation signal detection method using the wide-area observation window according to the present invention under the same conditions as in Table 1. In the activation signal detection method using the wide area observation window according to the present invention, the conditions are the same as in Table 1, and the local observation window exists in the wide area observation window.

Is set to 5 in the wide field of view, and the third threshold

Is set so that the number of local observation windows is three or more.

Referring to Table 2, the value of the activation is determined to be 211, which is less than the detection method of the activation signal by the double threshold, but the error due to noise is 8 and the effect of the noise compared to 103, which is the error due to the noise by the double threshold. It can be seen that the error is reduced by about 13 times.

Comparison of the determination ability of the activation signal with respect to Table 1, which is a result of the activation signal detection capability according to the double threshold, and Table 2, which is a result of the activation signal detection capability of the activation signal detection method using the wide-area observation window according to the present invention. Values for sensitivity, specificity, precision, and accuracy are summarized in Table 3 below.

을 더 증가시켜서 로그에 저장함으로써, 활성화 신호 정보의 카운트 수와 실제로 발생한 이벤트의 수의 차이를 줄일 수 있으며, 그 외의 특성은 본 발명에 따른 광역 관찰창을 이용한 활성화 신호 검출 방법이 월등히 우수한 성능을 나타내고 있음을 알 수 있다.Referring to Table 3, when comparing the activation signal detection method (TTD) using the wide area observation window according to the present invention and the activation signal detection method (DTD) according to the double threshold, the activation signal detection method according to the present invention is a double threshold value. The sensitivity seems to be lowered in the sensitivity compared to the activation signal detection method, but the number of local observation windows included in the global observation window

By further increasing and storing in the log, it is possible to reduce the difference between the number of counts of the activation signal information and the number of events actually occurred, and the other characteristics are that the activation signal detection method using the wide-area observation window according to the present invention has excellent performance. It can be seen that.

In particular, it can be seen that the number of signals indicating errors among the total number of signals 1001 is 28, and the probability of not making an error is 99%, indicating an excellent activation signal detection capability.

On the other hand, embodiments of the present invention can be configured in various ways.

Since the activation signal detection method using the wide viewing window according to the present invention can minimize the effects of noise without using a filter, the above-mentioned EMG measurement, elongation measurement, snoring measurement, as well as a vibration plate installed on the road vehicle It is possible to investigate the loading load of and to analyze the floating population of important commercial areas by installing the diaphragm on the sidewalk block as well.

In addition, it can be applied to medical devices such as blood pressure monitors, and since the blood pressure is measured periodically, applying the activation signal detection method based on this enables accurate blood pressure measurement, and does not constitute a filter. The activation signal detection apparatus according to the invention can be variously applied.

The invention can be implemented via software. When implemented in software, the constituent means of the present invention are code segments that perform the necessary work. The program or code segments may be stored on a processor readable medium or transmitted by a computer data signal coupled with a carrier on a transmission medium or network.

The computer-readable recording medium includes all kinds of recording devices in which data is stored which can be read by a computer system. Examples of computer-readable recording devices include ROM, RAM, CD-ROM, DVD ± ROM, DVD-RAM, magnetic tape, floppy disks, hard disks, optical data storage devices, and the like. The computer readable recording medium can also be distributed over network coupled computer devices so that the computer readable code is stored and executed in a distributed fashion.

Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary and will be understood by those of ordinary skill in the art that various modifications and variations can be made therefrom. However, such modifications should be considered to be within the technical protection scope of the present invention. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

1 illustrates an example of a typical sinusoidal signal.

2 is a graph illustrating an example of a signal detection method using a single threshold applied to signal measurement.

3 is a graph illustrating an example of a signal detection method using a double threshold applied to signal measurement.

4 is a flowchart illustrating a method of detecting an activation signal using a wide area observation window according to the present invention.

Figure 5 shows an example of the activation signal detection method using a wide area observation window according to the present invention.

6 illustrates an activation signal detection apparatus using a wide area observation window according to the present invention.

7 illustrates an activation signal determined according to a double threshold.

8 shows an example of an activation signal detection method using a wide-area observation window according to the present invention.

Claims (14)

Outputting a signal waveform measured from an object to be tested, determining start and end time points of the output signal waveform, and counting the number of signals within the determined time point; Setting a size of a wide area observation window including a local observation window and the local observation window using the object to be examined and the number of counted signals, and determining a first threshold according to an amplitude of the measured signal waveform; Setting a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and counting the number of local observation windows having a signal number exceeding the set secondary threshold in the global observation window; Determining a third threshold according to the counted number of local observation windows, and detecting an activation signal by detecting a wide area observation window exceeding the third threshold; And Activation signal information of the object under test including a start time and an end time of the counted wide area observation window, a start time and an end time of the output signal waveform, the count of the counted wide area observation window, and an identifier of the object to be examined are recorded in a log. Activation signal detection method using a wide area observation window comprising the step of storing. The method of claim 1, The signal waveform is An activation signal detection method using a wide area observation window, characterized in that the signal waveform having a low signal-to-noise ratio. The method of claim 1, Start and end time points of the output signal waveform is The activation signal detection method using a wide area observation window, characterized in that set based on the identifier of the object to be stored in the log. The method of claim 1, The step of storing the activation signal information in a log Among the counted wide area observation window, which is included in the counted wide area observation window, the signal having the fastest time value among the signals exceeding the first threshold of the counted local observation window is stored as a start time of the activation signal. step; And Storing, as an end point of the activation signal, a signal having a latest time value included in the last counted wide area observation window among the counted wide area observation windows and exceeding the first threshold of the counted local observation window; Activation signal detection method using a wide area observation window comprising a. The method of claim 1, The signal waveform is An activation signal detection method using a wide-area observation window, comprising a sinusoidal signal including an electrocardiogram, an electrocardiogram, an electrocardiogram, or a microphone output. The method of claim 1, Comprising: comparing the activation signal information of the subject and the event actually generated to calculate a reception operation characteristic, and resetting the number of local observation window of the third threshold in accordance with the reception operation characteristic and storing it in the log; Activation signal detection method using a wide area observation window, characterized in that. delete A computer-readable recording medium having recorded thereon a program for executing the method of any one of claims 1 to 6. An output signal counting unit configured to output a signal waveform measured from a test target, determine start and end time points of the output signal waveform, and count the number of signals within the determined time point; An observation window setting unit configured to set a size of a wide area observation window including a local observation window and the local observation window by using the object to be examined and the number of counted signals; Determine a primary threshold according to the amplitude magnitude of the measured signal waveform, set a secondary threshold of the local observation window according to the number of signals exceeding the determined primary threshold, and set the secondary threshold in the global observation window. A threshold setting unit for determining a third threshold according to the number of local observation windows having an excessive number of signals; An activation signal detection unit for counting the number of local observation windows having a signal number greater than or equal to the secondary threshold, and detecting an activation signal by detecting a wide area observation window in which the counted local observation windows exceed the third threshold; And Storing the activation signal information of the subject under test including a start time and an end time of the counted wide area observation window, a start time and an end time of the output signal waveform, a count of the counted wide area observation window, and an identifier of the subject to be examined; Activation signal detection apparatus using a wide area observation window including a log storage unit. The method of claim 9, The signal waveform is Activation signal detection apparatus using a wide area observation window, characterized in that the signal waveform having a low signal-to-noise ratio. The method of claim 9, Start and end time points of the output signal waveform is The activation signal detection apparatus using a wide-area observation window, characterized in that set based on the activation signal information of the object to be stored stored in the log storage unit. The method of claim 9, The log storage unit A signal having a fastest time value among the signals exceeding the first threshold of the counted local observation window, which is included in the counted wide area observation window among the counted wide area observation windows, is stored as a start time point of the activation signal; , A signal which is included in the last counted wide area observation window among the counted wide area observation windows and which has the latest time value among signals exceeding the first threshold of the counted local observation window, as an end point of the activation signal; The activation signal detection apparatus using a wide area observation window, characterized in that it further comprises a timing recording module. The method of claim 9, The log storage unit Comprising a wide range of observation characterized in that the feedback module for comparing the activation signal information of the test target and the event actually generated to calculate the reception operation characteristics, and resets the number of local observation window of the third threshold in accordance with the reception operation characteristics Activation signal detection device using a window. delete

Images