TW201900110A - Sleep analysis method capable of effectively and automatically analyzing sleep condition/habit of user - Google Patents

Abstract

The present invention provides a sleep analysis method, which includes the following steps: sensing a three-axis acceleration signal related to a user; calculating activity intensity values corresponding to multiple timings according to the three-axis acceleration signal; calculating a plurality of first sleep scores and second sleep scores according to the activity intensity values; determining a sleep interval according to the sleep scores; further determining a sleep timing and an awake timing corresponding to the sleep interval; and, analyzing the sleep quality corresponding to the sleep interval according to the activity intensity values corresponding to the sleep interval. In particular, with the aforementioned method, the present invention can effectively and automatically analyze the sleep condition/habit of the user.

Description

Sleep analysis method

The invention relates to a sleep analysis method, in particular to a sleep analysis method capable of automatically analyzing a user's sleep condition/habit.

With the rapid development of wearable electronic devices, it is now possible to record sleep information, including the start time and end time of sleep, through the wearable device to analyze the sleep habits of the user on a regular basis so that the user can analyze according to the analysis. The result is to understand and/or adjust your sleep habits; this can help to develop good habits and avoid chronic diseases.

However, when using the existing wearing device for sleep analysis, it is necessary to manually operate the wearing device before going to bed to manually set the start time of sleep, and it is necessary to operate the wearing device after waking up to manually set the end time of sleep, except for inconvenience. In addition, it is also possible to forget to set the aforementioned start time and end time without storing the corresponding time record.

Accordingly, it is an object of the present invention to provide a sleep analysis method that can automatically and automatically analyze a user's sleep condition/habit.

Thus, the sleep analysis method of the present invention is implemented by a detection system including a wearable device adapted to be worn on a user. The sleep analysis method comprises a step (a), a step (b), a step (c), a step (d) and a step (e).

The step (a) is that the wearing device senses an acceleration signal.

The step (b) is that the detecting system calculates a pair of active intensity values of the time points at each time point of the plurality of time points, wherein each time point corresponds to the plurality of signal samples of the acceleration signal and the activity intensity value is according to the The signal samples corresponding to the time points are calculated.

The step (c) is that the detecting system uses a time window that moves along the time axis, and for each time point, calculates a first sleep score and a second sleep score for a pair of time points, wherein the first sleep The score is the number of time points at which the activity intensity value corresponding to the time threshold is less than a first activity threshold at all time points covered by the time window, the second sleep score being the activity corresponding to all the time points covered by the time window The amount of time when the intensity value is less than a second active threshold, and the first active threshold is greater than the second active threshold.

The step (d) is when the detecting system determines that the first sleep score corresponding to the plurality of consecutive time points is greater than a first sleep interval threshold, and the second time point corresponding to the at least one time point of the consecutive time points When the sleep score is greater than a second sleep interval threshold, the detection system determines that the earliest time of the consecutive time points is a starting point of a sleep interval, wherein the second sleep interval threshold is greater than the first sleep interval valve value.

The step (e) is: when the detecting system determines that the first sleep score corresponding to a starting point of the sleep interval is less than a awake threshold, the detecting system determines that the time is the sleeping interval. The end point.

The effect of the invention is that the sleep condition/habit of the user can be automatically and automatically analyzed.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are denoted by the same reference numerals.

Referring to Figures 1 and 2, the sleep analysis method of the present invention is implemented by a detection system including a wearable device and a computing device. The wearing device is, for example, a smart watch suitable for being worn on a wrist of a user, and has a three-axis acceleration gauge; the computing device is, for example, a smart phone, and can be connected via a wired universal asynchronous transmission transmitter (Universal Asynchronous) Receiver/Transmitter, UART) or Universal Serial Bus (USB), or wireless WiFi or Bluetooth protocol to communicate with the wearable device.

Referring to Fig. 1, a flow chart of an embodiment of the sleep analysis method of the present invention will be described below.

First, in step S1, the wearable device senses a three-axis acceleration signal by the three-axis acceleration gauge, wherein the sampling frequency of the three-axis acceleration signal is 20 Hz, and includes an X-axis signal, a Y-axis signal, and a Z Axis signal.

Next, in step S2, the wearer for each signal sample S( t ) of the triaxial acceleration signal, according to the X-axis sample Ax( t ) of the signal sample, the Y-axis sample Ay( t ) and the Z-axis sample Az ( t ) calculating a signal strength value M( t ), wherein . The computing device calculates a pair of active intensity values for each of the plurality of time points based on the signal strength values. Preferably, the interval between two adjacent time points is one minute, that is, each time point corresponds to a plurality of signal samples and a plurality of signal strength values, and the computing device calculates the signal strengths corresponding to the minutes every one minute. The value calculates an activity intensity value. Referring to FIG. 2, for each time point, the activity intensity corresponding to the time point is the number of peak points P of the curve 1 formed by all the signal strength values corresponding to the time point; wherein, as shown in FIG. 2, each peak point P is A local maximum of curve 1 and a difference from each adjacent local minimum M exceeds a threshold.

Next, in step S3, referring to FIG. 3 and FIG. 4, the computing device uses a time window 2 moving along the time axis, and for each time point, calculates a first sleep score 4 and a second for a pair of time points. a sleep score of 5, wherein the first sleep score 4 is a number of time points at which the corresponding activity intensity value is less than a first activity threshold at all time points covered by the time window 2, and the second sleep score 5 is The activity intensity value corresponding to all time points covered by the time window 2 is less than the number of time points of a second activity threshold, and the first activity threshold is greater than the second activity threshold. For example, the time window 2 has a length of 20 minutes, and each time point corresponds to one minute. Therefore, the time window 2 corresponding to the 20th minute T ₂₀ covers the time from the first minute T ₁ to the twentieth minute T ₂₀ . 20 hours. Similarly, the time corresponding to a twenty-first minute time window 2 is covered by a second one minute to twenty minutes, ten minutes time T ₃₀ corresponding to the third time window 2 is covered by the eleventh to thirty minutes The time T2 corresponding to the time T2 (2) corresponding to the minute T ₃₀ and the fortieth minute T ₄₀ is the twenty-first minute to the fortieth minute T ₄₀ and the like. As shown in FIG. 3, the activity intensity value corresponding to the 20 time points covered by the time window 2a (2) corresponding to the _40th minute T ₄₀ is less than the number of the first activity threshold is 10, so the fortieth minute The first sleep score corresponding to T ₄₀ is 10.

Next, in step S4, referring to FIG. 4, the computing device determines whether the triaxial acceleration signal has a sleep interval. When the computing device determines that there are multiple consecutive time points, for example, thirty time points, and each time point corresponds to one minute, the corresponding first sleep score 4 is greater than a first sleep interval threshold, and the consecutive time points are When the second sleep score 5 corresponding to the at least one time point is greater than a second sleep interval threshold, the computing device determines that the three-axis acceleration signal exists in the sleep interval 3, and the earliest of the consecutive time points is the sleep a starting point T _{i of the} interval, wherein the second sleep interval threshold is greater than the first sleep interval threshold. And when the computing device determines that the first sleep score 4 corresponding to a time point shorter than the starting point T _i of the sleep interval is less than a awake threshold, the computing device determines that the time point is the end of the sleep interval 3 T _j .

In another embodiment, the wearing state of the wearing device can also be determined by the triaxial acceleration signal, and the wearing state is further utilized to assist in determining the starting point and the ending point of the sleeping interval. In detail, for each time point, if the computing device determines that the activity intensity value corresponding to the time point is greater than an activity intensity threshold value and all the signal samples corresponding to the time point meet the "first axis and gravity in the three-axis direction" The number of signal samples in which the direction is substantially parallel" and the "other signal values of the other axes different from the first axis approach zero" is less than a threshold value, and the computing device determines that the wearable device is at the time point The worn state; otherwise, the wearable device is in an unworn state at that point in time. For example, as shown in FIG. 5, when the wearing device 6 is flat, the Z axis of the wearing device 6 is parallel to the direction of gravity, so in this case, the X-axis sample and Y of the three-axis acceleration signal are sensed. The values of the axis samples approach zero. As shown in FIG. 6, when the wearing device 6 is placed sideways, the X axis of the wearing device 6 is parallel to the direction of gravity, so in this case, the values of the Z-axis sample and the Y-axis sample of the triaxial acceleration signal are sensed. Both are approaching zero. Therefore, for a point in time, if the signal sample corresponding to the time point corresponds to "the first axis in the triaxial direction is substantially parallel to the direction of gravity" and the signal value in the other axis different from the first axis The number of signal samples that are both near zero and the two conditions is less than a threshold value at which the wearing state of the wearable device is likely to be the worn state. Further using the wear state to assist in determining the start point and the end point of the sleep interval: when the computing device determines that there are multiple consecutive time points, the first sleep score is greater than the first sleep interval threshold and the consecutive time points When the second sleep score corresponding to at least one time point is greater than the second sleep interval threshold and at least one of the consecutive time points is corresponding to the wearable state, the computing device determines that the corresponding one of the consecutive time points is The earliest time of the wearing state is the starting point of the sleeping interval; when the computing device determines that there are a plurality of consecutive time points, for example, ten time points, all of which are corresponding to the wearing state, the computing device determines that the consecutive time points are The earliest time is the end of the sleep interval.

Next, in step S5, the computing device further determines a sleep time point and an awake time point corresponding to the sleep interval 3.

In detail, the computing device determines whether the curve formed by the first sleep scores 4 corresponding to the isochronous points has a valley point earlier in time than the starting point T _i of the sleep interval, as shown in FIG. When the curve is determined to have the valley point V, the computing device determines that the time point corresponding to the valley point V is a sleep time point T _s ; when it is determined that the curve does not have the valley point, the computing device determines that The point in time that is earlier than the starting point T _{i of} the sleep interval and the corresponding first sleep score 4 is greater than a point of entry into the sleep threshold is the closest to the starting point T _i .

The computing device determines that the end point T _j that is earlier than the end point T _{j of} the sleep interval and the corresponding first sleep score 4 is less than a awake threshold value is the closest to the end point T _j of the sleep interval. _e .

Moreover, in another embodiment, the sleep quality of the user in the sleep zone 3 can be further analyzed. In detail, for each time point in the sleep interval 3, when the activity intensity value corresponding to the time point is greater than a first threshold value, the computing device determines that the time point corresponds to a first awake score; when the time point corresponds to When the activity intensity value is greater than a second threshold value and less than the first threshold value, the computing device determines that the time point corresponds to a second awake score; when the activity intensity value corresponding to the time point is greater than a third threshold value and less than the When the second threshold is used, the computing device determines that the time point corresponds to a third awake score; when the activity intensity value corresponding to the time point is less than the third threshold value, the computing device determines that the time point corresponds to a fourth awake score; The threshold values are from the largest to the smallest, the first threshold, the second threshold, the third threshold, and the fourth threshold, and the awake scores are the first awake score and the second from the largest to the smallest. The awake score, the third awake score, and the fourth awake score, wherein the lower the awake score, the better the sleep quality.

In addition, in addition to the foregoing calculating the awake score corresponding to each time point, the corresponding awake score may also be calculated for a time period including a plurality of time points. For example, each time point corresponds to one minute, and the computing device calculates the sum of the awake scores corresponding to the ten time points covered by the ten minutes as the awake score corresponding to the ten minutes every ten minutes; thus, the figure 7 can be generated. The sleep quality map corresponding to sleep interval 3 is shown, wherein the lower the awake score, the better the sleep quality.

Further, although the wearable device and the computing device are each responsible for partial calculations in the foregoing embodiments, they are not limited thereto. For example, the detection system may also include only the wearable device, and the wear device is responsible for all operations.

In summary, the sleep analysis method of the present invention calculates a three-axis acceleration signal related to the user, and calculates an activity intensity value corresponding to the plurality of time points according to the three-axis acceleration signal, and according to the activity intensity Calculating a plurality of first sleep scores and a second sleep score, and determining a sleep interval according to the sleep scores, and further determining a sleep time point and an awake time point corresponding to the sleep interval, and according to the sleep interval corresponding to the sleep interval The activity intensity value analyzes the sleep quality corresponding to the sleep interval, and can effectively analyze the sleep condition/habit of the user automatically, so that the object of the present invention can be achieved.

However, the above is only the embodiment of the present invention, and the scope of the invention is not limited thereto, and all the equivalent equivalent changes and modifications according to the scope of the patent application and the patent specification of the present invention are still The scope of the invention is covered.

1‧‧‧ Curves formed by signal strength values

2‧‧‧Time window

2a‧‧‧Time window

3‧‧‧sleeping interval

4‧‧‧First sleep score

5‧‧‧Second sleep score

6‧‧‧Wearing device

The starting point of the sleep interval in T _i ‧‧

End point of the sleep interval of T _j ‧‧

T _s ‧‧‧ When you fall asleep

T _e ‧‧‧When awake

P‧‧‧ peak

V‧‧‧ Valley Point

M‧‧‧ local minimum

T ₁ ‧‧‧First minute

T ₂₀ ‧‧‧Twenty minutes

T ₃₀ ‧‧‧30th minute

T ₄₀ ‧‧‧ forty minutes

S1~S5‧‧‧Steps

Other features and effects of the present invention will be apparent from the embodiments of the present invention, wherein: FIG. 1 is a flow chart illustrating an embodiment of the sleep analysis method of the present invention; FIG. 2 is a schematic diagram illustrating Figure 3 is a schematic diagram illustrating the use of a time window moving along the time axis to calculate the sleep score; Figure 4 is a schematic diagram illustrating a sleep interval and a corresponding sleep time point and an awake time point; Figure 5 is a schematic view showing a situation in which a wearable device is placed flat; Figure 6 is a schematic view showing the side of the wearable device; and Figure 7 is a schematic view showing a sleep quality map corresponding to the sleep interval.

Claims (7)

A sleep analysis method is implemented by a detection system including a wearable device, the wearable device being adapted to be worn on a user, the sleep analysis method comprising the following steps: (a) the wear device senses an acceleration signal; b) the detection system calculates a pair of active intensity values for each time point for each time point of the plurality of time points, wherein each time point corresponds to a plurality of signal samples of the acceleration signal and the activity intensity value is corresponding to the time point The signal sample is calculated; (c) the detection system uses a time window that moves along the time axis, and for each time point, calculates a first sleep score and a second sleep score for a pair of time points, wherein The first sleep score is a number of time points when the corresponding activity intensity value is less than a first activity threshold value at all time points covered by the time window, and the second sleep score is at all time points covered by the time window The corresponding activity intensity value is less than a second activity threshold value, and the first activity threshold is greater than the second activity threshold; (d) when the detection system determines When the first sleep score corresponding to the plurality of consecutive time points is greater than a first sleep interval threshold and the second sleep score corresponding to at least one of the consecutive time points is greater than a second sleep interval threshold, the detect The measurement system determines that the earliest time in the consecutive time points is the starting point of a sleep interval, wherein the second sleep interval threshold is greater than the first sleep interval threshold; and (e) when the detection system determines that the night is late When the first sleep score corresponding to a certain point of the starting point of the sleep interval is less than a awake threshold, the detecting system determines that the time point is the end point of the sleep interval. The sleep analysis method of claim 1, further comprising a step (f): for each time point, if the detection system determines that the activity intensity value corresponding to the time point is greater than an activity intensity threshold, the detection system Determining that the wearing device is in a worn state at the time; wherein, in the step (d), the detecting system determines that the first sleep score corresponding to the plurality of consecutive time points is greater than the first sleep interval threshold And when the second sleep score corresponding to at least one of the consecutive time points is greater than the second sleep interval threshold and at least one of the consecutive time points is corresponding to the worn state, the detecting system determines the continuous The earliest time in the time point corresponding to the state of being worn is the starting point of the sleep interval. The sleep analysis method according to claim 2, wherein the wear device has a three-axis acceleration gauge, wherein in the step (a), the acceleration signal is sensed by the three-axis acceleration gauge and the acceleration signal is a three-axis acceleration signal And in the step (f), for each time point, if the detecting system determines that the activity intensity value corresponding to the time point is greater than the activity intensity threshold value and all the signal samples corresponding to the time point meet the three-axis direction The detection system determines that the wearable device is in the worn state at the point in time when the first axial direction is substantially parallel to the direction of gravity and the number of signal samples whose other axial signal values approach zero is less than a threshold value. The sleep analysis method according to claim 1, wherein the wear device has a three-axis acceleration gauge, wherein in the step (a), the acceleration signal is sensed by the three-axis acceleration gauge and the acceleration signal is a three-axis acceleration signal Each signal sample of the acceleration signal includes an X-axis sample, a Y-axis sample, and a Z-axis sample; in the step (b), for each signal sample corresponding to each time point, the detection system is based on the signal A signal intensity value is calculated for the X-axis sample, the Y-axis sample, and the Z-axis sample of the sample, and for each time point, the activity intensity corresponding to the time point is the peak point of the curve formed by the signal intensity values corresponding to the time point. quantity. The sleep analysis method of claim 1, further comprising a step (g): the detecting system determines whether the curve formed by the first sleep scores corresponding to the isochronous points has a time earlier than the sleep The valley point of the starting point of the interval, when it is determined that the curve has the valley point, the detecting system determines that the time point corresponding to the valley point is a sleep time point; when it is determined that the curve does not have the valley point, The detecting system determines that the closest to the starting point in the time point that is earlier than the starting point of the sleeping interval and the corresponding first sleep score is greater than a threshold value of entering the sleep threshold is the point of falling into sleep. The sleep analysis method according to claim 1, further comprising a step (h): the detecting system determines the time point that is earlier in time than the end of the sleep interval and the corresponding first sleep score is less than a awake threshold The closest to the end point is a waking point. The sleep analysis method of claim 1, further comprising a step (i): for each time point in the sleep interval, when the activity intensity value corresponding to the time point is greater than a first threshold value, the detection system Determining that the time point corresponds to a first awake score; when the activity intensity value corresponding to the time point is greater than a second threshold value and less than the first threshold value, the detecting system determines that the time point corresponds to a second awake score; When the activity intensity value corresponding to the time point is greater than a third threshold value and less than the second threshold value, the detection system determines that the time point corresponds to a third awake score; when the activity intensity value corresponding to the time point is less than the third threshold The value of the threshold value corresponds to a fourth awake score; The awake scores are the first awake score, the second awake score, the third awake score, and the fourth awake score from large to small.