CN109528159A - A kind of sleep quality and respiration monitoring system and method based on bed body - Google Patents
A kind of sleep quality and respiration monitoring system and method based on bed body Download PDFInfo
- Publication number
- CN109528159A CN109528159A CN201811243454.6A CN201811243454A CN109528159A CN 109528159 A CN109528159 A CN 109528159A CN 201811243454 A CN201811243454 A CN 201811243454A CN 109528159 A CN109528159 A CN 109528159A
- Authority
- CN
- China
- Prior art keywords
- signal
- original signal
- sleep
- pressure sensor
- original
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6887—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient mounted on external non-worn devices, e.g. non-medical devices
- A61B5/6891—Furniture
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
- A61B5/0826—Detecting or evaluating apnoea events
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1118—Determining activity level
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4806—Sleep evaluation
- A61B5/4809—Sleep detection, i.e. determining whether a subject is asleep or not
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Physiology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Dentistry (AREA)
- Anesthesiology (AREA)
- Pulmonology (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a kind of sleep quality based on bed body and respiration monitoring system and methods.Four cabinet base lower ends of bed body of this system are respectively mounted pressure sensor, pressure sensor is sequentially connected A/D conversion module, microprocessor, wireless module by conducting wire, the WiFi module and bluetooth module of wireless module obtain original signal from microprocessor by serial ports, original signal is sent network data base by WiFi module, and bluetooth module sends mobile phone for original signal by Bluetooth protocol.Monitoring method is mainly to microprocessor reset under empty bed state；Pressure sensor signal carries out anticipation by microprocessor through A/D conversion module and has no progeny, it is sent to network data base or mobile phone, to carry out feature extraction to original signal, to obtain sleep signal and breath signal, and then the detection of apnea phenomenon and the statistical analysis of respiratory rate are completed.The present invention has at low cost, the simple feature of structure, is suitable for the applications such as residential care and hospital's monitoring.
Description
Technical field
The present invention relates to a kind of sleep quality and respiration monitoring systems, more particularly, to a kind of can analyze based on bed body
Specific behavior of sleeping and the system and method for carrying out breath signal statistical analysis.
Background technique
There are many detection methods at present to detect sleep quality, and medically most common method is sleep analysis monitor
(PSG), sleep analysis monitor (PSG) is always treated as being sleep monitor " goldstandard ", and PSG is mainly medically using precision
The physiological characteristics such as equipment combination blood pressure, heartbeat, brain wave accurately make the evaluation of science to sleep quality state.But PSG
Patient is needed to wear many equipment, it is also desirable to slept the whole night in hospital, cost is also higher, both influences sensory experience, can not also answer
It is detected used in family.
In addition, there are also the methods for carrying out sleep detection using the dynamic meter of wrist, the dynamic meter of leg and 3-axis acceleration sensor.These
For method compared to PSG, cost wants much lower and simple portable, can be applied to family's detection, but does not still solve
The problem of needing wearable device to influence sensory experience.Meanwhile the dynamic meter of wrist and the dynamic meter of leg there is also an issue be exactly cannot be automatic
The time for determining the sleep of user's beginning and end, need user's manual setting.Therefore a kind of contactless sleep is needed
Behavioral value method, may be implemented the sleep detection of multi-angle, while have the function of breast rail, and system cost is also dropped
It is low.
Summary of the invention
For the deficiency in background technique, the purpose of the present invention is to provide a kind of sleep quality based on bed body and breathings
Monitoring system and method.
To achieve the goals above, the technical solution adopted by the present invention is that:
One, sleep quality and respiration monitoring system based on bed body
The lower end of four cabinet bases of bed body is respectively mounted pressure sensor, and four pressure sensors are commonly connected to A/ by conducting wire
D conversion module, A/D conversion module are connected to wireless module through microprocessor, and wireless module includes WiFi module and bluetooth module,
WiFi module and bluetooth module pass through serial ports and obtain original signal from microprocessor, and original signal is sent net by WiFi module
Network database, bluetooth module send mobile phone for original signal by Bluetooth protocol.
Pressure sensor is fixed under cabinet base, and pressure sensor is for acquiring pressure signal, by collected pressure signal
Resistance value is converted to, and then is converted to output voltage values.
A/D conversion module includes regulated power supply, on-chip clock oscillator, for by the output voltage values of pressure sensor into
Row amplification and A/D conversion.
Microprocessor include it is multiple rapidly input/export I/O mouthfuls, it is original after A/D module processing for reading
Signal is simultaneously judged.
Sleep behavior detection algorithm carries out multilayer decomposition using barycentric coodinates of the wavelet decomposition to original signal, extracts small echo
Detail coefficients and variance are classified, so that sleep signal is divided into just as judging characteristic using the method for support vector machines
Normal sleep signal and leg move signal.
It breathes statistic algorithm and windowing process is carried out to original signal, according to the threshold value of setting by breath signal from original signal
In extract, then filtered breath signal is obtained by filtering, and count respiration rate and detect apnea and show
As realizing a variety of unawares detection of sleep procedure.
Two, a kind of the step of monitoring method of sleep quality and respiration monitoring system based on bed body, this method, is as follows:
1) pressure sensor is resetted under empty bed state.
Using a foot of cabinet base as coordinate origin, the two cabinet base directions adjacent using its are positive as X-axis forward direction and Y-axis
Establish rectangular coordinate system, barycentric coodinates calculated according to the reading of the pressure sensor under four cabinet bases, barycentric coodinates it is horizontal, vertical
Coordinate x, y can be obtained according to lever principle:
Wherein, w1, w2, w3, w4 are respectively the reading of the pressure sensor under four cabinet bases.
The reading of empty bed state lower pressure sensor is recorded as fur weight, activity causes pressure to pass to user in bed
The reading of sensor changes, and the reading that pressure sensor changes every time is subtracted the reading of fur weight as original signal.
2) original signal of pressure sensor sentences original signal by microprocessor after the conversion of A/D conversion module
It is disconnected.
Be arranged flag bit S, if the original signal twice in succession received all greater than index threshold T, by flag bit S
Set 1, it is believed that user starts to sleep；If original signal twice in succession sets 0 all less than index threshold T, by flag bit S,
Think that user has got up.
In view of the weight and actual test of normal human, 2 kilograms are set by index threshold T.
3) network data base or mobile phone, net are sent by WiFi module or bluetooth module by original signal and flag bit S
Network database or mobile phone carry out feature extraction to original signal to obtain sleep signal and breath signal.
The extraction of sleep signal is based on sleep behavior detection algorithm and sleep signal is divided into ortho signal and the dynamic letter of leg
Number, specifically: using wavelet decomposition to the barycentric coodinates of the original signal of acquisition carry out multilayer decompose to obtain barycentric coodinates cross,
Approximation coefficient and detail coefficients after ordinate wavelet decomposition choose the detail coefficients of layer 7 as fisrt feature, by center of gravity
The cross of coordinate, ordinate variance constitute 18 dimensional feature vectors and be used as second feature, when there are slopes greater than the in detail coefficients
When the variance of the peak value of one preset threshold and the cross of barycentric coodinates, ordinate is greater than the second preset threshold, it is believed that original signal is
Leg moves signal；Detail coefficients (x, y respectively have 8 values) after specific seven layers of wavelet decomposition for choosing center of gravity cross, ordinate, center of gravity
Horizontal, ordinate variance (2 values)) 18 dimensional feature vectors are constituted, it is dynamic to carry out classification acquisition leg using the method for support vector machines
The feature of signal occurs to automatically identify leg in ortho signal and move signal in some period.
The separation that breath signal and leg move signal is based on breast rail algorithm, specifically: breath signal threshold value H is set, it is right
Original signal carries out windowing process, calculates the difference of the maxima and minima of the original signal in each window, and difference, which is greater than, exhales
Think that the original signal in window belongs to leg and moves signal when inhaling signal threshold value H, thinks in window when difference is less than breath signal threshold value H
Original signal belong to breath signal；It abandons all legs that belong in original signal and moves the data of signal, and then isolated
Breath signal after separation is removed random disturbances by moving average filtering by breath signal, then by FIR low-pass filtering it
Signal afterwards obtains filtered breath signal.
4) statistical analysis of the detection of apnea phenomenon and respiratory rate:
The detection method of apnea phenomenon is similar with breathing statistic algorithm, specifically: apnea threshold value D is set, it is right
Original signal carries out windowing process, calculates the difference of the maxima and minima of the original signal in each window, and difference, which is greater than, exhales
Think that the original signal in window belongs to breath signal when inhaling suspend threshold D, thinks in window when difference is less than apnea threshold value D
Original signal belong to apnea phenomenon.
The method that the statistics of respiratory rate uses statistical signal peak value, to the filtered breath signal meter described in step 3)
Slope is calculated, a minor peaks are counted as when slope is by just becoming negative, the sum of peak value is the number breathed.
5) statistical result of the respiratory rate obtained according to step 4) draws out every point in sleep-respiratory frequency histogram
Respiration rate in clock, network data base or mobile phone are sent to the processing result of original signal as sleep behavior testing result
Mobile terminal, user by mobile terminal check sleep during respiratory rate situation of change.
The invention has the advantages that:
1) whole system it is simpler than medically common sleep analysis monitor (PSG), conveniently, low cost, than on the market
The equipment accuracy such as the dynamic meter of wrist it is higher；
2) it carries out sleep behavior using support vector machines by moving the extraction that signal carries out feature to collected leg
Classification realize that automatic identification leg moves behavior, accuracy rate is very high；Collected breath signal is also very clear, then to isolating
The breath signal that leg moves behavior is filtered, and the detection of Lai Jinhang apnea phenomenon and the statistical analysis of respiratory rate are right
Ortho and the dynamic twitch of leg differentiate accuracy rate and are up to 90% or more, and breathing and apnea judge precision close to 100%；
3) it acquire in entire data, transmit and processing system in terms of, there is at low cost, the simple feature of structure, be applicable in
It is detected in family.
Detailed description of the invention
Fig. 1 is system architecture diagram.
Fig. 2 is sleep behavior detection algorithm flow chart.
Fig. 3 is breathing statistic algorithm flow chart.
Fig. 4 is each layer detail coefficients figure of wavelet decomposition.
Fig. 5 is breast rail effect picture.
Fig. 6 is the dynamic separating effect figure with breath signal of leg.
Fig. 7 is finite impulse response (FIR) low-pass filter amplitude-frequency response figure.
Fig. 8 is sleep-respiratory frequency histogram.
In figure: 1. bed bodies, 2. pressure sensors, 3.A/D conversion module, 4. microprocessors, 5.WiFi module, 6. bluetooth moulds
Block, 7. network data bases, 8. mobile phones.
Specific embodiment
Present invention will be further explained below with reference to the attached drawings and examples.
As shown in Figure 1, the lower end of 1 four cabinet bases of bed body is respectively mounted pressure sensor 2, four pressure sensors 2 are by leading
Line is commonly connected to A/D conversion module 3, and A/D conversion module 3 is connected to wireless module through microprocessor 4, and wireless module includes
WiFi module 5 and bluetooth module 6, WiFi module 5 and bluetooth module 6 obtain original signal from microprocessor 4 by serial ports,
Original signal is sent network data base 7 by WiFi module 5, and bluetooth module 6 is sent original signal by Bluetooth protocol in one's hands
Machine 8.
Pressure sensor 2 is fixed under cabinet base, and collected pressure signal is converted to resistance value by pressure sensor 2, into
And be converted to output voltage values.
A/D conversion module 3 includes a variety of required peripheral circuits, such as regulated power supply, on-chip clock oscillator, is integrated
Height, fast response time, strong interference immunity are spent, A/D conversion module 3 is for amplifying the output voltage values of pressure sensor 2
Original signal is obtained with after A/D conversion process.
Microprocessor 4 rapidly inputs/exports I/O mouthfuls including multiple, and the requirement for meeting system turns for reading through A/D
Mold changing block 3 treated original signal is simultaneously judged.
Sleep behavior detection algorithm carries out multilayer decomposition using barycentric coodinates of the wavelet decomposition to original signal, extracts small echo
Detail coefficients and variance are classified, so that sleep signal is divided into just as judging characteristic using the method for support vector machines
Normal sleep signal and leg move signal.
It breathes statistic algorithm and windowing process is carried out to original signal, according to the threshold value of setting by breath signal from original signal
In extract, filtered breath signal is obtained by moving average filtering and FIR low-pass filtering, and count respiration rate with
And the phenomenon that detection apnea, realize a variety of unawares detection of sleep procedure.
A specific embodiment of the invention includes embodiment one and embodiment two, the first two of embodiment one and embodiment two
Step is essentially identical, the difference is that the microprocessor 4 of embodiment one sends original signal to network number by Wifi module 5
The processing of signal is carried out according to library 7；The microprocessor 4 of embodiment two sends original signal by bluetooth module 6 and carries out letter to mobile phone 8
Number processing.
Embodiment one: microprocessor 4 sends original signal to network data base 7 by Wifi module 5
The pressure that the present embodiment is used using the 2 collection voltages data of pressure sensor for being placed on four 1 underfooting of bed, this system
The model YZC-167 of force snesor 2,75 kilograms of range, total range of four sensors 2 is 300 kilograms.YZC-167 is one
Tested pressure conversion can be the variation of respective resistivity values, and then be converted to output voltage values by money strain pressure transducer
Variation.A/D conversion module 3 selects HX711A/D conversion module, does amplification to this voltage value and A/D is converted and be can be obtained
Voltage data.Here the pressure reading obtained is not accurate numerical value, is calibrated.
1) it will do it primary reset under empty bed state, record reading at this time as fur weight, read every time later
Number Shi Douhui subtracts this fur weight, and the reading that processing obtains so is just entirely the nt wt net weight as caused by human body, avoids
The error due to caused by the extraneous factors such as placement position of article on bed.
It is the analysis based on center of gravity that empty bed 1, which resets the method used, it is therefore desirable to turn the reading of four pressure sensors 2
It is changed to the reading of center of gravity, using a foot of bed 1 as coordinate origin, two beds 1 are respectively that X-axis forward direction and Y-axis forward direction are established along direction
Rectangular coordinate system, four pressure sensors 2 are placed in four 1 underfooting of bed.According to the pressure sensor 2 of four 1 underfooting of bed
Reading can calculate barycentric coodinates, respectively remember center of gravity cross, ordinate x, y, according to lever principle:
Wherein w1, w2, w3, w4 are respectively the reading of four cabinet base lower pressure sensors.
Activity can cause the reading of pressure sensor 2 to change to user in bed, the reading that pressure sensor 2 is read every time
Number subtracts the reading of fur weight under initial empty bed state as original signal to be processed；
2) original signal of pressure sensor 2 by A/D conversion module 3 conversion after by microprocessor 4 to original signal into
Row judgement:
In view of at present on the market many sleep detection equipment cannot judge automatically user's beginning and end sleep when
Between, this system devises a kind of simple judgment method, and STM32 microprocessor can be selected in microprocessor 4, in STM32 micro process
After device gets original signal, judged first, if original signal twice in succession is incited somebody to action all greater than index threshold T
Flag bit S sets 1, it is believed that user starts to sleep；If original signal twice in succession, will mark all less than index threshold T
Position S sets 0, it is believed that user has got up.In view of the weight and actual test of normal human, index threshold T is arranged
It is 2 kilograms.
The sample frequency of this system is 10Hz, and leg moves the duration of signal at 2 seconds or so, and the frequency of breath signal exists
0.2 arrives 0.8Hz or so, so the sample frequency of 10Hz is fully able to meet system requirements.
3) microprocessor 4 is by reading w1, w2, w3, w4 of four pressure sensors 2 and flag bit S with object numbered musical notation
(Json) format is sent to network data base 7 by Wifi module 5, network data base 7 to original signal carry out feature extraction from
And obtain sleep signal and breath signal.
Network data base 7 has selected the message queue telemetering of open source to transmit (MQTT) message server EMQ, and EMQ is based on
Erlang/OTP language platform supports million grades of connections and distributed type assemblies to provide the MQTT message based on publish/subscribe mode
Server supports MQTT V3.1/V3.1.1 protocol specification.The upload of EMQ message server data can achieve 10~15 times often
Second, it also can satisfy the requirement of system.Data storage then uses MySQL database, and the processing of data then uses local various works
Tool, such as Matlab can thus build the completely free transmission of MQTT data and processing system.
Original signal is published to MQTT server by microprocessor 4, is ordered in computer end using the client that Python writes
Message is read, original signal is got.Then feature extraction is carried out to obtain sleep signal and breath signal to original signal.
As shown in Fig. 2, the extraction of sleep signal is based on sleep behavior detection algorithm, respectively from time domain and the angle of frequency domain from
Extract following characteristics in original signal, sleep signal be divided into ortho signal and leg and moves signal, its purpose is to from
The part that may represent that leg is dynamic and twitches is distinguished in original signal, sleep behavior is divided into following two categories: the first kind is just
Stable state often is kept when sleep, is generated without any Large Amplitude Motion；Second class includes swing and whole body or the office of leg
The twitch in portion (such as leg).
Sleep behavior detection algorithm specifically: carry out multilayer using barycentric coodinates of the wavelet decomposition to the original signal of acquisition
Approximation coefficient and detail coefficients after decomposing the cross for obtaining barycentric coodinates, ordinate wavelet decomposition, cross, the ordinate of barycentric coodinates
Detail coefficients (x, y respectively have 8 values) after seven layers of wavelet decomposition of (x, y), 18 Wei Te of variance composition of center of gravity cross, ordinate
Levy vector.Wavelet decomposition is a kind of multiple dimensioned multiresolution analysis method, is very suitable for analyzing unstable signal.
Wherein, a, b are respectively scale factor and shift factor, and f (t) is original signal,For morther wavelet, * is indicated altogether
Yoke,T fruit is tied for the f of wavelet transformation, t is the time.Original signal is decomposed by wavelet decomposition
Approximation coefficient and detail coefficients two parts reflect the general trend and high-frequency characteristic of signal respectively, are respectively original in data volume
The half or so of beginning data can continue to decompose with pairing approximation component later.
Original signal and first and third, five, seven layer of wavelet decomposition detail coefficient are as shown in Figure 4.The column of the left side one are centers of gravity in Fig. 4
The abscissa x of coordinate, the column of the right one are the ordinate y of barycentric coodinates.The first row is original signal, the first half of original signal
It is the stabilization signal of the first kind, latter half is that the leg of the second class moves and twitch signal, from the details system of first layer wavelet decomposition
It can be seen that, the wavelet details coefficient of the stationary signal of the first kind is all very small in number, and the second class leg moves signal and occurs every time
It is a peak value that leg will react when dynamic in detail coefficients, therefore wavelet decomposition can extract the high-frequency characteristic of original signal
Out, there is specific physical significance.With increasing for Decomposition order, although peak value gradually obscures, between two class signals
Difference it is still fairly obvious.Therefore, use the detail coefficients of wavelet decomposition also can be very high as the classification accuracy of feature.
The detail coefficients of layer 7 are chosen as fisrt feature, the variance of the cross of barycentric coodinates, ordinate is constituted into 18 dimensions
Feature vector is as second feature, when there are peak values and barycentric coodinates that slope is greater than the first preset threshold in detail coefficients
When horizontal, ordinate variance is greater than the second preset threshold, it is believed that original signal is that leg moves signal；It is horizontal, ordinate to choose center of gravity
Detail coefficients (x, y respectively have 8 values) after seven layers of wavelet decomposition, the variance (2 values) of center of gravity cross, ordinate) constitute 18 dimensions
Feature vector carries out classification using the method for support vector machines and obtains the feature that leg moves signal, thus by ortho signal
It automatically identifies leg and moves the classification that signal carries out sleep behavior, and then significantly behavior mentions by sleeping disorders correlations such as the dynamic twitch of leg
It takes out；
The classification method of this step is using support vector machines (SVM).Data are grouped with being one group per minute,
To corresponding 18 dimensional feature vector of each group of calculating, 45 groups of data are obtained using MQTT server, 20 groups of data therein are made
For training data, 25 groups of data obtain a SVM model as test data, using training data training, to test data into
Row classification.As a result there was only one group of data classification mistake, accuracy 96%.
Breath signal and the isolated breast rail algorithm flow chart that leg moves signal are as shown in Figure 3.Because leg moves signal
Amplitude is much larger than breath signal, and direct treatment effect is very poor, needs to separate.It is that data are carried out first by the method that the two separates
Adding window, the data that window length is 21, about 2 seconds calculate the difference of data maximums and minimum value in window, and set breath signal threshold
Value H thinks that the central point of this group of data belongs to leg and moves signal when difference is greater than breath signal threshold value H, when difference is less than breathing
Think that the central point of this group of data belongs to breath signal when signal threshold value H.It finds after all legs move data and directly counts these
According to discarding, the leg of early stage every night moved number less than hundred times, therefore directly abandoned at two seconds or so each leg dynamic duration
It influences little.The sleep detection algorithm of this system was done using one minute data classifies, and can only judge to have occurred in a certain minute
Leg moves phenomenon, if in conjunction with this upper breast rail algorithm can be accurate to a certain second taken place leg move phenomenon.Separation
Effect is as shown in fig. 6, point thing discrete in Fig. 6 originally belongs to the point that leg moves signal.
The breath signal isolated is filtered, 5 rank moving average filters are first passed through, if data length is n,
Preceding 4 values of data are successively mended in data finally, the serial data for constituting a n+4 takes later since first data
Current data and 4 data later calculate mean value, are assigned to current data.Moving average filtering is expressed such as with mathematical formulae
Under:
X [i] is original signal in formula, and y [i] is the signal after moving average filtering.
Moving average filtering removes a part of random noise and then carries out finite impulse response (FIR) low pass filtered
Wave extracts breath signal.FIR low pass filter is designed using window function metht, and sample frequency Fs is 500Hz, it is contemplated that breath signal
Frequency range in 0.2~0.8Hz, therefore select cutoff frequency Fc for 0.8Hz, window type is Caesar (Kaiser) window,
Beta is 2.5, filter order 30, and the filter amplitude-frequency response designed is as shown in Figure 7.
Filtered breath signal as shown in figure 5, the first row is original signal, the second row be by moving average filtering it
Signal afterwards, the third line are using the signal after FIR low-pass filtering, and dotted line is the apnea phenomenon detected, circle
What is sectioned out is the breathing detected.
4) statistical analysis of the detection of apnea phenomenon and respiratory rate:
The detection method of apnea phenomenon is similar with breathing statistic algorithm, since the duration of apnea phenomenon is big
It is approximately 2 seconds, therefore still selects length for 21 window, calculates the difference of the maximum value of 21 data points and minimum value in window, if
A fixed apnea threshold value D, thinks that the central point of this group of data belongs to breath signal when difference is greater than apnea threshold value,
Think that it belongs to apnea phenomenon when difference is less than apnea threshold value D.Likewise, signal has carried out at normalization first
Reason, so this apnea threshold value D can be applied to the detection of different crowd.
The method that the statistics of respiratory rate uses statistical signal peak value in this system calculates the signal after filtering oblique
, there are a minor peaks when slope is by just becoming negative in rate, and total peak value number is the number breathed.
5) result of MQTT server data processing can be sent to the displaying that processing result is carried out in a APP, sleep
The respiration rate in per minute is delineated in dormancy respiratory rate histogram, as shown in figure 8, user is facilitated to check oneself sleep period
Between respiratory rate situation of change.
Embodiment two: microprocessor 4 sends original signal by bluetooth module 6 and is handled to mobile phone 8
The present embodiment acquires data using the pressure sensor 2 for being placed on four 1 underfooting of bed, and the pressure that this system uses passes
The model YZC-167 of sensor 2,75 kilograms of range, total range of four sensors 2 is 300 kilograms.YZC-167 is a answers
Tested pressure conversion can be the variation of respective resistivity values, and then be converted to the change of output voltage values by variant pressure sensor
Change.A/D conversion module 3 selects HX711A/D conversion module, does amplification to this voltage value and A/D converts and can be obtained pressure
Data.Here the pressure reading obtained is not accurate numerical value, is calibrated.
1) it will do it primary reset under empty bed state, record reading at this time as fur weight, read every time later
Number Shi Douhui subtracts this fur weight, and the reading that processing obtains so is just entirely the nt wt net weight as caused by human body, avoids
The error due to caused by the extraneous factors such as placement position of article on bed.
It is the analysis based on center of gravity that empty bed 1, which resets the method used, it is therefore desirable to turn the reading of four pressure sensors 2
It is changed to the reading of center of gravity, using a foot of bed 1 as coordinate origin, two beds 1 are respectively that X-axis forward direction and Y-axis forward direction are established along direction
Rectangular coordinate system, four pressure sensors 2 are placed in four 1 underfooting of bed.According to the pressure sensor 2 of four 1 underfooting of bed
Reading can calculate barycentric coodinates, respectively remember center of gravity cross, ordinate x, y, according to lever principle:
Wherein, w1, w2, w3, w4 are respectively the reading of four cabinet base lower pressure sensors.
Activity can cause the reading of pressure sensor 2 to change to user in bed, the reading that pressure sensor 2 is read every time
Number subtracts the reading of fur weight under initial empty bed state as original signal to be processed；
2) original signal of pressure sensor 2 by A/D conversion module 3 conversion after by microprocessor 4 to original signal into
Row judgement:
In view of at present on the market many sleep detection equipment cannot judge automatically user's beginning and end sleep when
Between, this system devises a kind of simple judgment method, and STM32 microprocessor can be selected in microprocessor 4, in STM32 micro process
After device gets original signal, judged first, if original signal twice in succession is incited somebody to action all greater than index threshold T
Flag bit S sets 1, it is believed that user starts to sleep；If original signal twice in succession, will mark all less than index threshold T
Position S sets 0, it is believed that user has got up.In view of the weight and actual test of normal human, index threshold T is arranged
It is 2 kilograms.
The sample frequency of this system is 10Hz, and leg moves the duration of signal at 2 seconds or so, and the frequency of breath signal exists
0.2 arrives 0.8Hz or so, so the sample frequency of 10Hz is fully able to meet system requirements.
3) in view of the environment of no Wifi covering, microprocessor 4 is by the reading of four pressure sensors 2 in the present embodiment
W1, w2, w3, w4 and flag bit S are sent to bluetooth module 6 by serial ports, then transmit data to mobile phone 8 by bluetooth module 6,
The processing to original signal is realized in mobile terminal.
This system by bluetooth module 6 send be still five data-signals constitute Json formatted data, while
The accuracy of signal transmission is further enhanced before every group of data plus a start bit.By testing for a long time, with this
Kind data format carries out signal transmission accuracy rate can achieve 100% substantially.It is mobile-terminated receive original signal after, carry out special
Sign is extracted to obtain sleep signal and breath signal.
As shown in Fig. 2, the extraction of sleep signal is based on sleep behavior detection algorithm, respectively from time domain and the angle of frequency domain from
Extract following characteristics in original signal, sleep signal be divided into ortho signal and leg and moves signal, its purpose is to from
The part that may represent that leg is dynamic and twitches is distinguished in original signal, sleep behavior is divided into following two categories: the first kind is just
Stable state often is kept when sleep, is generated without any Large Amplitude Motion；Second class includes swing and whole body or the office of leg
The twitch in portion (such as leg).
Sleep behavior detection algorithm specifically: carry out multilayer using barycentric coodinates of the wavelet decomposition to the original signal of acquisition
Approximation coefficient and detail coefficients after decomposing the cross for obtaining barycentric coodinates, ordinate wavelet decomposition, cross, the ordinate of barycentric coodinates
Detail coefficients (x, y respectively have 8 values) after seven layers of wavelet decomposition of (x, y), 18 Wei Te of variance composition of center of gravity cross, ordinate
Levy vector.Wavelet decomposition formula calculates are as follows:
Wherein, a, b are respectively scale factor and shift factor, and f (t) is original signal,For morther wavelet, * is indicated altogether
Yoke,T fruit is tied for the f of wavelet transformation, t is the time.Original signal is decomposed by wavelet decomposition
Approximation coefficient and detail coefficients two parts reflect the general trend and high-frequency characteristic of signal respectively, are respectively original in data volume
The half or so of beginning data can continue to decompose with pairing approximation component later.
Original signal and first and third, five, seven layer of wavelet decomposition detail coefficient are as shown in Figure 4.The column of the left side one are centers of gravity in Fig. 4
The abscissa x of coordinate, the column of the right one are the ordinate y of barycentric coodinates.The first row is original signal, the first half of original signal
It is the stabilization signal of the first kind, latter half is that the leg of the second class moves and twitch signal, from the details system of first layer wavelet decomposition
It can be seen that, the wavelet details coefficient of the stationary signal of the first kind is all very small in number, and the second class leg moves signal and occurs every time
It is a peak value that leg will react when dynamic in detail coefficients, therefore wavelet decomposition can extract the high-frequency characteristic of original signal
Out, there is specific physical significance.
The detail coefficients of layer 7 are chosen as fisrt feature, the variance of the cross of barycentric coodinates, ordinate is constituted into 18 dimensions
Feature vector is as second feature, when there are peak values and barycentric coodinates that slope is greater than the first preset threshold in detail coefficients
When horizontal, ordinate variance is greater than the second preset threshold, it is believed that original signal is that leg moves signal；It is horizontal, ordinate to choose center of gravity
Detail coefficients (x, y respectively have 8 values) after seven layers of wavelet decomposition, the variance (2 values) of center of gravity cross, ordinate) constitute 18 dimensions
Feature vector carries out classification using the method for support vector machines and obtains the feature that leg moves signal, thus by ortho signal
It automatically identifies leg and moves the classification that signal carries out sleep behavior, and then significantly behavior mentions by sleeping disorders correlations such as the dynamic twitch of leg
It takes out；
The classification method of this step is using support vector machines (SVM).Data are grouped with being one group per minute,
To corresponding 18 dimensional feature vector of each group of calculating, 30 groups of data are obtained using bluetooth 6 and mobile phone 8, by 14 groups of data therein
As training data, 16 groups of data obtain a SVM model as test data, using training data training, to test data
Classify.As a result there was only one group of data classification mistake, accuracy 94%.By these data and pass through MQTT server process
Obtained data are mixed, and randomly select a portion training classifier, with remaining data test, final accuracy rate is complete
Portion has reached 90% or more, can reach 100%.
Breath signal and the isolated breast rail algorithm flow chart that leg moves signal are as shown in Figure 3.Because leg moves signal
Amplitude is much larger than breath signal, and direct treatment effect is very poor, needs to separate.It is that data are carried out first by the method that the two separates
Adding window, the data that window length is 21, about 2 seconds calculate the difference of data maximums and minimum value in window, and set a breathing letter
Number threshold value H thinks that the central point of this group of data belongs to leg and moves signal when difference is greater than breath signal threshold value H, when difference is less than
Think that the central point of this group of data belongs to breath signal when breath signal threshold value H.It finds all legs and moves data later directly by this
A little data abandon, and the leg of early stage every night moved number and lost less than hundred times, therefore directly at two seconds or so each leg dynamic duration
Abandoning also influences less.The sleep detection algorithm of this system was done using one minute data classifies, and can only judge hair in a certain minute
Given birth to leg move phenomenon, if in conjunction with this upper breast rail algorithm can be accurate to a certain second taken place leg move phenomenon.Point
From effect as shown in fig. 6, point thing discrete in Fig. 6 originally belong to leg move signal point.
The breath signal isolated is filtered.First pass through moving average filtering to remove random disturbances, here
Using 5 rank moving average filters, if data length is n, preceding 4 values of data are successively mended in data finally, constituting one
The serial data of a n+4 takes current data and 4 data later to calculate mean value later since first data, is assigned to current
Data.
Moving average filtering removes a part of random noise and then carries out finite impulse response (FIR) low pass filtered
Wave extracts breath signal.FIR low pass filter is designed using window function metht, and sample frequency Fs is 500Hz, it is contemplated that breath signal
Frequency range in 0.2~0.8Hz, therefore select cutoff frequency Fc for 0.8Hz, window type is Caesar (Kaiser) window,
Beta is 2.5, filter order 30, and the filter amplitude-frequency response designed is as shown in Figure 7.
Filtered breath signal as shown in figure 5, the first row is original signal, the second row be by moving average filtering it
Signal afterwards, the third line are using the signal after FIR low-pass filtering, and dotted line is the apnea phenomenon detected, circle
What is sectioned out is the breathing detected.
4) statistical analysis of the detection of apnea phenomenon and respiratory rate:
The detection method of apnea phenomenon is similar with breathing statistic algorithm, since the duration of apnea phenomenon is big
It is approximately 2 seconds, therefore still selects length for 21 window, calculates the difference of the maximum value of 21 data points and minimum value in window, if
A fixed apnea threshold value D, thinks that the central point of this group of data belongs to breath signal when difference is greater than apnea threshold value,
Think that it belongs to apnea phenomenon when difference is less than apnea threshold value D.Likewise, signal has carried out at normalization first
Reason, so this apnea threshold value H can be applied to the detection of different crowd.
The method that the statistics of respiratory rate uses statistical signal peak value in this system calculates the signal after filtering oblique
, there are a minor peaks when slope is by just becoming negative in rate, and total peak value number is the number breathed.
5) statistical result can be shown on a app, be delineated in sleep-respiratory frequency histogram per minute
Interior respiration rate, as shown in figure 8, facilitate user check oneself sleep during respiratory rate situation of change.
Claims (7)
1. a kind of sleep quality and respiration monitoring system based on bed body, including bed body (1), it is characterised in that: bed body (1) four
The lower end of cabinet base is respectively mounted pressure sensor (2), and four pressure sensors (2) are commonly connected to A/D conversion module by conducting wire
(3), A/D conversion module (3) is connected to wireless module through microprocessor (4), and wireless module includes WiFi module (5) and bluetooth mould
Block (6), WiFi module (5) and bluetooth module (6) obtain original signal, WiFi module (5) from microprocessor (4) by serial ports
It sends original signal to network data base (7), bluetooth module (6) sends mobile phone (8) for original signal by Bluetooth protocol.
2. a kind of sleep quality and respiration monitoring system based on bed body according to claim 1, it is characterised in that: described
Pressure sensor (2) be fixed under cabinet base, collected pressure signal is converted to resistance value by pressure sensor (2), in turn
Be converted to output voltage values.
3. a kind of sleep quality and respiration monitoring system based on bed body according to claim 1, it is characterised in that: described
A/D conversion module (3) include regulated power supply, on-chip clock oscillator, for by the output voltage values of pressure sensor (2) into
Row amplification and A/D conversion.
4. a kind of sleep quality and respiration monitoring system based on bed body according to claim 1, it is characterised in that: described
Microprocessor (4) include it is multiple rapidly input/export (I/O) mouth, for reading, through A/D conversion module (3), that treated is former
Beginning signal is simultaneously judged.
5. a kind of sleep quality and respiration monitoring system based on bed body according to claim 1, it is characterised in that: sleep
Behavioral value algorithm carries out multilayer decomposition to the barycentric coodinates of original signal using wavelet decomposition, extracts wavelet details coefficient and again
The variance of heart coordinate is classified as judging characteristic using the method for support vector machines, so that sleep signal is divided into normally
Sleep signal and leg move signal.
6. a kind of sleep quality and respiration monitoring system based on bed body according to claim 1, it is characterised in that: breathing
Statistic algorithm carries out windowing process to original signal, is extracted breath signal from original signal according to the threshold value of setting,
Filtered breath signal is obtained by filtering again, and counts respiration rate and detection apnea, realizes the more of sleep procedure
Kind unaware detection.
7. being applied to a kind of monitoring side of any sleep quality and respiration monitoring system based on bed body of claim 1-6
Method, it is characterised in that method includes the following steps:
1) pressure sensor (2) is resetted under empty bed state:
The reading of empty bed state lower pressure sensor (2) is recorded as fur weight, activity causes pressure to pass to user in bed
The reading of sensor (2) changes, and the reading that pressure sensor (2) changes every time is subtracted the reading of fur weight as original letter
Number；
Using a foot of cabinet base as coordinate origin, the two cabinet base directions adjacent using its are established as X-axis forward direction and Y-axis forward direction
Rectangular coordinate system calculates barycentric coodinates according to the original signal of the pressure sensor (2) under four cabinet bases, barycentric coodinates
Horizontal, ordinate (x, y) can be obtained according to lever principle:
Wherein, w1, w2, w3, w4 are respectively the original signal value of the pressure sensor (2) under four cabinet bases；
2) original signal of pressure sensor (2) is judged after A/D conversion module (3) conversion by microprocessor (4): being set
Flag bit S is set, if flag bit S is set 1, it is believed that make all greater than index threshold T by the original signal twice in succession received
User starts to sleep；If original signal twice in succession sets 0 all less than index threshold T, by flag bit S, it is believed that user
It has got up；
3) network data base (7) or hand are sent by WiFi module (5) or bluetooth module (6) by original signal and flag bit S
Machine (8), network data base (7) or mobile phone (8) carry out feature extraction to original signal to obtain sleep signal and breath signal；
The extraction of sleep signal is based on sleep behavior detection algorithm and sleep signal is divided into ortho signal and the dynamic signal of leg, tool
Body are as follows: carry out multilayer using barycentric coodinates of the wavelet decomposition to the original signal of acquisition and decompose to obtain horizontal, the vertical seat of barycentric coodinates
Approximation coefficient and detail coefficients after marking wavelet decomposition, choose the detail coefficients of layer 7 as fisrt feature, by barycentric coodinates
Cross, ordinate variance constitutive characteristic vector as second feature, when there are slopes to be greater than the first default threshold in detail coefficients
When the peak value and the cross of barycentric coodinates of value, the variance of ordinate are greater than the second preset threshold, it is believed that original signal is that leg moves signal；
Classification is carried out using the method for support vector machines and obtains the feature that leg moves signal, thus by being automatically identified in ortho signal
Leg moves signal；
The separation that breath signal and leg move signal uses breast rail calculation method, specifically: breath signal threshold value H is set, it is right
Original signal carries out windowing process, calculates the difference of the maxima and minima of the original signal in each window, and difference, which is greater than, exhales
Think that the original signal in window belongs to leg and moves signal when inhaling signal threshold value H, thinks in window when difference is less than breath signal threshold value H
Original signal belong to breath signal；It abandons all legs that belong in original signal and moves the data of signal, and then isolated
Breath signal after separation is removed random disturbances by moving average filtering by breath signal, then by FIR low-pass filtering it
Signal afterwards obtains filtered breath signal；
4) statistical analysis of the detection of apnea phenomenon and respiratory rate:
Set apnea threshold value D, windowing process carried out to original signal, calculate the maximum value of the original signal in each window with
The difference of minimum value, difference thinks that the original signal in window belongs to breath signal when being greater than apnea threshold value D, when difference is less than
Think that the original signal in window belongs to apnea phenomenon when apnea threshold value D；
The method that the statistics of respiratory rate uses statistical signal peak value calculates the filtered breath signal described in step 3) oblique
Rate, is counted as a minor peaks when slope is by just becoming negative, and the sum of peak value is the number breathed；
5) statistical result of the respiratory rate obtained according to step 4) is drawn out in sleep-respiratory frequency histogram in per minute
Respiration rate, the processing result to original signal of network data base (7) or mobile phone (8) sends out as sleep behavior testing result
It is sent to mobile terminal, user checks the situation of change of respiratory rate during sleep by mobile terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811243454.6A CN109528159B (en) | 2018-10-24 | 2018-10-24 | Human sleep and respiration monitoring system and method based on bed body |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811243454.6A CN109528159B (en) | 2018-10-24 | 2018-10-24 | Human sleep and respiration monitoring system and method based on bed body |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109528159A true CN109528159A (en) | 2019-03-29 |
| CN109528159B CN109528159B (en) | 2020-08-14 |
Family
ID=65844967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811243454.6A Active CN109528159B (en) | 2018-10-24 | 2018-10-24 | Human sleep and respiration monitoring system and method based on bed body |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109528159B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110575139A (en) * | 2019-08-13 | 2019-12-17 | 珠海市万瑙特健康科技有限公司 | Sleep monitoring method and equipment |
| CN111227792A (en) * | 2020-01-10 | 2020-06-05 | 京东方科技集团股份有限公司 | Apnea detection method and system, electronic device and storage medium |
| CN111310599A (en) * | 2020-01-20 | 2020-06-19 | 重庆大学 | Sleep action recognition system capable of quickly adapting to various change factors |
| CN112545851A (en) * | 2020-11-03 | 2021-03-26 | 未来穿戴技术有限公司 | Massage method and device, electronic device and computer readable storage medium |
| CN112617821A (en) * | 2020-12-30 | 2021-04-09 | 山东沃尔德生物技术有限公司 | System and method for monitoring getting on and off of bed of old people |
| TWI773562B (en) * | 2021-10-14 | 2022-08-01 | 雲云科技股份有限公司 | Breath detecting system and breath detecting mat thereof |
| CN117409537A (en) * | 2023-11-20 | 2024-01-16 | 河北大学附属医院 | Ward sensing and calling system for night use |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102078184A (en) * | 2010-12-31 | 2011-06-01 | 深圳清华大学研究院 | Method for monitoring body position by using weighing bed and weighing bed |
| US20170143269A1 (en) * | 2006-09-22 | 2017-05-25 | Sleepiq Labs Inc. | Systems and methods for monitoring a subject at rest |
| CN107041730A (en) * | 2017-03-07 | 2017-08-15 | 杭州医电园科技有限公司 | A kind of contactless sleep monitoring device and its monitoring method |
| CN108135486A (en) * | 2015-08-17 | 2018-06-08 | 瑞思迈传感器技术有限公司 | Sleep disordered breathing disease screening instrument |
| CN108289639A (en) * | 2015-09-29 | 2018-07-17 | 美蓓亚三美株式会社 | Biometric information monitors system |
| CN108697241A (en) * | 2015-12-30 | 2018-10-23 | 德沃特奥金有限公司 | Sleep with sensor or dependence furniture |
-
2018
- 2018-10-24 CN CN201811243454.6A patent/CN109528159B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170143269A1 (en) * | 2006-09-22 | 2017-05-25 | Sleepiq Labs Inc. | Systems and methods for monitoring a subject at rest |
| CN102078184A (en) * | 2010-12-31 | 2011-06-01 | 深圳清华大学研究院 | Method for monitoring body position by using weighing bed and weighing bed |
| CN108135486A (en) * | 2015-08-17 | 2018-06-08 | 瑞思迈传感器技术有限公司 | Sleep disordered breathing disease screening instrument |
| CN108289639A (en) * | 2015-09-29 | 2018-07-17 | 美蓓亚三美株式会社 | Biometric information monitors system |
| CN108697241A (en) * | 2015-12-30 | 2018-10-23 | 德沃特奥金有限公司 | Sleep with sensor or dependence furniture |
| CN107041730A (en) * | 2017-03-07 | 2017-08-15 | 杭州医电园科技有限公司 | A kind of contactless sleep monitoring device and its monitoring method |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110575139A (en) * | 2019-08-13 | 2019-12-17 | 珠海市万瑙特健康科技有限公司 | Sleep monitoring method and equipment |
| CN111227792A (en) * | 2020-01-10 | 2020-06-05 | 京东方科技集团股份有限公司 | Apnea detection method and system, electronic device and storage medium |
| CN111227792B (en) * | 2020-01-10 | 2022-11-01 | 京东方科技集团股份有限公司 | Apnea detection method and system, electronic device and storage medium |
| CN111310599A (en) * | 2020-01-20 | 2020-06-19 | 重庆大学 | Sleep action recognition system capable of quickly adapting to various change factors |
| CN112545851A (en) * | 2020-11-03 | 2021-03-26 | 未来穿戴技术有限公司 | Massage method and device, electronic device and computer readable storage medium |
| CN112617821A (en) * | 2020-12-30 | 2021-04-09 | 山东沃尔德生物技术有限公司 | System and method for monitoring getting on and off of bed of old people |
| TWI773562B (en) * | 2021-10-14 | 2022-08-01 | 雲云科技股份有限公司 | Breath detecting system and breath detecting mat thereof |
| CN117409537A (en) * | 2023-11-20 | 2024-01-16 | 河北大学附属医院 | Ward sensing and calling system for night use |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109528159B (en) | 2020-08-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109528159A (en) | A kind of sleep quality and respiration monitoring system and method based on bed body | |
| CN108042108B (en) | Sleep quality monitoring method and system based on body vibration signals | |
| CN106937808B (en) | Data acquisition system of intelligent mattress | |
| CN105118236B (en) | Paralysis falls to monitor and preventing mean and its processing method | |
| CN104545844B (en) | Multi-parameter sleep monitoring and intelligent diagnosis system based on 4G mobile communication technology and application method of multi-parameter sleep monitoring and intelligent diagnosis system | |
| WO2017056476A1 (en) | Biometric information monitoring system | |
| CN108670200A (en) | A kind of sleep sound of snoring classification and Detection method and system based on deep learning | |
| EP3329845A1 (en) | Biological condition determination device and biological condition determination method | |
| CN106419845B (en) | A kind of sleep monitoring device and method based on piezoceramic transducer | |
| CN110432908A (en) | Perception and activity classification to baby | |
| CN103845038B (en) | A kind of sign signal acquisition method and apparatus | |
| CN105852864A (en) | Wearable respiratory monitoring system and monitoring method thereof | |
| CN104473648A (en) | Physiological parameter monitoring-combined human body tumble warning and detecting method | |
| CN107887032A (en) | A kind of data processing method and device | |
| CN113688720B (en) | Method for predicting sleeping gesture based on neural network recognition | |
| CN106344034A (en) | Sleep quality evaluation system and method | |
| CN110811647A (en) | Multi-channel hidden lie detection method based on ballistocardiogram signal | |
| CN109044275B (en) | Non-invasive sensing sleep quality analysis system and method based on fuzzy inference system | |
| CN111481207A (en) | Sleep posture recognition device and method based on cardiac shock signal | |
| CN206063138U (en) | The healthy critical warning cloud monitoring system of naturalness night burst | |
| CN110693510A (en) | Attention deficit hyperactivity disorder auxiliary diagnosis device and using method thereof | |
| CN104274158A (en) | Furniture physical sign detection device and furniture type physical sign monitoring system | |
| CN206080518U (en) | Wearable breathes monitor system | |
| CN210990210U (en) | Non-inductive sleep monitoring device and monitoring system | |
| Jeng et al. | Stream data analysis of body sensors for sleep posture monitoring: An automatic labelling approach |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |