CN113198067A

CN113198067A - Automatic medical care monitoring system

Info

Publication number: CN113198067A
Application number: CN202110628881.1A
Authority: CN
Inventors: 郭昆仑; 金晖; 何洁
Original assignee: Hangzhou City University
Current assignee: Hangzhou City University
Priority date: 2021-06-03
Filing date: 2021-06-03
Publication date: 2021-08-03

Abstract

The invention relates to an automatic medical care monitoring system, which comprises the following steps: the disease state monitoring module collects face signals by using a signal collecting device and transmits the collected face signals to the raspiberry pi4b for disease state monitoring; the transfusion monitoring module detects the transfusion dripping speed and displays the transfusion dripping speed on the display module by using the resistance voltage change corresponding to the signal change received by the infrared geminate transistors. The invention has the beneficial effects that: the invention designs an automatic medical care monitoring system which has the functions of dropping speed detection, dropping speed display, automatic alarm, transfusion stopping, body temperature monitoring, blood pressure monitoring, disease state monitoring, wireless transmission and the like. The automatic medical care monitoring system takes a raspiberry pi4b as a core, utilizes infrared geminate transistors to detect liquid drop dripping, then transmits data to a singlechip through a circuit to realize signal processing and control, and finally realizes the alarm function of threshold crossing of the dripping speed and blood backflow and the function of stopping transfusion.

Description

Automatic medical care monitoring system

Technical Field

The invention belongs to the field of medical care, and particularly relates to an automatic medical care monitoring system.

Background

Automated medical monitoring systems have developed rapidly in recent years due to the rapid development of artificial intelligence. An automatic medical monitoring system refers to a system that can replace medical staff to monitor the medical care of a patient. The common medical care behaviors include transfusion, temperature measurement, heart rate measurement and the like, and all belong to simple and high-repeatability medical care behaviors; for many years, in order to better monitor patients at home and abroad, a great number of medical care systems are designed, such as automatic heart rate measuring machines, electronic thermometers, intelligent droppers and the like. However, generally speaking, the function is single, the innovation is not strong, and a large-scale medical care system such as an intelligent surgical robot is high in cost and is not suitable for the simple medical care behaviors. In the new coronary pneumonia epidemic situation of 2020, the defect is amplified, on one hand, the number of patients is large due to strong pneumonia spreading capacity, on the other hand, medical staff is limited, most of patients are light patients, and simple but labor-consuming medical actions such as isolation, temperature measurement and heart rate measurement at regular time, infusion on time and the like are needed. Therefore, there is a need for a medical care system that can meet the above needs and at the same time has a certain degree of intelligence.

Disclosure of Invention

The invention aims to overcome the defects in the prior art and provide an automatic medical care monitoring system.

The automated healthcare monitoring system comprising: the infusion monitoring device comprises a main module, a disease state monitoring module, an infusion monitoring module, a body temperature monitoring module, a heart rate monitoring module and a wireless transmission module; the condition monitoring module, the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are connected with the main module through the wireless transmission module in a Bluetooth mode;

wherein the disease state monitoring module is composed of a raspiberry pi4b, and a signal acquisition device is arranged on the disease state monitoring module;

the infusion monitoring module is provided with a main control module, an audible and visual alarm module, a display module, a key module, a blood return alarm module, a dripping speed detection module, a Bluetooth module, a motor module and a power socket switch module; the sound and light alarm module, the display module, the key module, the blood return alarm module, the dripping speed detection module and the Bluetooth module are all electrically connected with the main control module; the main control module adjusts the dropping speed of the infusion tube through the motor module; a power socket switch module is connected between the main control module and the power supply; the dripping speed detection module is provided with an infrared pair tube (an infrared transmitting tube and an infrared receiving tube) and a voltage comparator; the blood return alarm module is also provided with an infrared pair tube (an infrared transmitting tube and an infrared receiving tube) and a voltage comparator; a stepping motor is arranged on the motor module;

the body temperature monitoring module is provided with an infrared sensing chip and a signal processing chip;

be equipped with integrated pulse blood oxygen and heart rate on the heart rate monitoring module and detect integrative sensor, the optical signal processor and the analog signal processor of optimization, integrated pulse blood oxygen and heart rate detect integrative sensor and include two LED lamps and a high accuracy photoelectric sensor.

Preferably, the main module is a mobile phone.

Preferably, the wireless transmission module is a serial port Bluetooth module or a raspberry self-contained Bluetooth module; the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are in Bluetooth connection with the main module through the serial port Bluetooth module; the disease state monitoring module is connected with the main module through a Bluetooth module carried by the raspberry, so that data such as dripping speed, body temperature and heart rate can be transmitted to equipment such as a mobile phone or a computer and the like which can be connected with Bluetooth through Bluetooth.

Preferably, the main control module on the infusion monitoring module is an STC89C52RC single-chip microcomputer, the Bluetooth module is an HC-05 Bluetooth module, the display module is a 1602 liquid crystal display, the motor module is ULN2003, and the voltage comparator is an LM 393; the stepper motor is 28BYJ 48.

Preferably, the infrared sensing chip on the body temperature monitoring module is MLX 90614.

Preferably, the sensor integrating pulse blood oxygen and heart rate detection on the heart rate monitoring module is Max 30100.

The working method of the automatic medical care monitoring system comprises the following steps:

step 1, a disease state monitoring module collects face signals by using a signal collecting device and transmits the collected face signals to a raspiberry pi4b for disease state monitoring;

step 2, the transfusion monitoring module detects the transfusion dripping speed and displays the detected transfusion dripping speed on the display module to judge whether blood returns during transfusion of a patient, and when blood flows back, the infrared receiving tube receives a corresponding signal and alarms through the sound-light alarm module; the motor module is used for driving the transfusion circuit and regulating and controlling the dripping speed: when the infusion speed exceeds a set threshold value, the motor is started, then the infusion tube is clamped, and the dropping speed is reduced; when the infusion speed is lower than a set threshold value, the motor is driven reversely to accelerate the dropping speed;

step 3, the body temperature monitoring module measures the temperature of the target by using an infrared sensing chip and then calculates the body temperature by using a built-in signal processing chip;

step 4, the heart rate monitoring module adopts an optical volume method, heart rate signals are collected through a sensor integrating pulse blood oxygen and heart rate detection, an internal low-noise analog signal processor performs ADC (analog-to-digital converter) processing on the heart rate signals, then the heart rate signals enter an array filter, and finally the heart rate signals are output through an IIC (inter integrated circuit) bus; the heart rate is calculated from the change in the amount of light reflected from the cells to the sensor.

Preferably, the signal acquisition device in the condition monitoring module in step 1 is pi camera; in the step 2, a DC power socket is adopted as a power socket switch module in the transfusion monitoring module, and the on-off of the power supply is controlled by a key.

Preferably, step 1 specifically comprises the following steps:

step 1.1, the state of illness monitoring module carries out face detection: calling a special human face characteristic point detector (shape _ predictor _68_ face _ landmark of a dlib computer deep learning open source library) of a dlib, detecting eye, mouth, eyebrow and nose areas in a human face, returning corresponding coordinates, calculating Euclidean geometric distances among the eye, mouth, eyebrow and nose areas in the human face by combining the coordinates, then collecting the changes of the Euclidean geometric distances among the eye, mouth, eyebrow and nose areas when the state of the human face changes, comparing the changes of the Euclidean geometric distances with a set judgment threshold value, and detecting and recognizing facial areas and expressions; if the change of the Euclidean geometric distance reaches a set judgment threshold value, the facial area in the face has an expression corresponding to the judgment threshold value;

step 1.2, the state of illness monitoring module carries on the fatigue detection; the fatigue detection contents comprise blink detection, mouth opening and closing detection, eyebrow picking degree detection and nose spacing detection;

step 1.2.1, blink detection: collecting the change of Euclidean geometric distances between 12 characteristic points on human eyes when the state of a human face changes, wherein each eye has 6 characteristic points; calculating the human eye opening and closing threshold value:

in the above formula, EAR is eye aspect ratio, and p1 to p6 respectively represent 6 eye feature points clockwise from the left; two modulo sum p on the molecule₂-p₆||+||p₃-p₅The | is the distance of the characteristic point of the human eye in the vertical direction, and the modulo calculation on the denominator is | p₁-p₄| is the distance of the characteristic point of the human eye in the horizontal direction, and | p₁-p₄Matching the number of terms of the numerator by multiplying 2 to ensure that the weight of the numerator is the same as that of the denominator;

comparing the EAR obtained by calculation with an EAR reference value, and if the EAR obtained by calculation is far smaller than the EAR reference value, judging that blinking occurs;

step 1.2.2, mouth opening and closing detection: acquiring the change of Euclidean geometric distances between 4 mark points of upper and lower lips and 2 feature points of left and right mouth corners when the state of a human face changes, and calculating a mouth opening and closing threshold value:

the above formula is to use two norms to calculate the threshold of the feature points of the mouth, specifically to calculate the square sum and root number of the distance between the corresponding feature points;

if the calculated mouth opening and closing threshold is larger than the normal mouth threshold, the mouth is judged to be opened, and if the calculated mouth opening and closing threshold is smaller than the normal mouth threshold, the mouth is judged to be closed;

step 1.2.3, eyebrow picking degree detection: performing linear fitting on ten feature points on the eyebrows by using a polyfit function in numpy, wherein the left and right eyebrows are respectively provided with five feature points; after a linear function is fitted, judging the eyebrow picking degree through the slope; establishing two lists of line _ brow _ x and line _ brow _ y for storing the x and y position information of the eyebrows, setting a variable brow _ sum for storing the sum of the heights of the feature point coordinates of the eyebrows, and setting a variable frown _ sum for storing the sum of the distances of the eyebrows on two sides; taking the line on the rectangular frame as a horizontal axis, and summing the height sum of the coordinates of the eyebrow feature points as the sum of the difference between each coordinate and the vertical coordinate of the coordinate at the upper left corner of the rectangular frame; the sum of the eyebrow distances at the two sides is the sum of the differences between each coordinate and the abscissa of the coordinate at the upper left corner of the rectangular frame; the lists line _ break _ x and line _ break _ y are grouped by an array function in numpy, the grouped lists line _ break _ x and line _ break _ y are subjected to least square polynomial fitting by an np.

In the above formula, | p (x)_j)-y_jL represents the square of the difference between the absolute values of the coefficients minimizing the square error between the feature point and the fixed point; wherein p (x)_j) Coefficient representing the minimum squared error of the feature point, y_jCoefficients representing a fixed-point minimized square error; judging the proportion of the eyebrow length and the eyebrow height in the face frame by detecting the length and the height of the eyebrow feature points from the face frame, and analyzing the eyebrow picking degree in combination; if the calculated final slope fitting value of the eyebrows is lower than a critical threshold value, eyebrow selection occurs;

step 1.2.4, counting the relationship between the change degree of the nose bridge length and the nose width of the face and the change of the facial expression when the facial expression of the face changes;

and step 1.3, the expression recognition is carried out by the condition monitoring module according to the step 1.2.

Preferably, step 3 specifically comprises the following steps:

step 3.1, calculating the output temperature of the thermopile:

V_ir(Ta，To)＝A.(To⁴-Ta⁴)，

wherein To is the absolute temperature of the object To be measured, and the unit is Kelvin; ta is the absolute temperature of the environment, A is the sensitivity;

step 3.2, calculating corresponding ambient temperature Ta and object temperature To;

conversion of RAM content to ambient absolute temperature Ta:

Ta[K]＝Tareg×0.02，or 0.02K/LSB

in the above formula, Tareg is the temperature of the environment measured by the internal linear sensor;

the absolute temperature To of the object To be measured is:

To[K]＝Toreg×0.02，or 0.02K/LSB

in the above equation, Toreg is the temperature of the object measured by the internal linear sensor.

The invention has the beneficial effects that:

the invention designs an automatic medical care monitoring system which has the functions of dropping speed detection, dropping speed display, automatic alarm, transfusion stopping, body temperature monitoring, blood pressure monitoring, disease state monitoring, wireless transmission and the like. The automatic medical care monitoring system takes a raspiberry pi4b as a core, utilizes infrared geminate transistors to detect liquid drop dripping, then transmits data to a singlechip through a circuit to realize signal processing and control, and finally realizes the alarm function of threshold crossing of the dripping speed and blood backflow and the function of stopping transfusion.

The invention is based on the starting point of edge calculation and strong real-time and applicability, automatically deploys the library files required by face recognition on the embedded equipment, improves the existing algorithm, innovating the algorithm from a new idea, and can pertinently analyze the face state of the person in various occasions. The invention combines the wireless transmission technology, the embedded system technology and the computer programming technology, has strong comprehensiveness and good reconstruction and portability.

The invention can be widely applied to a plurality of occasions, is not limited by network, region and space, has extremely simple hardware equipment, can realize the monitoring of the illness state of the patient only by combining and transforming the raspberry pie with the size similar to a credit card and a camera, and has low hardware cost.

The body temperature and the heart rate are monitored by non-invasive infrared temperature measurement and photoelectric heart rate measurement, so that the practicability is high; meanwhile, the main control of the infusion module can be realized only by using an STM89C52RC single chip, the modules of body temperature, heart rate and the like only adopt Arduino nano chips, the disease state monitoring module adopts raspberry pi4b development, one module is used for one core, the cost is proper, and each module can be independently used and can be divided and combined.

Drawings

FIG. 1 is a block diagram of the module components of the present invention;

FIG. 2 is a partial circuit diagram of raspiberry pi4 b;

FIG. 3 is a schematic diagram of 68 person face feature point detections;

FIG. 4 is a graph of eye opening and closing threshold variation;

FIG. 5 is a graph of mouth opening and closing thresholds;

FIG. 6 is a graph of eyebrow curvature threshold for anger and nature;

FIG. 7 is a graph of nasal bridge length variation thresholds for both anger and nature;

FIG. 8 is a circuit connection diagram of the MLX90614 module;

FIG. 9 is a logical functional diagram of MLX 90614;

FIG. 10 is an MLX90614 SMBUS connection diagram;

FIG. 11 is a schematic diagram of the PWM output mode configuration of MLX 90614;

FIG. 12 is a PWM output mode configuration diagram of MLX 90614;

FIG. 13 is a circuit diagram of a MAX30100 module;

FIG. 14 is a basic schematic diagram of a photoplethysmography;

FIG. 15 is a logical functional diagram of Max 30100;

FIG. 16 is a circuit connection diagram of a main control module of the single chip microcomputer;

FIG. 17 is a circuit diagram of the audible and visual alarm module;

FIG. 18 is a circuit diagram of a liquid crystal display module 1602;

FIG. 19 is a circuit diagram of the key module;

FIG. 20 is a circuit diagram of a drop velocity detection module;

FIG. 21 is a circuit diagram of the blood return alarm module;

FIG. 22 is a circuit diagram of a Bluetooth transmission module;

FIG. 23 is a circuit connection diagram of ULN 2003;

fig. 24 is a circuit diagram of a power outlet switch.

Detailed Description

The present invention will be further described with reference to the following examples. The following examples are set forth merely to aid in the understanding of the invention. It should be noted that, for a person skilled in the art, several modifications can be made to the invention without departing from the principle of the invention, and these modifications and modifications also fall within the protection scope of the claims of the present invention.

Example 1:

as shown in fig. 1, an automated healthcare monitoring system includes: the infusion monitoring device comprises a mobile phone, a disease state monitoring module, an infusion monitoring module, a body temperature monitoring module, a heart rate monitoring module and a wireless transmission module; the condition monitoring module, the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are connected with the mobile phone Bluetooth through the wireless transmission module; the wireless transmission module is a serial port Bluetooth module or a raspberry self-carrying Bluetooth module; the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are in Bluetooth connection with the main module through the serial port Bluetooth module; the disease state monitoring module is connected with the main module through a Bluetooth module carried by the raspberry, so that data such as dripping speed, body temperature and heart rate can be transmitted to equipment such as a mobile phone or a computer and the like which can be connected with Bluetooth through Bluetooth. Each module can work independently to improve efficiency and can transmit data to a mobile phone terminal.

Wherein the disease state monitoring module is composed of a raspiberry pi4b shown in fig. 2, and a signal acquisition device is arranged on the disease state monitoring module;

the infusion monitoring module is provided with a main control module (STC89C52RC singlechip), an audible and visual alarm module, a display module, a key module, a blood return alarm module, a dropping speed detection module, a Bluetooth module (HC-05 Bluetooth module), a motor module (ULN2003) and a power socket switch module; the sound and light alarm module, the display module (1602 liquid crystal display), the key module, the blood return alarm module, the dripping speed detection module and the Bluetooth module are all electrically connected with the main control module; the main control module adjusts the dropping speed of the infusion tube through the motor module; a power socket switch module is connected between the main control module and the power supply; the dripping speed detection module is provided with an infrared pair tube (an infrared transmitting tube and an infrared receiving tube) and a voltage comparator (LM 393); the blood return alarm module is also provided with an infrared pair tube (an infrared transmitting tube and an infrared receiving tube) and a voltage comparator; a stepping motor (28BYJ48) is arranged on the motor module;

an infrared sensing chip (MLX90614) is arranged on the body temperature monitoring module, and a signal processing chip is also arranged in the body temperature monitoring module;

be equipped with integrated pulse blood oxygen and heart rate on the heart rate monitoring module and detect integrative sensor (Max30100), the optical signal processor and the analog signal processor of optimization, integrated pulse blood oxygen and heart rate detect integrative sensor and include two LED lamps and a high accuracy photoelectric sensor.

Example 2:

a working method of an automatic medical care monitoring system comprises the following implementation steps:

a condition monitoring module:

the disease condition monitoring module is composed of raspiberrypi4B, raspiberrypi4B adopts a boston BCM2711B0 as an SoC, the microcomputer mainboard is based on an ARM framework, the maximum running memory can reach 8GB at present, and an SD card is used as the memory. The USB interface device is provided with 4 USB ports, 2 MircoHDMI ports, a 3.5mmAV port and a CSI port, and supports gigabit Ethernet, dual-frequency wireless networks and Bluetooth. A partial circuit schematic diagram is shown in FIG. 2;

the algorithm of the disease condition monitoring module is roughly divided into three parts, namely a face detection algorithm, a fatigue monitoring algorithm and an expression recognition algorithm; the fatigue monitoring algorithm is divided into blink detection, mouth closing detection, eyebrow picking degree detection and nose spacing detection.

The face detection algorithm is to obtain 68 face feature point detection maps shown in fig. 3 by calling shape _ predictor _68_ face _ landmark, i.e. a face feature point detector specific to dlib of the dlib computer deep learning open source library. The characteristics of the areas such as eyes, mouth, eyebrows, nose and the like in the human face are very obvious, and the main morphological change of the human face is concentrated in the areas; the shape _ predictor _68_ face _ maps feature point model can well detect the key parts of the face and return the coordinates of the key parts, the Euclidean geometric distance between the key parts can be calculated by combining the coordinates, then the change of the Euclidean geometric distance of the key areas when the state of the face changes is collected, and the detection and expression recognition of the face areas are realized by judging a threshold value.

1. The blink detection algorithm is implemented as follows:

eyes are used as key feature parts of a human face and are very important to detect, when people have different expression changes and fatigue, the blinking frequency and amplitude of the eyes of the people can be changed, and the change condition of the European geometric distance between 12 feature points (6 feature points of each eye) on the eyes of the people is collected when the state of the faces of the people is changed; then, setting a proper threshold value, the relationship between the change of the blink frequency and the human face state can be counted, and the following algorithm design about the eyebrow bending degree and the mouth closing frequency is the same principle. The principle of feature point selection is to select points that describe the contours and general motion features of the eyes, eyebrows, and mouth. The 68 feature points are automatically selected when the model is called using the dlib library as in fig. 3. The threshold formula for the mouth is the same principle as the eye, but in a somewhat different form. The eyebrow requires a more complex fitting curve function formula.

The blink detection algorithm used by the method is different from the traditional blink image processing method for calculating, the traditional blink detection is carried out by comparing the threshold value of the human eye white area in the shot picture with the threshold value in the model, and the method is greatly interfered by external factors such as picture quality, light rays and the like. However, the eye feature point acquisition method used in the project is a method which is simpler and more efficient to use and has strong robustness.

As shown in fig. 4, the aspect ratio of the eye is approximately constant when the eye is open, according to a designed functional algorithm. The eye threshold at opening is about 0.25; but when eye closure occurs, the threshold rapidly drops to around 0.1 or even close to 0. This simple equation can be derived to avoid image processing techniques, simply relying on the ratio of eye landmark distances to determine whether a person is blinking. The algorithm adopted for blink detection will be described in detail below.

As can be seen from the above formulas, p1 to p6 respectively represent 6 eye feature points of the human eye starting from the left clockwise; the distance of the human eye characteristic points in the vertical direction is calculated by two models of the numerator, the distance of the human eye characteristic points in the horizontal direction is calculated by a model of the denominator, and the multiplication of 2 is used for fitting the number of terms of the numerator, so that the numerator and the denominator are ensured to be the same in weight. When the real-time face detection is carried out, a real-time EAR is obtained through the formula, and then the EAR is compared with a reference EAR, so that whether the eye blinks or not can be judged.

2. Mouth closure detection algorithm implementation:

the mouth closing detection principle is the same as the blink detection, and accurate detection of mouth opening and closing and bending degrees can be realized by using a plurality of characteristic points on the mouth provided by shape _ predictor _68_ face _ landworks. Meanwhile, a correlation function method is designed, and then data are tested for many times, so that an optimal function can be obtained through continuous iteration.

As shown in fig. 5, it can be seen that the normal mouth threshold is about 0.45, and the mouth threshold can reach above 0.5 when the mouth is opened, i.e. the mouth opening threshold is considered to be 0.5, and when the threshold is greater than 0.5, the mouth is considered to be opened. The formula is as follows:

the above formula is to use two norms to calculate the threshold of the feature points of the mouth, specifically to calculate the square sum and root number of the distance between the corresponding feature points; the principle here is the same as the establishment of the eye threshold, but the difference is that the selected feature points are different, so the result of the critical threshold for mouth openness will also be different, and 4 landmark points for the upper and lower lips are selected here, because mouth openness is most obviously effective here.

3. The eyebrow bending degree detection algorithm is realized:

in the eyebrow bending detection, linear fitting needs to be carried out on ten feature points (five feature points of the left eyebrow and the right eyebrow) on the eyebrows, and fitting of a linear function can be easily realized by using a polyfit function in numpy. After fitting a linear function, judging the eyebrow picking degree through the slope. First, two array lists are required to be established to store the x and y position information of the eyebrows, here two lists of line _ brow _ x and line _ brow _ y are established. Secondly, two variables, brow _ sum and brown _ sum, are also needed to be set, and are respectively used for storing the sum of the heights of the feature point coordinates of the eyebrows and the sum of the distances between the eyebrows. And taking the line on the rectangular frame as a horizontal axis, summing the height, namely the sum of the differences between each coordinate and the ordinate of the upper left corner coordinate of the rectangular frame, and summing the length, namely the sum of the differences between each coordinate and the abscissa of the upper left corner coordinate of the rectangular frame, so that the finally-obtained slope needs to be converted. Then, the lists line _ break _ x and line _ break _ y are grouped by the array function in numpy, and are subjected to least squares polynomial fitting by the np. This fitting equation is mathematically expressed as:

in the above formula, | p (x)_j)-y_jL represents the square of the difference between the absolute values of the coefficients minimizing the square error between the feature point and the fixed point; wherein p (x)_j) Coefficient representing the minimum squared error of the feature point, y_jCoefficients representing a fixed-point minimized square error; this formula is the best solution to minimize the squared error, and is divided by the degree of the appropriate polynomial degree (x, y). The vector of coefficients p that minimizes the squared error is returned. Meanwhile, since the eyebrow is curved, it is also necessary to detect characteristics of the eyebrowThe length and height of the point from the face frame are shown in fig. 6, and the eyebrow picking degree is analyzed by judging the proportion of the length and height of the eyebrow in the face frame.

The eyebrow threshold value when angry is generated is calculated, and the actual test is 0.1, which is more suitable; namely, the threshold value is around 0.25 in normal times, and when gas is generated, the threshold value rapidly drops below 0.1, and 0.1 is taken as a critical threshold value.

4. The nose bridge distance detection algorithm is realized as follows:

the length of the nose bridge, the width of the nose and the like of the face can be changed when the facial expression is changed, as shown in fig. 7, the length of the nose bridge is shortened when the face is angry, the width of the nose is increased, the nose is relaxed naturally, and the length of the nose bridge is longer. The accuracy of facial expression recognition can be increased by counting the relationship between the degree of change and facial expression change.

Body temperature monitoring module

MLX90614 is an infrared measurement non-contact thermometer, and an infrared induction thermopile detector chip and a signal processing special integrated chip are packaged and integrated inside the MLX 90614. Meanwhile, a low-noise amplifier, a 17-bit analog-to-digital converter and a strong digital signal processing unit are integrated, so that high-precision and high-resolution measurement can be realized. MLX90614 controls the detection and calculation of object and ambient temperatures by an internal state machine. The end result is output via either the PWN or SMBus bus. The module schematic diagram is shown in FIG. 8;

as shown in fig. 9, MLX90614 controls the measurement and calculation of the object temperature and the ambient temperature by an internal state machine, performs temperature post-processing, and outputs the result in a PWM or SMBus mode (SMBus output is selected here, PWM is selected if specific digital body temperature data is desired, and PWM output mode is selected if the measurement data is represented by a fixed frequency and a certain duty ratio). MLX90614 controls the measurement and calculation of object temperature and environment temperature by an internal state machine, performs temperature post-processing, and outputs the result in PWM or SMBus mode. The internal working logic diagram is shown in fig. 9, and the working principle is that the infrared thermoelectric sensor measures the external temperature and outputs signals to an internal operational amplifier for processing and amplification, then analog signals are converted into digital signals through an A \ D converter, then the digital signals are modulated, processed and output through a digital filter of a DSP, and then the results are stored in an internal RAM. 32 16B EEPROM memory cells are arranged in the MLX90614, and some basic settings of temperature measurement can be changed by rewriting the addresses of the memory cell registers.

The IR sensor includes several serially connected thermocouples with cold junction set on the chip substrate and hot junction set on the film. The film cools by heating itself by absorption or radiation of IR. The output signals of the thermopile are:

V_ir(Ta，To)＝A.(To⁴-Ta⁴)，

where To is the absolute temperature of the object (kelvin), Ta is the absolute temperature of the sensor chip, and a is the sensitivity.

An additional on-chip temperature sensor is used to measure the temperature of the chip. After measuring the outputs of the two sensors, the corresponding ambient temperature and object temperature are calculated. The calculations are processed by an internal DSP which produces a digital output and is linearly proportional to the measured temperature. The temperature of the sensor chip is measured by a PTC or PTAT element, all states and data processing of the sensor are carried out in the chip, and the processed linear sensor temperature Ta is stored in the chip memory. The calculated temperature output resolution is 0.02 ℃ and the factory calibration range of the sensor is 40 … +125 ℃. In RAM cell address 006h, 2DE4h corresponds to-38.2 ℃ (linear output minimum), and 4DC4h (19908d) corresponds to 125 ℃. The RAM contents are converted to the actual Ta temperature by:

Ta[K]＝Tareg×0.02，or 0.02K/LSB

and for the object temperature To, the output result resolution is 0.02 ℃, and is stored in the RAM. The actual temperature of To is:

To[K]＝Toreg×0.02，or 0.02K/LSB

in the above equation, Toreg is the temperature of the object measured by the internal linear sensor; based on the above measurement results, the corresponding ambient temperature Ta and object temperature To are calculated, both with a temperature resolution of 0.01 ℃. Ta and To can be read in two ways: the RAM cell is read (0.02 ℃ resolution, fixed range) through a two-wire interface or output through a PWM digital mode. (10 bit resolution, range configurable)

The last step of the measurement cycle is: the measured Ta and To are readjusted To the output resolution required for PWM and this data is stored in registers of the PWM state machine, which can generate fixed frequency and a certain duty cycle representing the measured data;

as shown in fig. 10, the 3.3V power supply voltage connects the circuits of MLX90614 and SMBus. The MLX90614 has a clamp diode connected between the SDA/SCL and Vdd to power the MLX90614 device, while leaving the SMBus line as no load. The PWM output mode of MLX90614 is relatively simple, with the PWM mode being free running after EEPROM configuration as PWM, POR. For PWM mode operation, the SCL pin must be high. (may be shorted to the Vdd pin.

For example, with PWM as the default mode as shown in fig. 11, a pull-up resistor may be used to maintain the output mode of the SMBus,

similarly, as shown in fig. 12, when SMBus is present, POR mode is configured in EEPROM, and at this time, SCL line is high impedance. The purpose of the external pull-up resistor R1 is to ensure that the SCL line is high and that the PWMPOR default mode is also active. SMBus is still available. (e.g., to further reconfigure MLX90614, or to power management for sleep mode).

Third, heart rate monitoring module

Heart rate monitoring module:

max30100 is an integrated pulse blood oxygen and heart rate detection sensor, and the sensor is provided with two LEDs and a high-precision photoelectric sensor, and an optimized optical signal processor and an optimized analog signal processor are arranged in the sensor. The module is mainly applied to wearable equipment, fitness equipment and medical monitoring equipment, and a schematic circuit diagram of the module is shown in fig. 13.

MAX30100 detects the blood oxygen saturation and heart rate of a person using the principle of photoplethysmography. The basic principle is to measure the pulse and the blood oxygen saturation by utilizing the difference of the light transmittance of human body cells to the arterial blood and the venous blood. The sensor is composed of a photoelectric emitter and a photoelectric converter. It can then be secured to the patient's ear lobe, finger, wrist, etc. When measuring the blood oxygen saturation, calculating according to the absorption amount of the oxyhemoglobin (HBO2) and the Hemoglobin (HB) to infrared light; because the wavelength range of the external red light is approximately (850-1000nm), and the wavelength range of the red light is (600-750 nm). The absorption rates of red light and infrared light by blood cells and oxygenated red blood cells are different. Blood vessels will expand and contract with the heartbeat to generate different amounts of oxyhemoglobin, and when the blood volume is large, the amount of oxyhemoglobin is large, and the amount of absorbed infrared light is large. When the patient contracts, the amount of blood is small, and the amount of blood protein is large, so that the infrared light absorbed is small. The heart rate can be calculated from the change in the amount of light reflected by the cells to the sensor.

As shown in fig. 14, when the blood volume is large at vasodilation, i.e., oxygenated hemoglobin (HBO2) is large, it absorbs infrared light (IR) much; when the blood vessel is constricted, hemoglobin is high, and RED light (RED) is absorbed much; because the oxygen content of blood is sensitive to the emitted red light, Max30100 emits light with a certain wavelength by using specific red light, then receives the light by using a receiving tube (the emitting and receiving are all 16 bits of original data), and can determine the blood oxygen content according to the intensity of the received reflected light;

as can be seen from fig. 15, RED and IR are RED light and infrared light, and the photoelectric sensor collects the photoelectric signal and inputs the signal to the internal ADC for data acquisition, and then the signal enters the array filter and enters the data register, and finally the signal is output through the IIC bus. It is to be understood that the IIC bus actually requires only SCL and SDA, i.e., the system clock line and the system data line, for transmission. Compared with UART and SPI, the IIC bus is simpler and has small transmission data volume.

IV, transfusion monitoring module

The module takes STC89C52RC as a core and utilizes infrared geminate transistors to detect whether liquid drops drop or not; during dripping, the resistance value of a resistor connected with the infrared pair tube changes, and then a liquid drop dripping signal is output through the LM393 voltage comparator; after receiving the information, the single chip microcomputer calculates the liquid drop speed and displays the calculation result on a liquid crystal display 1602; meanwhile, the information is also transmitted to the mobile phone for display through HC-05 Bluetooth. The module is roughly divided into 9 small modules which are respectively a STC89C52RC singlechip main control module, a sound-light alarm module, a 1602 liquid crystal display module, a key module, a blood return alarm module, a dropping speed detection module, a Bluetooth module, an ULN2003 motor module and a power socket module. The hardware design is as follows:

1. main control module of single chip microcomputer

As shown in fig. 16, the STC89C52RC single chip microcomputer is used, and although the single chip microcomputer is a product produced more than ten years ago, the single chip microcomputer has excellent and stable performance, low power consumption and low price, so the single chip microcomputer is still widely applied to teaching and scientific research and manufacture in colleges and universities. The crystal oscillator is black in appearance, has forty pins, is generally externally connected with a 12MHz crystal oscillator, and has the working voltage of 3V to 5V. The infusion monitoring module is a singlechip which has an 8-bit CPU and a 4K-byte EEPROM and uses an MCS-51 classic kernel, and simultaneously has an 8K-byte program storage unit, and the program storage space required by the infusion monitoring module is completely sufficient. And it has UART, the serial interface of solitary RXD and TXD, this provides very big convenience for using serial ports next to change bluetooth module. It has three timers T0, T1 and T2, compared with 8 16-bit timers at short time of STM series single chip microcomputer, the STC single chip microcomputer is better started, and the programming uses C-based single chip microcomputer language, which is convenient for rewriting.

2. Acousto-optic alarm module

The circuit structure diagram of the sound and light alarm module is shown in fig. 17, and it can be seen from the figure that one end of the resistor R10 is connected with the base stage of the PNP triode 8550, and the other end is connected with the pin P2.1 of the single chip microcomputer. For the PNP transistor, when the base input voltage is high, the transistor is turned off. When the input voltage of the base stage is low level, the triode is conducted. Therefore, when the dropping speed exceeds the set threshold value, the P2.1 can output low level by the singlechip, then the triode is conducted, the buzzer gives an alarm, and the alarm system gives an alarm. When the dripping speed is in a reasonable interval, the P2.1 port is in a high level due to the unique design of the single chip microcomputer, so that an alarm cannot be triggered.

3.1602 liquid crystal display module

As shown in fig. 18, the LCD1602 is a basic liquid crystal display module often used by a single chip, and it means that the display module can display 16 bytes and 2 lines, which is a character type display module LCD 1602. Meanwhile, the backlight type LED lamp is divided into a backlight type and a non-backlight type, and the purpose is to adapt to a dim environment. The reason for selecting 1602 LCD is that it has low cost, simple and stable interface with single chip, and simple programming. When the display is used, the display can display different characters only by enabling the corresponding interface of the singlechip to output corresponding levels. It can display English letters and some common symbols. However, if the P0 interface of 51 scm is used, the corresponding interface must be connected with a pull-up resistor to provide stable output capability.

4. Key module

As shown in fig. 19, the keys are respectively connected to pins P3.5, P3.6, and P3.7 of the single chip, and the corresponding pin is connected to a low level when the key is correspondingly pressed. The simplest key module unit is selected here. The keys are respectively connected to the pins P3.5, P3.6 and P3.7 of the single chip microcomputer, when the keys are pressed down, the corresponding interface of the single chip microcomputer can input low level, and when the single chip microcomputer detects the low level, the corresponding keys are pressed down. It is noted that such a key is not provided with the mechanical debounce function of the advanced key, and when a person presses the key, the input level should ideally be instantaneously changed to a low level. However, it takes time for the single chip to detect the conversion, and if the finger pressed in the time is released and the level conversion is not completed yet, the input level is not converted into the low level all the time, so that the single chip does not detect that the key is dropped, and the detection fails. Therefore, a time delay program is required to be designed, so that the single chip microcomputer detects twice, and a time delay is arranged in the middle of the single chip microcomputer, so that software dithering can be removed, and the reliability of the key module is improved.

5. Dripping speed detection module

As shown in fig. 20, the module is mainly composed of an infrared transmitting tube, an infrared receiving tube and an LM393 voltage comparator. As shown in FIG. 22, the infrared transmitting tube D8 transmits an infrared signal, and D1 receives the signal. The change of the signal causes the resistance of the internal circuit of the infrared receiving tube D1 to change, and the voltage obtained by the resistor R3 and the voltage obtained by the infrared receiving tube D1 are different; i.e., the 6 pin input voltage of LM393 changes. When a droplet is dropped, the resistance value of the infrared receiving tube D1 becomes large, and the voltage divided by pin 6 of the LM393 becomes large. When the voltage at pin 6 of LM393 is greater than the reference voltage at pin 5 set by LM939, pin 7 of LM393 will output a low level causing led D3 to light. Meanwhile, the P3.3 port of the singlechip receives a low level signal, and the drop speed of the liquid drop can be measured by counting the level change of the port within a certain time.

6. Blood return alarm module

As shown in fig. 21, the module principle is the same as the drop rate detection module principle, but the threshold design is different when programming. Because the threshold values of the blood and the dropper medium are greatly different from the threshold values of the liquid drops and the dropper medium, by setting the threshold values, when the blood flows back, the infrared receiving tube receives a corresponding signal, and the LM393 outputs a low level to the port P1.0. In programming, when the port P1.0 detects low level and lasts for a period of time, an alarm is sent out to remind.

7. Bluetooth module

As shown in fig. 22, the HC-05 module is a common simple high-performance bluetooth-to-serial module, which eliminates the trouble and complexity of the intermediate communication protocol, so that the user only needs to know the serial related knowledge to realize the bluetooth transmission of the system. The Bluetooth wireless transceiver is generally connected with the TXD and the RXD of the single chip microcomputer, so that the single chip microcomputer is communicated with other equipment through Bluetooth connection. The module has two working modes of a master mode and a slave mode, and can be applied to various applications. HC-05 uses standard AT commands and the user needs to enter a special command mode when starting the device. If the special command mode is not started, the equipment starts the data mode, and at the moment, the single chip microcomputer can be in wireless communication with other equipment. The Slave role, Master role, and Slave-Loop role are the names of its three working roles. The Slave is a passive connection, i.e. someone else connects me. Master is the Master, and I actively search and connect with others. The Slave-Loop means a passive connection, in which case the module will receive bluetooth master data and transmit it to a remote bluetooth master. When the Bluetooth connection device is used, Bluetooth connection with other devices can be carried out only by setting an AT mode capable of being connected in a matched mode and then setting a serial port baud rate.

8. Drive motor module

The stepping motor is used for regulating and controlling the dripping speed within a reasonable interval, namely when the dripping speed is too fast or too slow, the motor can automatically rotate forwards or backwards to regulate the dripping speed. A stepper motor is an actuator that converts electrical pulses into angular displacements. The stepping motor is a motor which can convert electric pulse into angular displacement. In short, the angle and speed of the motor can be controlled by controlling the number and frequency of the electric pulses, so that the rotation angle and speed of the motor can be adjusted.

The stepping motor used here is 28BYJ48, which is a common four-phase eight-beat motor on a single chip microcomputer, and the working voltage is generally 5V-12V DC. After a series of continuous electric pulses are applied to the stepping motor, the motor can rotate continuously. The motor module is used for achieving the purposes of driving the infusion circuit and regulating and controlling the dripping speed. When the infusion speed exceeds a set threshold value, the motor is started, then the infusion tube is clamped, and the dropping speed is reduced. When the infusion speed is lower than the set threshold value, the motor is driven reversely, and then the dropping speed is increased properly. Meanwhile, because the used interface signal of the single chip microcomputer is not large enough, the control signal of the single chip microcomputer is amplified through the ULN2003 and then output to the corresponding motor interface, and the connection diagram is shown in fig. 23.

9. Power switch module

The DC power socket is adopted, and the self-locking switch principle is utilized, so that the on-off of the power supply can be controlled by a key, and the DC power socket is very convenient. As shown in fig. 24, the power switch has 6 pins, but when in use, only two pins are needed to be connected with 2 pins, and the total number of the pins is four. Wherein 2 feet are connected with high level, 5 feet are connected with low level, 1 foot is connected with the singlechip, and 4 feet are connected with the other end of the singlechip. This has the advantage that when the switch is pressed, the self-locking switch will not automatically return to the original position, but will keep the pressed state, so that the power can be continuously switched on or off, which is the difference between the self-locking switch and the key.

Claims

1. An automated healthcare monitoring system, comprising: the infusion monitoring device comprises a main module, a disease state monitoring module, an infusion monitoring module, a body temperature monitoring module, a heart rate monitoring module and a wireless transmission module; the condition monitoring module, the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are connected with the main module through the wireless transmission module in a Bluetooth mode;

the infusion monitoring module is provided with a main control module, an audible and visual alarm module, a display module, a key module, a blood return alarm module, a dripping speed detection module, a Bluetooth module, a motor module and a power socket switch module; the sound and light alarm module, the display module, the key module, the blood return alarm module, the dripping speed detection module and the Bluetooth module are all electrically connected with the main control module; the main control module adjusts the dropping speed of the infusion tube through the motor module; a power socket switch module is connected between the main control module and the power supply; the dripping speed detection module is provided with an infrared pair tube and a voltage comparator; the blood return alarm module is also provided with an infrared geminate transistor and a voltage comparator; a stepping motor is arranged on the motor module;

be equipped with integrated pulse blood oxygen and integrative sensor of heart rate detection, optical signal processor and analog signal processor on the heart rate monitoring module, integrated pulse blood oxygen and integrative sensor of heart rate detection include two LED lamps and a photoelectric sensor.

2. The automated healthcare monitoring system of claim 1, wherein: the main module is a mobile phone.

3. The automated healthcare monitoring system of claim 1, wherein: the wireless transmission module is a serial port Bluetooth module or a raspberry self-carrying Bluetooth module; the infusion monitoring module, the body temperature monitoring module and the heart rate monitoring module are in Bluetooth connection with the main module through the serial port Bluetooth module; the disease state monitoring module is connected with the main module in a Bluetooth mode through a raspberry self-carrying Bluetooth module.

4. The automated healthcare monitoring system of claim 1, wherein: the main control module on the infusion monitoring module is an STC89C52RC single-chip microcomputer, the Bluetooth module is an HC-05 Bluetooth module, the display module is a 1602 liquid crystal display, the motor module is ULN2003, and the voltage comparator is an LM 393; the stepper motor is 28BYJ 48.

5. The automated healthcare monitoring system of claim 1, wherein: the infrared sensing chip on the body temperature monitoring module is MLX 90614.

6. The automated healthcare monitoring system of claim 1, wherein: the heart rate monitoring module is integrated with a pulse blood oxygen and heart rate detection integrated sensor which is Max 30100.

7. A method of operating an automated healthcare monitoring system according to claim 1, comprising the steps of:

8. The method of operation of an automated healthcare monitoring system according to claim 7, wherein: in the step 1, a signal acquisition device in the disease state monitoring module is pi camera; in the step 2, a DC power socket is adopted as a power socket switch module in the transfusion monitoring module, and the on-off of the power supply is controlled by a key.

9. The method of operating an automated healthcare monitoring system according to claim 7, wherein step 1 specifically comprises the steps of:

step 1.1, the state of illness monitoring module carries out face detection: calling a special human face characteristic point detector of dlib, detecting eye, mouth, eyebrow and nose areas in a human face and returning corresponding coordinates, calculating Euclidean geometric distances among the eye, mouth, eyebrow and nose areas in the human face by combining the coordinates, then collecting changes of the Euclidean geometric distances among the eye, mouth, eyebrow and nose areas when the state of the human face changes, comparing the changes of the Euclidean geometric distances with a set judgment threshold value, and detecting and recognizing the facial area and the expression; if the change of the Euclidean geometric distance reaches a set judgment threshold value, the facial area in the face has an expression corresponding to the judgment threshold value;

in the above formula, EAR is eye aspect ratio, and p1 to p6 respectively represent 6 eye feature points clockwise from the left; two modulo sum p on the molecule₂-p₆||+||p₃-p₅The | is the distance of the characteristic point of the human eye in the vertical direction, and the modulo calculation on the denominator is | p₁-p₄I is the distance P of the characteristic point of the human eye in the horizontal direction₁-p₄Matching the number of terms of the numerator by multiplying 2 to ensure that the weight of the numerator is the same as that of the denominator;

10. The method of operation of an automated healthcare monitoring system according to claim 7, wherein step 3 specifically comprises the steps of:

step 3.1, calculating the output temperature of the thermopile:

V_ir(Ta，To)＝A.(To⁴-Ta⁴)，

conversion of RAM content to ambient absolute temperature Ta:

Ta[K]＝Tareg×0.02，or 0.02K/LSB

the absolute temperature To of the object To be measured is:

To[K]＝Toreg×0.02，or 0.02K/LSB