Portable, multi-scenario, multi-functional emergency chest compression guidance device
By using portable sensor flexible consumables and a split host, combined with Kalman filtering and deep neural network models, the monitoring data is calibrated in real time, solving the problems of applicability and data error in existing emergency rescue devices, and realizing intelligent and efficient emergency rescue guidance in multiple scenarios.
Patent Information
- Application Number
- CN202510758187.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-09
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2045-06-09
AI Technical Summary
Existing emergency chest compression devices have limited functionality, cannot adapt to patients of different ages and body types, and the monitoring data is easily affected by temperature changes, resulting in large errors and failing to achieve intelligent and efficient emergency guidance.
It employs portable sensor flexible consumables and a split main unit, combined with a Kalman filter model with multiple temperature drift and a deep neural network analysis model, to calibrate and monitor data in real time and provide emergency chest compression guidance in multiple scenarios.
It enables intelligent, efficient, and precise emergency care guidance in multiple scenarios, reduces temperature drift errors, is suitable for patients of different body types and ages, prevents cross-infection, and improves the success rate of emergency care.
Smart Images

Figure CN120713745B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of medical assistive products, wearable devices, sensor technology, human-computer interaction, and robotics, and specifically to a method, system, medium, and device for a portable, multi-scenario, and multi-functional emergency chest compression guidance device. Background Technology
[0002] In scenarios such as battlefields, ordinary communities, wilderness emergencies, and disaster relief, there are often situations where professional first aid personnel are not on site or the number of professionals is insufficient. It is necessary for people without professional first aid knowledge to participate in rescue efforts. However, due to the lack of knowledge, experience, and emergency preparedness of ordinary people, there are often situations where manual operation fails, rescue fails, or secondary injuries are caused by operational errors when rescue is carried out. Therefore, professional auxiliary equipment is needed to guide non-professionals in emergency rescue.
[0003] Mainstream emergency chest compression guidance devices on the market have limited functionality, generally employing compression feedback devices for CPR. They can only be used for CPR on adults. These devices are modular and placed on the chest; when the rescuer presses on the module, the pressure is transmitted to the patient's chest through the device. Simultaneously, the device contains displacement and pressure sensors to monitor CPR parameters in real time, which are then displayed and alerted via a host computer. The structure is as follows: Figure 1 As shown. The above scheme has the following problems: (1) Single function: It is only for cardiopulmonary resuscitation emergency care. At the same time, the cardiopulmonary resuscitation CPR compression feedback device is large in size and can only be used for adult cardiopulmonary resuscitation emergency care. It cannot be used for infants and young children. Moreover, the monitoring parameters are single and cannot be intelligently adjusted according to the real-time physical condition of the rescued person. (2) Inadequate structural design: It can only be placed on the patient's chest. When used for other patients, it needs to be disinfected first and then the double-sided adhesive on the device needs to be replaced. It is not suitable for the rescue of multiple patients and is prone to cross-infection. (3) Sensor devices and structural components will cause data drift and structural wear during high-intensity compression, resulting in feedback data errors, affecting the monitoring of real-time rescue indicators and causing patient injury.
[0004] Chinese invention patent CN116019443B discloses a system and method for multi-dimensional compliance testing of chest compression indicators. The system includes a device housing, rubber straps, a voice interaction module, an inertial measurement unit module, a microcontroller module, a communication module, a storage module, a touch and display module, and a power supply module. The method includes compression depth detection, compression frequency detection, arm compression verticality detection, and elbow flexion detection methods, enabling the assessment of the rescuer's compression depth, frequency, and the standardization of arm and elbow posture. Through voice interaction, the rescuer triggers and guides the system's startup and operation via voice information, and the system provides guidance to the rescuer through voice communication. However, this system and method only monitor and guide the rescuer's actions; they cannot monitor the rescued person's various indicators in real time, nor can they dynamically adjust the rescuer's actions based on the rescued person's condition, thus failing to achieve the expected results in assisted rescue. Furthermore, while the aforementioned system and method measure the three-axis inertial data of the attached body using the inertial measurement module, including acceleration, angular velocity, and attitude quaternions, and transmit the real-time acquired inertial data to the central control processing module via a data transmission bus, they do not perform data calibration on the data measured by the inertial measurement module. This ignores data drift and other errors caused by temperature changes under high-intensity compressions, thus affecting the monitoring of real-time rescue indicators and impacting monitoring quality. Therefore, emergency compression guidance devices not only need to monitor the compliance of the rescuer's actions while simultaneously monitoring the state feedback of the rescued person, but also need to improve the accuracy of the monitoring data to reduce monitoring errors and improve the quality and efficiency of emergency guidance. Summary of the Invention
[0005] The purpose of this invention is to overcome the shortcomings of the prior art and to provide a portable, multi-scenario, multi-functional emergency compression guidance device, including a method, system, medium, and equipment, for scenarios such as cardiopulmonary resuscitation, Heimlich maneuver, and expectoration. This device receives patient feedback during the monitoring of rescue actions, adjusts the action guidance content in a timely manner based on the feedback, and calibrates the monitoring data to improve its accuracy.
[0006] To achieve the above objectives, the present invention adopts the following technical solution:
[0007] A portable, multi-scenario, multi-functional emergency chest compression guidance device, comprising a sensor flexible consumable and a split main unit;
[0008] The flexible sensor consumable is used to install monitoring positions on the rescued person and the rescuer. The flexible sensor consumable includes a flexible sensor module and a medical-grade silicone layer. The flexible sensor module adopts a press-fit packaging structure. The medical-grade silicone layer is disposed on the outer layer of the flexible sensor module, and an adhesive layer is disposed on the medical-grade silicone layer for bonding the monitoring position. The flexible sensor module includes a flexible sensor acquisition module and a data processing unit. The flexible sensor acquisition module is connected to the data processing unit. The flexible sensor acquisition module is used to acquire motion characteristic data and temperature data of the monitoring position. The data processing unit is used to calibrate the motion characteristic data of the monitoring position based on the acquired temperature data. The data processing unit is also connected to a separate main unit.
[0009] The split-type host is worn by the rescuer. The split-type host is equipped with an intelligent guidance processing unit and a display and voice module. The intelligent guidance processing unit is connected to the data acquisition and processing unit and the display and voice module respectively. The intelligent guidance processing unit is used to analyze the motion characteristic data of the calibrated monitoring position to obtain motion guidance data.
[0010] Preferably, the flexible sensor module further includes a flexible battery, and the flexible sensor acquisition module includes an inertial measurement unit, a press-type flexible pressure sensor, and a flexible temperature sensor. The data acquisition and processing unit is connected to the inertial measurement unit, the press-type flexible pressure sensor, the flexible temperature sensor, and the flexible battery, respectively. The motion characteristic data includes speed and pressure data.
[0011] Preferably, the split-type host adopts a smart bracelet and / or smart glasses. A first opening is provided on the medical silicone layer, and a data acquisition and transmission interface is provided at the first opening. The data acquisition and transmission interface is connected to the data acquisition and processing unit. The data acquisition and transmission interface is used to connect an external monitoring device or a power supply. The smart guidance processing unit is also connected to a data transmission interface. The data transmission interface is used to connect an external monitoring device, or is connected to the data acquisition and transmission interface to power the flexible sensor module.
[0012] Preferably, S1: Acquire real-time motion characteristic data and temperature data of the monitoring position, wherein the motion characteristic data includes acceleration, angular velocity and pressure;
[0013] S2: Perform local weighted regression filtering on the real-time acquired motion feature data; construct a Kalman filter model with multiple temperature drift, and calibrate the motion feature data after local weighted regression filtering using the Kalman filter model with multiple temperature drift based on the temperature data to obtain calibrated motion feature data;
[0014] S3: Construct a deep neural network analysis model and train the deep neural network analysis model based on historical data;
[0015] S4: After preprocessing the patient's body data and calibrated motion feature data, the body parameters and motion feature parameters are respectively input into the trained deep neural network analysis model. After analysis by the deep neural network analysis model, motion parameter indicators are obtained and the motion parameter indicators are output and displayed. The motion parameter indicators are used to reflect standard rescue action parameters.
[0016] The steps for constructing the Kalman filter model with multiple temperature drifts in step S2 are as follows:
[0017] S2.1: Set temperature range, dividing the set temperature range into multiple temperature intervals.
[0018] S2.2: Using the acceleration, angular velocity, and pressure data for each temperature range as input, and the displacement, pressure, and velocity data for each temperature range as output, construct a Kalman filter model for each temperature range. The specific formula is as follows:
[0019] ;
[0020] ;
[0021] ;
[0022] Where α=(α1, α2, α3), β=(β1, β2, β3), and γ=(γ1, γ2, γ3) are the Kalman filter model matrix coefficients corresponding to acceleration a, angular velocity w, and pressure p, respectively. a, w, and p are the input acceleration, angular velocity, and pressure, respectively. depth, speed, and pressure are the output displacement, velocity, and pressure, respectively. k is the current prediction period, and k-1 is the previous prediction period.
[0023] S2.3: Based on historical data, the Kalman filter model for each temperature range is trained to obtain a Kalman filter model with multiple temperature drifts.
[0024] Preferably, the method for training the Kalman filter model for each temperature range in step S2.3 is as follows: iterative training is performed using nonlinear regression fitting based on the Gauss-Newton iterative method, and the formula used for iterative training is:
[0025]
[0026] Among them, L T () represents the target residual function for a single temperature range T, where i is the iteration data number and M is the total number of data.
[0027] Preferably, the method for constructing the deep neural network analysis model in step S3 includes:
[0028] The deep neural network analysis model includes an input layer, three hidden layers, and an output layer. The input layer vector includes body parameters and motion feature parameters, and the output layer vector includes motion parameter indices.
[0029] The hidden layer includes a first hidden layer, a second hidden layer, and a batch normalization layer arranged sequentially. The batch normalization layer is connected to the output layer. The activation functions of the first hidden layer, the second hidden layer, and the batch normalization layer are all Leaky ReLU activation functions. The activation function of the output layer is Sigmoid activation function for parameter regression output. The output layer also maps the output to the target level range [a, b] through scaling and translation.
[0030] Preferably, the shaping preprocessing in step S4 adopts single-precision floating-point shaping, and the action parameter indicators include the pressure guidance level, the compression depth guidance level, the chest recoil guidance level, the ventilation rate guidance level, and the compression rate guidance level.
[0031] It also includes step S5: classifying the calibrated action feature data into levels, obtaining the action parameter implementation level when the rescuer performs the action, comparing the action parameter implementation level with the action parameter index to obtain action adjustment information, and outputting and displaying the action adjustment information;
[0032] The implementation levels of the action parameters include the implementation level of compression force, the implementation level of compression depth, the implementation level of chest recoil, the implementation level of ventilation rate, and the implementation level of compression frequency.
[0033] A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the device on which the computer-readable storage medium is located performs the portable, multi-scenario, multi-functional emergency chest compression guidance method.
[0034] An electronic device includes a memory and a processor, wherein the memory stores a program executable on the processor, and the processor executes the program to implement the portable, multi-scenario, multi-functional first aid chest compression guidance method.
[0035] A portable, multi-scenario, multi-functional emergency chest compression guidance system includes an emergency chest compression guidance device and the aforementioned electronic device.
[0036] The beneficial effects of this invention are:
[0037] The device of this invention detects the vital signs and movement information of the rescuer and the rescued at the monitoring position through a flexible sensor module. The data acquisition and processing unit calibrates the movement feature data of the monitoring position based on the collected temperature data. The first and second intelligent guidance processing units on the smart bracelet and smart glasses analyze the calibrated movement feature data of the monitoring position to obtain movement guidance data. The display and voice modules provide visual and voice display to guide the rescuer to adjust their movements. It realizes emergency compression guidance in multiple scenarios such as cardiopulmonary resuscitation, Heimlich maneuver, and percussion and expectoration. At the same time, the independent design of the flexible sensor consumables and the separate main unit makes it easy to stick or wear, enabling one-to-many emergency rescue solutions and preventing cross-infection among patients in the treatment situation.
[0038] The device of the present invention sets the flexible consumable material of the sensor as a press-type packaging structure, which is small and thin and can be attached to patients of different ages and body types, not limited to infants and adults; and it is easy to attach to different monitoring positions of rescuers and multiple rescued persons to realize one-to-many emergency rescue plan and prevent cross-infection of patients in some treatment situations.
[0039] This invention provides intelligent, efficient, and precise emergency treatment guidance. By receiving patient feedback during the monitoring of rescue actions, the method adjusts the guidance content promptly based on this feedback. Simultaneously, the monitoring data is calibrated to improve accuracy. A Kalman filter model with multiple temperature drift is used to calibrate the acquired monitoring data, reducing errors caused by temperature drift, improving data acquisition accuracy, and ensuring data validity in complex temperature scenarios. The calibrated data is then intelligently analyzed using a deep neural network analysis model, enabling intelligent adjustments to the rescuer's actions. Training the deep neural network analysis model with clinical big data further enhances the accuracy and success rate of the guidance. Attached Figure Description
[0040] The present invention will now be described in further detail with reference to the accompanying drawings:
[0041] Figure 1 This is an installation diagram for testing a traditional emergency chest compression guidance device;
[0042] Figure 2 This is a schematic diagram of the flexible sensor consumable material of the device of the present invention being placed on the chest of a human body;
[0043] Figure 3 This is a schematic diagram of the flexible sensor consumable material of the device of the present invention being placed at the base of the human palm.
[0044] Figure 4This is a schematic diagram of the flexible sensor consumable material of the device of the present invention being placed at the fingertip of a human body;
[0045] Figure 5 This is a schematic diagram of the flexible sensor consumable material of the present invention being placed at the web of the human body;
[0046] Figure 6 This is a schematic diagram of the structure of the flexible sensor consumable of the device of the present invention;
[0047] Figure 7 This is an exploded view of the flexible sensor consumable material of the device of the present invention;
[0048] Figure 8 This is a schematic diagram of the flexible sensor module of the device of the present invention;
[0049] Figure 9 This is a schematic diagram of the structure of the smart bracelet of the present invention;
[0050] Figure 10 This is a cross-sectional structural schematic diagram of the rotating shaft device of the smart bracelet of the present invention;
[0051] Figure 11 This is a schematic diagram of the structure of the smart glasses of the present invention;
[0052] Figure 12 This is a flowchart illustrating the method of the present invention for obtaining calibrated motion feature data;
[0053] Figure 13 This is a flowchart of the deep neural network analysis model of the method of the present invention;
[0054] Figure 14 This is a training effect curve of the deep neural network analysis model of the method of the present invention;
[0055] Figure 15 This is a display diagram showing the effect of the action parameter indicators of the method of the present invention.
[0056] Explanation of reference numerals in the attached figures:
[0057] 1-Flexible sensor consumable; 101-Flexible sensor module; 102-Medical silicone layer; 103-Adhesive layer; 1011-Data acquisition and processing unit; 1012-Flexible battery; 1013-Inertial measurement unit; 1014-Pressure-type flexible pressure sensor; 1015-Flexible temperature sensor; 104-Data acquisition and transmission interface; 201-Display housing; 202-Dial; 203-Display; 204-First speaker; 205-Strap; 2061-Hinge; 2062-Semi-circular base; 207-First data transmission interface; 301-Eyeglass lens; 302-Second speaker; 303-Second intelligent guidance processing unit; 304-Second data transmission interface; 305-Power switch. Detailed Implementation
[0058] The present invention provides a portable, multi-scenario, multi-functional emergency compression guidance device, including a sensor flexible consumable 1 and a split main unit.
[0059] like Figure 2-5 As shown, the flexible sensor material 1 is used to install monitoring positions on the rescued person and the rescuer. These monitoring positions include a vital sign monitoring position attached to the rescued person and a movement monitoring position attached to the rescuer. The vital sign monitoring position for the rescued person is located at the patient's emergency site, including the chest, abdomen, or back. The movement monitoring position for the rescuer includes the heel of the hand, fingers, and the web of the hand. Figure 6 and Figure 7 As shown, the flexible sensor consumable 1 includes a flexible sensor module 101 and a medical silicone layer 102. The flexible sensor module 101 adopts a press-fit packaging structure, and the medical silicone layer 102 is disposed on the outer layer of the flexible sensor module 101. An adhesive layer 103 is disposed on the medical silicone layer 102, which is used to bond the monitoring position. In this embodiment, the adhesive layer 103 is medical double-sided tape.
[0060] In this embodiment, the flexible sensor consumable 1 is attached to the emergency site of the rescued person according to the actual rescue scenario. For example, when performing cardiopulmonary resuscitation, the flexible sensor consumable 1 is attached to the chest of the rescued person; when performing the Heimlich maneuver, the flexible sensor consumable 1 is attached to the abdomen of the rescued person; when performing the percussion and expectoration maneuver, the flexible sensor consumable 1 is attached to the back of the rescued person, realizing multi-scenario and portable rescue.
[0061] like Figure 8As shown, the flexible sensor module 101 includes a flexible sensor acquisition module, a data processing unit 1011, and a flexible battery 1012. The flexible sensor acquisition module is connected to the data processing unit 1011. The flexible sensor acquisition module is used to acquire motion characteristic data and temperature data of the monitoring position. The data processing unit 1011 is used to calibrate the motion characteristic data of the monitoring position based on the acquired temperature data. The data processing unit 1011 is also connected to the split-type host. The flexible sensor acquisition module includes an inertial measurement unit 1013, a pressure-plate type flexible pressure sensor 1014, and a flexible temperature sensor 1015. The data processing unit 1011 is connected to the inertial measurement unit 1013, the pressure-plate type flexible pressure sensor 1014, the flexible temperature sensor 1015, and the flexible battery 1012, respectively. The motion characteristic data includes speed and pressure data. In this embodiment, the flexible battery 1012 is a rechargeable thin-film battery to prevent battery failure caused by compression or bending during the rescue process. The medical silicone layer 102 has a first opening, and a data acquisition and transmission interface 104 is provided at the first opening. The data acquisition and transmission interface 104 is connected to the data acquisition and processing unit 1011. The data acquisition and transmission interface 104 is used to connect external monitoring equipment or power supply. The monitoring equipment can be a third-party pulse oximeter and a ventilation oxygen concentration meter.
[0062] In this embodiment, the data acquisition and processing unit 1011 uses a flexible substrate to package the MCU module based on chip-scale package (CSP) technology. The MCU contains a Bluetooth module, a wireless transmission module, and a wired transmission module, which is suitable for application scenarios that require bending, folding, and squeezing. It can provide better mechanical strength and electronic function stability. At the same time, it is combined with a medical silicone layer 102 to encapsulate the flexible sensor module 101, so that it can be pasted on the emergency site and the rescuer's rescue site without foreign body sensation. At the same time, during the rescue pressure measurement, it can effectively transmit the pressure without hindering the rescue process.
[0063] In this embodiment, the flexible sensor consumable 1 adopts a press-pack packaging structure, which is small and thin, and can be attached to patients of different ages and body types, not limited to infants or adults. It is also easy to attach to different monitoring positions of rescuers and multiple patients, enabling a one-to-many emergency rescue plan. When the device needs to be reused, only the flexible sensor consumable 1 needs to be replaced, which not only buys more time for rescue but also prevents cross-infection among patients in certain treatment situations. By using flexible materials, it will not cause skin damage to rescuers and patients during rescue. At the same time, the press-pack thin sheet structure does not affect the rescuer's wearing of protective gloves, and as a disposable consumable, it eliminates the need to place the same sensor module on different patients' bodies, protecting both rescuers and patients while effectively providing rescue.
[0064] In this embodiment, the split-type main unit is worn by the rescuer and adopts a smart bracelet and smart glasses.
[0065] In this embodiment, as Figure 9 As shown, the smart bracelet is equipped with a first intelligent guidance processing unit and a first display and voice module. The first intelligent guidance processing unit is connected to the data acquisition and processing unit 1011 and the first display and voice module, respectively. The first intelligent guidance processing unit is used to analyze the motion feature data of the calibrated monitoring position to obtain motion guidance data. In this embodiment, the first display and voice module includes a first display device, a first speaker 204, and a first power module. The first display device is a display screen 203. The display screen 203, the first speaker 204, and the first power module are all connected to the first intelligent guidance processing unit. The first display and voice device realizes dual feedback reminders of visual and voice control. The first intelligent guidance processing unit is also connected to a first data transmission interface 207. The first data transmission interface 207 is used to connect external monitoring equipment, connect to the data acquisition and transmission interface 104 to power the flexible sensor module 101, or connect to a host computer to transmit data. The data acquisition and processing unit 1011 and the first intelligent guidance processing unit are wirelessly connected for data acquisition and transmission.
[0066] like Figure 10 As shown, the smart bracelet includes a display screen housing 201 and a watch face 202 rotatably connected via a pivot device. The display screen 203 is located on the front of the display screen housing 201, and the first speaker 204 and the first data transmission interface 207 are both located on the side of the display screen housing 201. The smart bracelet also includes a watch strap 205 connected to the watch face 202, and the watch strap 205 uses a Velcro assembly. By using a Velcro assembly, it is easy for people of different body types to wear, and it can be quickly put on and taken off, which can shorten the preparation time before rescue and gain more golden rescue time. The pivot device includes a pivot 2061 and a semi-circular base 2062. A ball is provided at the first end of the pivot 2061, and a cylindrical boss is provided at the second end of the pivot 2061. A first circular hole is opened in the center of the semi-circular base 2062, and a second circular hole is opened in the center of the dial 202 at the corresponding position of the semi-circular base 2062. The rotating shaft 2061 passes through the first and second circular holes. The sphere at the first end of the rotating shaft 2061 is tangent to the spherical hole at the bottom of the display screen housing 201, and the cylindrical boss at the second end of the rotating shaft 2061 is in close contact with the bottom surface of the dial 202. Through the rotating shaft device, the display screen 203 on the display screen housing 201 can rotate freely on the dial 202, making it convenient for rescue personnel to view.
[0067] In this embodiment, as Figure 11As shown, the smart glasses are equipped with a second intelligent guidance processing unit 303 and a second display and voice module. The second intelligent guidance processing unit 303 is connected to the data acquisition and processing unit 1011 and the second display and voice module, respectively. The second intelligent guidance processing unit 303 is used to analyze the motion feature data of the calibrated monitoring position to obtain motion guidance data. In this embodiment, the second display and voice module includes a second display device, a second speaker 302, and a second power module. The second display device, the second speaker 302, and the second power module are all connected to the second intelligent guidance processing unit 303. The second display and voice device realizes dual feedback reminders of vision and voice control. The second intelligent guidance processing unit 303 is also connected to a second data transmission interface 304. The second data transmission interface 304 is used to connect external monitoring equipment, or to connect to the data acquisition and transmission interface 104 to power the flexible sensor module 101, or to connect to a host computer to transmit data. The data acquisition and processing unit 1011 and the second intelligent guidance processing unit 303 are wirelessly connected for data acquisition and transmission.
[0068] In this embodiment, in certain emergency rescue scenarios, when it is inconvenient for the rescuer to wear a wristband, or when the wristband display is not clear enough, smart glasses can be used simultaneously for data collection and display. The smart glasses include an eyeglass frame and lenses 301. A second display device is integrated onto the lenses 301. A second speaker 302, a second intelligent guidance processing unit 303, and a second data transmission interface 304 are mounted on the eyeglass frame. A power switch 305 is also provided on the eyeglass frame, connected to the second intelligent guidance processing unit 303, and used to turn the smart glasses on and off.
[0069] This embodiment also includes a protection module for safety protection. The protection module includes insulating gloves, which are used to protect the rescuer during AED emergency care, preventing the rescuer from being accidentally shocked by the AED during CPR. The gloves also allow CPR to be performed while the AED is discharging, thus improving the success rate of cardiopulmonary resuscitation.
[0070] In this embodiment, the flexible sensor module 101 detects the vital signs and movement information of the rescuer and the rescued at the monitoring position. The data acquisition and processing unit 1011 calibrates the movement feature data of the monitoring position based on the collected temperature data. The first and second intelligent guidance processing units on the smart bracelet and smart glasses analyze the calibrated movement feature data of the monitoring position to obtain movement guidance data. The display and voice modules provide visual and voice display to guide the rescuer to adjust their movements. This realizes emergency compression guidance in multiple scenarios such as cardiopulmonary resuscitation, Heimlich maneuver, and percussion for expectoration. At the same time, the independent design of the flexible sensor consumable 1 and the split host makes it easy to stick or wear, enabling one-to-many emergency rescue solutions and preventing cross-infection among patients in the treatment situation.
[0071] In this embodiment, the specific process is as follows: First, the flexible sensor consumable 1 is attached to the emergency site of the rescued person according to the actual rescue scenario. For example, when performing cardiopulmonary resuscitation, the flexible sensor consumable 1 is attached to the chest of the rescued person; when performing the Heimlich maneuver, the flexible sensor consumable 1 is attached to the abdomen of the rescued person; when performing the percussion and expectoration maneuver, the flexible sensor consumable 1 is attached to the back of the rescued person. At the same time, the flexible sensor consumable 1 is attached to the palm, fingers, and web of the hand of the rescuer. The rescuer wears a smart bracelet and smart glasses, and then begins the rescue. During the rescue process, the vital signs and movement information of the rescuer and the rescued person detected by multiple flexible sensor modules 101 are calibrated by the data acquisition and processing unit 1011. The calibrated data is transmitted to the first intelligent guidance processing unit and the second intelligent guidance processing unit. After analysis, the action guidance data is output to the display screen 203 or the second display device, and voice prompts are provided. The rescuer adjusts the rescue actions according to the displayed action guidance data and voice broadcast information until the rescue is completed.
[0072] The device of this invention provides guidance for rescue actions in various rescue scenarios. The independent design of the disposable sensor flexible consumable 1 is suitable for monitoring different body types, ages, and locations, facilitating rapid rescue and avoiding cross-infection. It protects both rescuers and rescued while effectively providing rescue, thus ensuring safety for emergency treatment.
[0073] This invention provides a portable, multi-scenario, and multi-functional first aid chest compression guidance method, comprising the following steps:
[0074] S1: Acquire real-time motion characteristic data and temperature data of the monitoring position. The motion characteristic data includes acceleration, angular velocity and pressure.
[0075] S2: Perform local weighted regression filtering on the real-time acquired motion feature data; construct a Kalman filter model with multiple temperature drift, and calibrate the locally weighted regression filtered motion feature data using the Kalman filter model with multiple temperature drift based on the temperature data to obtain calibrated motion feature data. For example... Figure 12 As shown, the steps for constructing a Kalman filter model with multiple temperature drifts are as follows:
[0076] S2.1: Set a temperature range, and divide the set temperature range into multiple temperature intervals; in this embodiment, the set temperature range is 4-40℃, and the multiple temperature intervals are [4,5), [5,6), ..., [40,41).
[0077] S2.2: Take the acceleration, angular velocity and pressure data of each temperature range as input, and the displacement, pressure and velocity of each temperature range as output, and construct a Kalman filter model for each temperature range, such as the 4℃ Kalman filter model, the 5℃ Kalman filter model, ..., the 40℃ Kalman filter model, which are used to process data with temperatures in [4,5), [5,6), ..., [40,41), respectively.
[0078] The specific formulas for the Kalman filter model in each temperature range are as follows:
[0079] ;
[0080] ;
[0081] ;
[0082] Where α=(α1, α2, α3), β=(β1, β2, β3), and γ=(γ1, γ2, γ3) are the Kalman filter model matrix coefficients corresponding to acceleration a, angular velocity w, and pressure p, respectively. a, w, and p are the input acceleration, angular velocity, and pressure, respectively. depth, pressure, and speed are the output displacement, velocity, and pressure, respectively. k is the current prediction period, and k-1 is the previous prediction period.
[0083] S2.3: Based on historical data, the Kalman filter model for each temperature range is trained to obtain a Kalman filter model with multiple temperature drifts.
[0084] The method for training the Kalman filter model for each temperature range in step S2.3 is as follows: Iterative training is performed using nonlinear regression fitting based on the Gauss-Newton iterative method. The formula used for iterative training is:
[0085] ;
[0086] Among them, LT () represents the target residual function for a single temperature range T, where i is the iteration data number and M is the total number of data.
[0087] S3: As Figure 13 As shown, a deep neural network analysis model is constructed and trained based on historical data; the method for constructing the deep neural network analysis model in step S3 includes:
[0088] The deep neural network analysis model includes an input layer, three hidden layers, and an output layer. The input layer vector includes body parameters and motion feature parameters. The body parameters include the patient's height, weight, and gender, while the motion feature parameters include displacement, pressure, and velocity. The output layer vector includes motion parameter indices, such as guidance levels for compression force, compression depth, chest recoil, ventilation rate, and compression frequency. In this embodiment, the input layer vector also includes SpO2 concentration detected by a third-party pulse oximeter and FiO2 concentration detected by a third-party expiratory oxygen saturation meter. When the third-party pulse oximeter and expiratory oxygen saturation meter are not connected, both SpO2 and FiO2 concentrations are set to default values, such as 80% for SpO2 and 17% for FiO2.
[0089] The hidden layer consists of a first hidden layer, a second hidden layer, and a batch normalization layer arranged sequentially. The batch normalization layer is connected to the output layer. Each of the first hidden layer, the second hidden layer, and the batch normalization layer has 16 neurons. The activation functions of the first hidden layer, the second hidden layer, and the batch normalization layer are all Leaky ReLU activation functions. The activation function of the output layer is Sigmoid activation function for parametric regression output. The output layer also maps the output to the target level interval [a, b] through scaling and translation, where a is the minimum value in the target level interval and b is the maximum value in the target level interval.
[0090] In this embodiment, the formula for the Leaky ReLU activation function is:
[0091] ;
[0092] Where, x j For input variables; y j For output variables; a j It is a fixed parameter in the interval (1, +∞).
[0093] In this embodiment, the output layer uses scaling to map the output to the target level range [a, b] using the following formula:
[0094] ;
[0095] Among them, yscaled Let S(x) be the scaled action parameter index, and let S(x) be the Sigmoid activation function.
[0096] In this embodiment, the training dataset collected in the preclinical experiment contains nearly 5,000 data points, which can effectively train a high-precision traffic estimation model. Furthermore, during subsequent use of the device, historical data stored in the split-type host can be exported for secondary training. The model converges during the training process, achieving good prediction results. Figure 14 The training results shown demonstrate that after 200 iterations, the deep neural network analysis model has learned the aforementioned nonlinear factors well, and the average cost function value calculated from the training set data is within 0.6171, indicating that the deep neural network analysis model has excellent predictive performance.
[0097] S4: After preprocessing the patient's body data and calibrated motion feature data, the body parameters and motion feature parameters are respectively input into the trained deep neural network analysis model. After analysis by the deep neural network analysis model, the motion parameter index is obtained and output and displayed. The motion parameter index is used to reflect the standard rescue action parameters.
[0098] In this embodiment, the shaping preprocessing uses single-precision floating-point shaping, which is represented as:
[0099] ;
[0100] Where value is the number before integer preprocessing, sign represents 1 sign bit, e represents 8 exponent bits, and b0~b 22 This represents the 23-bit mantissa, where n is the index of the mantissa digits, and * represents the multiplication operation.
[0101] Step S5: Classify the calibrated action characteristic data into levels to obtain the action parameter implementation levels during rescue. These levels include compression force implementation level, compression depth implementation level, chest recoil implementation level, ventilation rate implementation level, and compression rate implementation level. In this embodiment, the compression force implementation level, compression depth implementation level, and chest recoil implementation level are each divided into 10 levels to suit the vital signs of infants, young children, adults, and the elderly. The ventilation rate implementation level and compression rate implementation level are each divided into 5 levels.
[0102] The motion adjustment information is obtained by comparing the implementation level and the index of the motion parameters, and then output and displayed. In this embodiment, key motion parameter indices and implementation levels are filtered according to the importance of the motion features, and key motion adjustment information is output. When the implementation level of a key motion parameter is less than the index of the key motion parameter, the key motion adjustment information includes information on increasing force, depth, rebound, and frequency. When the implementation level of a key motion parameter is greater than the index of the key motion parameter, the key motion adjustment information includes information on decreasing force, depth, rebound, and frequency. When the implementation level of a key motion parameter is equal to the index of the key motion parameter, the key motion adjustment information includes information on maintaining the current force, depth, rebound, and frequency.
[0103] In this embodiment, the quality is evaluated based on the action parameter implementation level and action parameter index according to the evaluation mechanism, and the emergency evaluation quality is displayed through the first display device and the second display device.
[0104] like Figure 15 As shown, the action parameters, implementation level, action parameter indicators, and action adjustment information obtained through the above method are displayed on a screen or a second display device of smart glasses. This displays the emergency assessment quality and key action adjustment information, providing simple instructions to guide the rescuer. This further guides the rescuer to immediately adjust the action characteristic parameters, while the separate main unit records relevant real-time data for subsequent analysis and secondary training of the deep neural network analysis model based on historical data.
[0105] This invention provides intelligent, efficient, and precise emergency treatment guidance. By receiving patient feedback during the monitoring of rescue actions, the method adjusts the guidance content promptly based on this feedback. Simultaneously, the monitoring data is calibrated to improve accuracy. A Kalman filter model with multiple temperature drift is used to calibrate the acquired monitoring data, reducing errors caused by temperature drift, improving data acquisition accuracy, and ensuring data validity in complex temperature scenarios. The calibrated data is then intelligently analyzed using a deep neural network analysis model, enabling intelligent adjustments to the rescuer's actions. Training the deep neural network analysis model with clinical big data further enhances the accuracy and success rate of the guidance.
[0106] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it causes the device containing the computer-readable storage medium to perform the aforementioned portable, multi-scenario, and multi-functional first aid chest compression guidance method. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), random access memory, and other memories.
[0107] The present invention provides an electronic device comprising: a memory and a processor, wherein the memory stores a program that can run on the processor, and the processor executes the program to implement the above-mentioned portable, multi-scenario, multi-functional emergency chest compression guidance method.
[0108] If the modules / units integrated in the electronic device described in this application are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can also be implemented by a computer program instructing related hardware devices. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above.
[0109] The present invention provides a portable, multi-scenario, multi-functional emergency chest compression guidance system, including an emergency chest compression guidance device and the aforementioned electronic device.
[0110] Furthermore, the computer-readable storage medium may primarily include a stored program area and a stored data area, wherein the stored program area may store the operating system, an application program required for at least one function, etc.; and the stored data area may store data created based on the use of blockchain nodes, etc.
[0111] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.
Claims
1. A portable, multi-scenario, multi-functional first aid chest compression guidance device, characterized in that, This includes flexible sensor consumables and separate main unit components; The flexible sensor consumable is used to install monitoring positions on the rescued person and the rescuer. The flexible sensor consumable includes a flexible sensor module and a medical-grade silicone layer. The flexible sensor module adopts a press-fit packaging structure. The medical-grade silicone layer is disposed on the outer layer of the flexible sensor module, and an adhesive layer is disposed on the medical-grade silicone layer for bonding the monitoring position. The flexible sensor module includes a flexible sensor acquisition module and a data processing unit. The flexible sensor acquisition module is connected to the data processing unit. The flexible sensor acquisition module is used to acquire motion characteristic data and temperature data of the monitoring position. The data processing unit is used to calibrate the motion characteristic data of the monitoring position based on the acquired temperature data. The data processing unit is also connected to a separate main unit. The split-type host is worn by the rescuer. The split-type host is equipped with an intelligent guidance processing unit and a display and voice module. The intelligent guidance processing unit is connected to the data acquisition and processing unit, the display and voice module respectively. The intelligent guidance processing unit is used to analyze the motion characteristic data of the calibrated monitoring position to obtain motion guidance data. The portable, multi-scenario, multi-functional emergency chest compression guidance device performs the following steps: S1: Acquire real-time motion characteristic data and temperature data of the monitoring position, wherein the motion characteristic data includes acceleration, angular velocity and pressure; S2: Perform local weighted regression filtering on the real-time acquired motion feature data; construct a Kalman filter model with multiple temperature drift, and calibrate the motion feature data after local weighted regression filtering using the Kalman filter model with multiple temperature drift based on the temperature data to obtain calibrated motion feature data; S3: Construct a deep neural network analysis model and train the deep neural network analysis model based on historical data; S4: After preprocessing the patient's body data and calibrated motion feature data, the body parameters and motion feature parameters are respectively input into the trained deep neural network analysis model. After analysis by the deep neural network analysis model, motion parameter indicators are obtained and the motion parameter indicators are output and displayed. The motion parameter indicators are used to reflect standard rescue action parameters. The steps for constructing the Kalman filter model with multiple temperature drifts in step S2 are as follows: S2.1: Set temperature range, dividing the set temperature range into multiple temperature intervals. S2.2: Using the acceleration, angular velocity, and pressure data for each temperature range as input, and the displacement, pressure, and velocity data for each temperature range as output, construct a Kalman filter model for each temperature range. The specific formula is as follows: ; ; ; Where α=(α1, α2, α3), β=(β1, β2, β3), and γ=(γ1, γ2, γ3) are the Kalman filter model matrix coefficients corresponding to acceleration a, angular velocity w, and pressure p, respectively. a, w, and p are the input acceleration, angular velocity, and pressure, respectively. depth, speed, and pressure are the output displacement, velocity, and pressure, respectively. k is the current prediction period, and k-1 is the previous prediction period. S2.3: Based on historical data, the Kalman filter model for each temperature range is trained to obtain a Kalman filter model with multiple temperature drifts.
2. The portable, multi-scenario, multi-functional emergency chest compression guidance device according to claim 1, characterized in that, The flexible sensor module also includes a flexible battery. The flexible sensor acquisition module includes an inertial measurement unit, a press-type flexible pressure sensor, and a flexible temperature sensor. The data acquisition and processing unit is connected to the inertial measurement unit, the press-type flexible pressure sensor, the flexible temperature sensor, and the flexible battery, respectively.
3. The portable, multi-scenario, multi-functional emergency chest compression guidance device according to claim 1, characterized in that, The split-type host adopts a smart bracelet and / or smart glasses. A first opening is provided on the medical silicone layer. A data acquisition and transmission interface is provided at the first opening. The data acquisition and transmission interface is connected to the data acquisition and processing unit. The data acquisition and transmission interface is used to connect external monitoring equipment or power supply. The smart guidance and processing unit is also connected to a data transmission interface. The data transmission interface is used to connect external monitoring equipment or is connected to the data acquisition and transmission interface to power the flexible sensor module.
4. The portable, multi-scenario, multi-functional emergency chest compression guidance device according to claim 1, characterized in that, The method for training the Kalman filter model for each temperature range in step S2.3 is as follows: iterative training is performed using nonlinear regression fitting based on the Gauss-Newton iterative method, and the formula used for iterative training is: ; Among them, L T () represents the target residual function for a single temperature range T, where i is the iteration data number and M is the total number of data.
5. The portable, multi-scenario, multi-functional emergency chest compression guidance device according to claim 1, characterized in that, The method for constructing the deep neural network analysis model in step S3 includes: The deep neural network analysis model includes an input layer, three hidden layers, and an output layer. The input layer vector includes body parameters and motion feature parameters, and the output layer vector includes motion parameter indices. The hidden layer includes a first hidden layer, a second hidden layer, and a batch normalization layer arranged sequentially. The batch normalization layer is connected to the output layer. The activation functions of the first hidden layer, the second hidden layer, and the batch normalization layer are all LeakyReLU activation functions. The activation function of the output layer is Sigmoid activation function for parameter regression output. The output layer also maps the output to the target level range [a, b] through scaling and translation.
6. The portable, multi-scenario, multi-functional emergency chest compression guidance device according to claim 1, characterized in that, The shaping preprocessing in step S4 uses single-precision floating-point shaping, and the action parameter indicators include the pressure level guidance, the compression depth guidance, the chest recoil guidance, the ventilation rate guidance, and the compression rate guidance. It also includes step S5: classifying the calibrated action feature data into levels, obtaining the action parameter implementation level when the rescuer performs the action, comparing the action parameter implementation level with the action parameter index to obtain action adjustment information, and outputting and displaying the action adjustment information; The implementation levels of the action parameters include the implementation level of compression force, the implementation level of compression depth, the implementation level of chest recoil, the implementation level of ventilation rate, and the implementation level of compression frequency.
Citation Information
Patent Citations
Cardiopulmonary resuscitation chest compression compliance detection system and method
CN116019443B
Rescue administration aid device
CA2012614A1
Cardiopulmonary resuscitation assistant device
CN106726437A
