A multi-interface edge computing digital base for cardiopulmonary function assessment

Abstract

The application provides a multi-interface edge computing digital base for cardiopulmonary function evaluation, comprising: a multi-interface data receiving module for receiving cardiopulmonary sensing terminal data and environmental sensing data; a data preprocessing module for performing preprocessing operations such as device protocol analysis, filtering and feature extraction on the received data; an edge computing inference module for deploying a cardiopulmonary function evaluation artificial intelligence model and performing real-time inference evaluation based on the preprocessed data; a data storage and management cloud service module for storing the collected data and evaluation results in a cloud time series database; and a human-computer interaction module for displaying user information input and cardiopulmonary function evaluation results. The application can connect various existing cardiopulmonary function sensing terminals through the multi-interface data receiving module, perform protocol analysis, filtering and feature extraction through the data preprocessing module, and then perform real-time evaluation of cardiopulmonary function through the edge computing inference module.

Description

TECHNICAL FIELD

[0001] The present application relates to a heart-lung function diagnosis and evaluation device, in particular to a multi-interface edge computing digital base for heart-lung function evaluation. BACKGROUND

[0002] At present, the incidence and mortality of heart-lung function decline related diseases represented by ischemic heart disease and chronic obstructive pulmonary disease are among the top in the world. However, heart-lung function monitoring mainly relies on clinical diagnosis and treatment, and heart and lung are managed separately, with single examination and focus on single organ evaluation. Guided by the concept of active health and the theory of integrated medical physiology, it is particularly important to develop related equipment for overall evaluation of heart-lung function for early warning, delaying the occurrence and development of heart-lung function decline related diseases.

[0003] With the progress of wearable technology and Internet of Things technology, there are many wearable devices on the market for monitoring heart function or lung function, such as smart bracelets, heart rate bands, non-invasive blood glucose detectors, etc. These devices have reached the level of professional medical equipment in the measurement accuracy of some physiological parameters, which provides the possibility for portable monitoring of heart-lung function in living space. However, these devices have single function and less integration of gas, electrophysiology, blood oxygen, blood lipid, vascular parameters and other physiological parameters, which cannot meet the evaluation needs in multiple scenes (exercise, rest, sleep). In order to use existing devices to evaluate heart-lung function as a whole, it is necessary to collect and integrate the data generated by existing devices. However, the communication methods and protocols of sensing terminals are different, and the diversity and heterogeneity of sensing terminals bring difficulties to data collection and integration. On the other hand, the amount of multi-source data is huge, and the storage, analysis, integration and evaluation of data will consume more time, which brings certain challenges to real-time evaluation of heart-lung function. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a multi-interface edge computing digital base for heart-lung function evaluation, which can connect various existing heart-lung function sensing terminals through a multi-interface data receiving module, receive multi-source heart-lung sensing data, and perform protocol analysis, filtering and feature extraction through a data preprocessing module, and then perform real-time evaluation of heart-lung function through a heart-lung function evaluation algorithm based on multi-source information fusion deployed by an edge computing inference module.

[0005] To achieve the above technical solutions, the application provides a multi-interface edge computing digital base for cardiopulmonary function evaluation, comprising: a multi-interface data receiving module, providing a plurality of wired and wireless interfaces, for receiving cardiopulmonary sensing terminal data and environmental sensing data; a data preprocessing module, for performing preprocessing operations such as device protocol analysis, filtering, and feature extraction on the received data; an edge computing inference module, for deploying a cardiopulmonary function evaluation artificial intelligence model and performing real-time inference evaluation based on the preprocessed data; a data storage and management cloud service module, for storing the collected data and evaluation results in a cloud time series database, and providing a data management and visualization UI interface; and a human-computer interaction module, for displaying user information input and cardiopulmonary function evaluation results.

[0006] Preferably, the multi-interface data receiving module includes a plurality of wired interface modules and wireless interface modules, the wired interface modules include but are not limited to ADC, UART, USB, SPI, IIC, and CSI interfaces; and the wireless interface modules include but are not limited to BLE, ZigBee, and WiFi interfaces, wherein the BLE interface function is realized based on the Espressif ESP32-C3, the ZigBee interface function is realized based on the TI CC2530, and the WiFi function is realized based on the Intel AX210.

[0007] Preferably, the wireless interface module saves a mac address list of the cardiopulmonary sensing terminal required to be connected, and initiates a connection as soon as the scanned device mac address exists in the list.

[0008] Preferably, the BLE interface module runs in a master-slave mode, and there is a device connection upper limit for a single BLE interface module, and the system increases the number of connectable BLE devices by realizing multi-channel BLE interface module access function, and the specific steps are as follows: initially, the first BLE interface module performs scanning and connection, and when the number of connections reaches the connection upper limit, the digital base is notified to enable the next BLE interface module to perform scanning. The connection upper limit is between 8 and 20 depending on the used BLE protocol stack.

[0009] Preferably, the BLE interface module and the ZigBee interface module simultaneously receive a plurality of sensing terminal device data, and to distinguish the data sources, the received data is encapsulated again and then forwarded to the data preprocessing module through the same interface, wherein the encapsulation format is: a 2-byte length frame header, which is 0xAA, 0xAA, a 1-byte length data source identifier, a 1-byte length data number, a 1-byte length received data size identifier, and N-byte length received original data, where N is the size of the received data.

[0010] Preferably, the data preprocessing module is implemented based on the RK3588 embedded Linux platform of Ruijie Micro, and can perform protocol analysis, filtering and feature extraction on the received data. The protocol analysis process is to extract target data from the original data according to the data protocol of each sensing terminal. For the secondary packet data transmitted by the wireless interface, the data source and the original data are obtained by unpacking, and then the target data is extracted according to the data source using the corresponding device protocol. The module reserves a protocol setting interface. When a new type of device is added, the device type identification number, the position of the target data in the data packet, the length of the target data, and the byte sequence are filled in the configuration file. The filtering process filters out high-frequency noise, power frequency interference, baseline drift, motion artifacts and other interference of the time series data. The module reserves a setting interface for filtering algorithm related parameters. The feature extraction process extracts physiological feature data related to heart and lung function from the data, including heart rate, heart rate variability, blood pressure, blood oxygen, oxygen uptake, and carbon dioxide output.

[0011] Preferably, the edge computing inference module is implemented based on the RK3588 embedded Linux platform of Ruijie Micro. The platform has a built-in convolution processing unit. The original model of the required deployment algorithm is converted into an RKNN model using RKNN Toolkit2 and deployed on the platform to speed up the algorithm inference speed. The algorithm model is designed using a 1DCNN-LSTM model architecture. The model input is the signal features obtained by the preprocessing module, including heart rate, heart rate variability, blood pressure, blood oxygen, oxygen uptake, and carbon dioxide output. The data length is 30 seconds. Each type of data is processed by 1DCNN and LSTM respectively. Then the output of the LSTM is spliced and input into the fully connected layer to obtain the evaluation results of heart and lung function, including normal heart and lung function, reduced heart and lung function, and abnormal heart and lung function.

[0012] Preferably, the data storage and management cloud service module stores part of the important data and the evaluation results in the Influx DB time series database on the cloud server through WiFi, Ethernet or 4G / 5G mobile communication, using the API interface provided by the influxdb-client-python package. At the same time, a time series database background interface is provided for professionals to query and retrieve data.

[0013] Preferably, the human-computer interaction module displays through the HDMI interface connected display. The interface is implemented based on PyQt5, and is used for inputting user information, displaying the connection state of the heart and lung sensing terminal, displaying heart and lung function parameters and evaluation results. At the same time, a wireless Web UI display interface is provided to improve the portability and ease of use of the device. When the user's mobile phone or computer is in the same local area network as the digital base, the specified port address can be accessed to enter the Web UI. The function of the Web UI interface is the same as that of the wired display interface.

[0014] Preferably, the multi-interface edge computing digital base for heart-lung function evaluation includes two modes of data acquisition mode and normal working mode, when entering the data acquisition mode, the received raw data is directly stored after preliminary analysis, at this time, the data preprocessing module only realizes the protocol analysis function, and the edge computing inference module is in a disabled state; when entering the normal working mode, the data is stored after being collected, preprocessed and obtained by the edge computing inference module.

[0015] The multi-interface edge computing digital base for heart-lung function evaluation has the advantages that:

[0016] (1) The multi-interface edge computing digital base for heart-lung function evaluation connects various existing heart-lung function sensing terminals through a multi-interface data receiving module, receives multi-source heart-lung sensing data, and performs protocol analysis, filtering and feature extraction through a data preprocessing module, and then performs real-time evaluation of heart-lung function through a heart-lung function evaluation algorithm based on multi-source information fusion deployed by an edge computing inference module. At the same time, the data storage function is realized, so that professional personnel can use the data for research.

[0017] (2) The multi-interface data receiving module and the data preprocessing module realized by the multi-interface edge computing digital base for heart-lung function evaluation solve the problems of data collection, analysis and fusion of heterogeneous terminals, and based on the collected data, the multi-source data fusion method can effectively improve the accuracy of the heart-lung function evaluation algorithm; on the other hand, the edge computing inference module realized by the base effectively shortens the analysis, fusion and evaluation time of multi-source heart-lung data, and ensures the real-time performance of the algorithm operation. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is the overall system block diagram of the present application.

[0019] Figure 2 is the data flow diagram in the data acquisition mode and the normal working mode of the embodiment of the present application.

[0020] Figure 3 is the process of connecting multiple heart-lung sensing terminals by the BLE interface module in the embodiment of the present application.

[0021] Figure 4 is the secondary packaging format of the received data by the wireless interface module in the embodiment of the present application.

[0022] Figure 5 is a schematic diagram of the heart-lung function evaluation algorithm model in the embodiment of the present application.

[0023] Figure 6The storage form of the central lung sensing data in the time series database in the embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application.

[0025] Embodiment: A multi-interface edge computing digital base for cardiopulmonary function evaluation.

[0026] At present, some wearable devices can monitor the heart or lung function, such as electrocardiogram clothes, smart bracelets, breathing masks, etc., which can measure physiological parameters such as electrocardiogram, heart rate, blood oxygen, oxygen uptake, carbon dioxide output, etc. In this embodiment, such devices are collectively referred to as cardiopulmonary sensing terminals. The purpose of this embodiment is to use a multi-interface edge computing digital base for cardiopulmonary function evaluation proposed by the present application to collect, fuse, analyze and store the multi-source physiological parameters generated by these devices, so that researchers can use these multi-source data to establish a cardiopulmonary function evaluation algorithm. The base also includes an edge computing inference module to deploy related algorithms, so as to accurately evaluate the cardiopulmonary function in real time in actual application.

[0027] As shown in Figure 1 A multi-interface edge computing digital base for cardiopulmonary function evaluation includes a multi-interface data receiving module, a data preprocessing module, an edge computing inference module, a data storage and management cloud service module, and a human-computer interaction module. The multi-interface data receiving module provides multiple wired and wireless interfaces for receiving cardiopulmonary sensing terminal data and environmental sensing data. The data preprocessing module is used for device protocol analysis, filtering, feature extraction and other preprocessing operations on the received data. The edge computing inference module is used to deploy a cardiopulmonary function evaluation artificial intelligence model and perform real-time inference evaluation based on preprocessed data. The data storage and management cloud service module is used to store the collected data and evaluation results in a cloud time series database and provide data management and visualization UI interface. The human-computer interaction module is used to display user information input and cardiopulmonary function evaluation results.

[0028] As shown in Figure 2As shown, the system includes two modes of data collection mode and normal working mode. In the data collection mode, the digital base connects the cardiopulmonary perception terminal, and the received raw data is directly stored after preliminary analysis. At this time, the data preprocessing module only realizes the protocol analysis function, and the edge computing inference module is in a disabled state. In the normal working mode, the data is collected, preprocessed, and the cardiopulmonary function evaluation result is obtained through the edge computing inference module, and the preprocessed data and the evaluation result are stored. The data collection mode is used to establish the data set in the algorithm research, and the normal working mode is used for daily use after the algorithm research and deployment.

[0029] In the multi-interface data receiving module, there are various wired and wireless data interface modules that can receive cardiopulmonary perception terminal data and environmental sensor data of different protocols and formats. In this embodiment, the wired interface module includes ADC, UART, USB, SPI, IIC, and CSI interfaces; the wireless interface module includes BLE, ZigBee, and WiFi interfaces. In some other embodiments, interfaces can be added as needed. Preferably, the BLE interface function is realized based on the Espressif ESP32-C3, the ZigBee interface function is realized based on the TI CC2530, and the WiFi function is realized based on the Intel AX210.

[0030] In the wireless interface module, the mac address list of the required connected cardiopulmonary perception terminal is saved, and once the scanned device mac address exists in the list, the connection is initiated actively. In this embodiment, the mac address list is saved in the nvs memory, and in some embodiments, the mac address can be saved in the program body or can be sent to the wireless interface module by the host computer before each run.

[0031] The BLE interface module runs in a master-slave mode, and there is a connection limit for a single BLE interface module. According to different protocol stacks used, the connection limit is often between 8 and 20. In this embodiment, the BLE protocol stack of a single ESP32-C3 can only support up to 8 BLE devices to be connected simultaneously, and the system realizes the function of multi-channel ESP32-C3 controller access, that is, when the number of currently connected ESP32-C3 is equal to 8, the next BLE interface module is enabled. Figure 3 The process of connecting multiple cardiopulmonary perception terminals by the BLE interface module is shown.

[0032] The BLE interface module and the ZigBee interface module can simultaneously receive multiple sensing terminal device data. Since the wireless interface module forwards the received multiple-source cardiopulmonary sensing terminal data to the data preprocessing module through the same interface, in order to enable the digital base to distinguish the data sources, the received data needs to be encapsulated again. In the embodiment, the encapsulation format of the received data by the wireless interface module is as shown in the following table: Figure 4 Specifically, the encapsulation format is: a 2-byte length frame header, 0xAA, 0xAA, a 1-byte length data source identifier, a 1-byte length data number, a 1-byte length received data size identifier, and N-byte length received original data, where N is the size of the accepted data.

[0033] The data preprocessing module in the embodiment is implemented based on the Rui-Kai Micro RK3588 embedded Linux platform, and can perform protocol analysis, filtering, feature extraction on the received data. In some embodiments, time alignment, data segmentation, signal denoising, normalization processing, peak detection, data rejection and other preprocessing operations can also be performed.

[0034] The protocol analysis is to extract target data from the original data according to the protocol of each sensing terminal. For example, in a PPG module that needs to be connected in the embodiment, the length of a frame of data sent by the module is 20 bytes, of which the 12th to 15th bytes store the original value of a 24-bit ADC, which is the target data, and the known array is stored in little-endian byte order. Therefore, the target data value can be restored according to the 4-byte data. For the secondary encapsulated data transmitted by the wireless interface, the data source and the original data need to be obtained by unpacking, and then the target data is extracted according to the data source using the corresponding device protocol. In the embodiment, after detecting two consecutive 0xAA in the data stream, the next byte is the data source, the next byte is the data number, the next byte is the data length N, and then the N-byte length sequence read in succession is the original data. Then the target data can be obtained according to the corresponding protocol. In addition, the module reserves a protocol setting interface. When a new type of device is added, the device type identifier, the position of the target data in the data packet, the length of the target data, and the byte order and other information need to be filled in the configuration file so that the program can complete the extraction of the target data.

[0035] The filtering is to filter out high-frequency noise, power frequency interference, baseline drift, motion artifacts and other interference of the time series data. The module reserves a setting interface for filter-related parameters.

[0036] The feature extraction is to extract physiological feature data related to cardiopulmonary function from the data, including heart rate, heart rate variability, blood pressure, blood oxygen, oxygen uptake, carbon dioxide output and other features.

[0037] Further, the edge computing inference module is implemented based on the Rockchip RK3588 embedded Linux platform, which has a built-in convolution processing unit (NPU). By using RKNN Toolkit2, the required deployment algorithm model is converted into an RKNN model and deployed on the platform, effectively accelerating the algorithm inference speed. Preferably, before deploying the algorithm, data should be collected using the data acquisition module of the digital base and the original algorithm model should be developed based on the stored data. The original algorithm model can be developed based on Caffe, TensorFlow, TensorFlow Lite, ONNX, Pytorch, MXNet, etc. After that, it is converted into an RKNN model for deployment to speed up the inference speed. In this embodiment, the algorithm model is designed using a 1DCNN-LSTM model architecture, as shown in Figure 5 The model input is the signal features obtained by the preprocessing module, including heart rate, heart rate variability, blood pressure, blood oxygen, oxygen uptake, carbon dioxide output, etc., and the data length is 30 seconds. Each type of data is processed by 1DCNN and LSTM respectively, and then the output of the LSTM is spliced and input into the fully connected layer to obtain the evaluation result of the cardiopulmonary function, including normal cardiopulmonary function, reduced cardiopulmonary function, and abnormal cardiopulmonary function.

[0038] Further, the data storage and management cloud service module stores some important data and evaluation results in the Influx DB time series database on the cloud server through WiFi, Ethernet or 4G / 5G mobile communication, using the API interface provided by the influxdb-client-python package. At the same time, a time series database background interface is provided for professionals to query and retrieve data.

[0039] A complete time series data entry contains time, fields and tags. They represent key-value pairs corresponding to time. The difference between them is that the key of tags has an index, while the key of fields does not have an index, so in actual use, the fields that need to be frequently queried are set as tags, and fields are used to store continuous numerical values. In this embodiment, the time series database data storage format is that time is the data sampling time, measurement is set as the user number to indicate the ownership of the time series data table, tag is set as the sensor source device name to quickly find the data of the same device, and field is set as the value of all sensors of the device. When a device has multiple sensors, there will be multiple field key-value pairs at the same time point. In this embodiment, the smart bracelet is used as one of the cardiopulmonary sensing terminals. This device measures the heart rate and blood oxygen values at the same time, Figure 5The saving form of the data of the smart bracelet at the same time in the time series database is illustrated. It should be noted that this table is only a form for people to view data, and the table is not directly stored in the database, but a variety of compression strategies are used to reduce the data storage amount.

[0040] Further, the man-machine interaction module displays through the HDMI interface connected display, the interface is realized based on PyQt5, and is used for inputting user information, displaying the connection state of the cardiopulmonary sensing terminal, displaying cardiopulmonary function parameters and evaluation results. At the same time, a wireless Web UI display interface is provided to improve the portability and ease of use of the device. When the user's mobile phone or computer is in the same local area network as the digital base, the specified port address can be accessed to enter the Web UI, and the function of the Web UI interface is the same as that of the wired display interface.

[0041] Based on the above description of the implementation of each module, a use step of the digital base is given:

[0042] S1, determine the sensing terminal to be connected, and store the mac addresses of the cardiopulmonary sensing terminals in the NVS so that the multi-interface data receiving module can successfully connect and receive data.

[0043] S2, fill in the device type identification number, the position of the target data in the data packet, the target data length and the byte sequence and other information in the configuration file in the data preprocessing module so that the program can complete protocol analysis and target data extraction.

[0044] S3, under the data acquisition module, establish a data set and complete algorithm research, and then convert the original algorithm model into an RKNN model using RKNN Toolkit2, and deploy it to the edge computing inference module. This step can be skipped according to actual use.

[0045] S4, in the normal working mode, the user wears the related equipment, and can see the collected data curve and the cardiopulmonary function evaluation result on the man-machine interaction module.

[0046] Therefore, the multi-interface data receiving module and the data preprocessing module realized by the base solve the problems of data collection, analysis and fusion of heterogeneous terminals, and based on the collected data, the accuracy of the cardiopulmonary function evaluation algorithm can be effectively improved by using multi-source data fusion. On the other hand, the edge computing inference module realized by the base effectively shortens the analysis, fusion and evaluation time of multi-source cardiopulmonary data, and ensures the real-time performance of the algorithm.

[0047] The above is a preferred embodiment of the present application, but the present application should not be limited to the contents disclosed in the embodiment and the drawings, so any equivalent or modification made without departing from the spirit of the present application falls within the scope of protection of the present application.

Claims

1. A multi-interface edge computing digital dock for cardiopulmonary function assessment, characterized in that... include: The multi-interface data receiving module provides multiple wired and wireless interfaces for receiving data from the cardiopulmonary sensing terminal and environmental sensing data. The multi-interface data receiving module includes various wired and wireless interface modules. The wired interface modules include ADC, UART, USB, SPI, IIC, and CSI interfaces. The wireless interface modules include BLE, ZigBee, and WiFi interfaces. The BLE interface function is implemented based on Espressif ESP32-C3, the ZigBee interface function is implemented based on TICC2530, and the WiFi function is implemented based on Intel AX210. The BLE interface module operates in a master-slave mode, and there is a device connection limit for a single BLE interface module. The system increases the number of BLE devices that can be connected by implementing the function of accessing multiple BLE interface modules. The specific steps are as follows: Initially, the first BLE interface module scans and connects. When the number of connections reaches the connection limit, the digital dock is notified to enable the next BLE interface module to scan. The maximum number of connections varies between 8 and 20, depending on the BLE protocol stack used. The BLE interface module and ZigBee interface module simultaneously receive data from multiple sensing terminal devices. To distinguish the data source, the received data is encapsulated twice and then forwarded to the data preprocessing module through the same interface. The encapsulation format is as follows: a 2-byte frame header, which is 0xAA, 0xAA; a 1-byte data source identifier; a 1-byte data number; a 1-byte received data size identifier; and N bytes of the received original data, where N is the size of the received data. The data preprocessing module is used to perform preprocessing operations such as device protocol parsing, filtering, and feature extraction on the received data; The data preprocessing module is implemented based on the Rockchip RK3588 embedded Linux platform and can perform protocol parsing, filtering, and feature extraction on the received data. The protocol parsing process involves extracting the target data from the original data based on the data protocols of each sensing terminal. For secondary packet data transmitted from the wireless interface, the data source and original data are first unpacked, and then the target data is extracted according to the corresponding device protocol based on the data source. The module reserves a protocol setting interface. When a new type of device is added, the device type identifier, the position of the target data in the data packet, the length of the target data, and the byte order information are filled in the configuration file. The filtering process removes high-frequency noise, power frequency interference, baseline drift, and motion artifacts from the time-series data. The module reserves an interface for setting parameters related to the filtering algorithm. The feature extraction process extracts physiological feature data related to cardiopulmonary function from the data, including heart rate, heart rate variability, blood pressure, blood oxygen, oxygen uptake, and carbon dioxide excretion. An edge computing inference module is used to deploy an AI model for cardiopulmonary function assessment and perform real-time inference assessment based on preprocessed data. This module is implemented on the Rockchip RK3588 embedded Linux platform, which has a built-in convolutional processing unit. The original model of the algorithm to be deployed is converted into an RKNN model using RKNNToolkit2 and deployed to the platform, accelerating the algorithm's inference speed. The algorithm model is designed using a 1DCNN-LSTM model architecture. The model input consists of signal features obtained from the preprocessing module, including heart rate, heart rate variability, blood pressure, blood oxygen saturation, oxygen uptake, and carbon dioxide excretion. The data length is 30 seconds. Each type of data is processed by both 1DCNN and LSTM, and the LSTM outputs are concatenated before being input into a fully connected layer to obtain the cardiopulmonary function assessment results, including three categories: normal cardiopulmonary function, decreased cardiopulmonary function, and abnormal cardiopulmonary function. The data storage and management cloud service module is used to store the collected data and evaluation results in a cloud time-series database and provides data management and visualization UI interface; The human-computer interaction module is used to display user information input and cardiopulmonary function assessment results; There are two modes: data acquisition mode and normal operation mode. When entering data acquisition mode, the received raw data is directly stored after preliminary parsing. At this time, the data preprocessing module only performs protocol parsing, while the edge computing inference module is disabled. When entering normal operation mode, the data is acquired, preprocessed, and then processed by the edge computing inference module to obtain the cardiopulmonary function evaluation results. The preprocessed data and evaluation results are then stored.

2. The multi-interface edge computing digital base for cardiopulmonary function assessment as described in claim 1, characterized in that, The wireless interface module stores a list of MAC addresses of the cardiopulmonary sensing terminals that need to be connected. Once the MAC address of the scanned device is in the list, it will actively initiate a connection.

3. The multi-interface edge computing digital base for cardiopulmonary function assessment as described in claim 1, characterized in that, The data storage and management cloud service module uses the API interface provided by the influxdb-client-python package to store some important data and evaluation results in the InfluxDB time series database on the cloud server via WiFi, Ethernet or 4G / 5G mobile communication. At the same time, it provides a backend interface for the time series database so that professionals can query and retrieve data.

4. The multi-interface edge computing digital base for cardiopulmonary function assessment as described in claim 1, characterized in that, The human-computer interaction module connects to a monitor via an HDMI interface. The interface is implemented using PyQt5 and is used to input user information, display the connection status of the cardiopulmonary sensing terminal, display cardiopulmonary function parameters and evaluation results. At the same time, a wireless WebUI display interface is provided to improve the portability and ease of use of the device. When the user's mobile phone or computer is on the same local area network as the digital dock, it can access the WebUI through a specified port address. The WebUI interface has the same functions as the wired display interface.

Images