Patient real-time data automatic acquisition method based on Internet of Things

Abstract

The invention discloses a patient real-time data automatic acquisition method based on the Internet of Things, and relates to the technical field of Internet of Things technology and medical health data analysis, and the method comprises the steps: employing multiple sensors to detect the physiological data of a patient; transmitting the physiological data to an edge computing unit, preprocessing the physiological data, and analyzing the preprocessed physiological data; judging whether the physiological data of the patient is abnormal or not according to a comprehensive abnormal score obtained according to an analysis result, and marking; and storing the preprocessed physiological data and the abnormal result in a cloud database, and establishing an electronic health record of the patient. According to the method, the vital signs of the patient are detected by adopting multiple sensors, and the edge calculation and encryption technology is combined, so that real-time automatic acquisition and anomaly detection of the physiological data of the patient are realized, the data security is guaranteed, the real-time performance and accuracy of data processing are greatly improved, and the data processing efficiency is improved. And more comprehensive and accurate health monitoring is provided for the patient.

Description

Technical Field

[0001] The present invention relates to the field of Internet of Things technology and medical health data analysis technology, and in particular to a method for automatically collecting real-time patient data based on the Internet of Things. Background Art

[0002] With the rapid development of Internet of Things (IoT) technology, the field of medical health is gradually moving towards intelligence and automation. In recent years, the use of IoT technology to monitor patients' physiological data in real time, especially in patient health monitoring and disease prevention, has made significant progress. As one of the key technologies in the Internet of Things, multi-sensor technology can obtain patients' vital signs in real time in multiple dimensions, such as heart rate, blood pressure, body temperature, blood oxygen and other physiological data. These data are collected and transmitted to the cloud platform through sensors, providing doctors with real-time and accurate information on patients' health status, which helps to detect changes in patients' conditions in a timely manner and improve the efficiency of diagnosis and treatment. Especially in telemedicine and intelligent health management systems, real-time data collection and analysis technology based on the Internet of Things has become an indispensable part.

[0003] However, the existing patient data collection technology based on the Internet of Things still has some shortcomings. First, the traditional physiological data collection method often relies on cloud servers for data storage and processing, resulting in high data transmission delays, especially in environments with limited bandwidth, which may not reflect the patient's health status in real time. Secondly, in the existing technology, the analysis of physiological data is often concentrated in the cloud, lacking sufficient edge computing support. As an emerging technology, edge computing can perform data processing on computing devices near the data source, reduce the burden of data transmission, and improve the response speed, but its application in the automatic collection and analysis of medical and health data is not yet widespread. In addition, many existing systems have relatively simple processing methods for abnormal data detection and labeling, often relying on traditional threshold settings, and fail to fully utilize the comprehensive analysis of multidimensional data, resulting in low diagnostic accuracy and failure to timely detect potential health risks. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention proposes a method for automatically collecting real-time patient data based on the Internet of Things.

[0005] Therefore, the present invention provides a method for automatically collecting real-time patient data based on the Internet of Things, which can solve the problems mentioned in the background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for automatically collecting real-time patient data based on the Internet of Things, which comprises: A multi-sensor is used to detect the patient's physiological data; the multi-sensor includes a first physiological sensor, a second physiological sensor, a third physiological sensor, a fourth physiological sensor and a fifth physiological sensor, the first physiological sensor collects the patient's electrocardiogram data, the second physiological sensor collects the patient's blood pressure data, the third physiological sensor collects the patient's blood oxygen saturation data and pulse rate data, the fourth physiological sensor collects the patient's body temperature data, and the fifth physiological sensor collects the patient's respiratory rate data; Transmitting the physiological data to an edge computing unit, preprocessing the physiological data, and building a physiological parameter association model based on the preprocessed physiological data; Determining whether the patient's physiological data is abnormal according to the comprehensive abnormality score output by the physiological parameter association model, and marking the abnormal data to obtain an abnormal result; The pre-processed physiological data and abnormal results are stored in a cloud database, and an electronic health record of the patient is established.

[0007] As a preferred solution of the method for automatic real-time data collection of patients based on the Internet of Things of the present invention, wherein: the first physiological sensor is installed on the first fixing belt through a buckle structure; the first fixing belt is set at the middle clavicle line position of the fourth intercostal space on the left side of the patient's chest, and the inner surface thereof is coated with a conductive gel layer; the collection electrode of the first physiological sensor is in close contact with the conductive gel layer to collect the electrocardiogram data of the patient; The second physiological sensor is installed in the pressure cavity of the second fixing belt; the second fixing belt is set at the radial artery pulsation position of the patient's left wrist, and adopts an adjustable strap structure; the second physiological sensor measures the blood pressure data of the patient by means of air pump pressurization; The third physiological sensor is arranged at the end of the index finger of the patient's right hand to obtain the patient's blood oxygen saturation data and pulse rate data; The fourth physiological sensor is arranged at the patient's armpit to collect the patient's body temperature data; The fifth physiological sensor is arranged at the chest position of the patient, and detects chest vibration through a piezoelectric sensor element to collect respiratory frequency data.

[0008] As a preferred solution of the method for automatic collection of real-time patient data based on the Internet of Things of the present invention, the preprocessing of the physiological data includes: Filtering the ECG data in the received physiological data by the digital signal processor of the edge computing unit to obtain filtered ECG data; Performing QRS wave group detection on the filtered ECG data; The peak position of the R wave was determined by the threshold comparison method, and the RR interval between adjacent R waves was calculated to obtain the heart rate variability index; Using a median filter method to eliminate pulse interference from the blood pressure data and correct missing data points; The blood oxygen saturation data is smoothed using a moving average filtering method; The body temperature data is also filtered.

[0009] As a preferred solution of the method for automatic collection of real-time patient data based on the Internet of Things of the present invention, the electrocardiogram data in the physiological data is filtered as shown in the following formula: ;in, is the gain coefficient, is the damping ratio, is the natural frequency, is the transfer function of the bandpass filter, is the Laplace transform complex variable; The QRS wave group detection of the filtered ECG data is performed using a wavelet transform method, wherein the calculation of the wavelet transform coefficient is shown in the following formula: ;in, is the scale parameter, is the translation parameter, is the wavelet mother function, is the input ECG signal, is the time variable, is the wavelet transform coefficient; The calculation of the heart rate variability index is shown in the following formula: ;in, , , is the weight coefficient, is the standard deviation of the RR interval, is the root mean square of the difference between adjacent RR intervals, is the percentage of adjacent RR intervals with a difference greater than 50ms, An indicator of heart rate variability.

[0010] As a preferred solution of the method for automatic collection of real-time patient data based on the Internet of Things of the present invention, the blood pressure data is subjected to a median filtering method to eliminate pulse interference as shown in the following formula: ;in, , , is the adaptive weight coefficient, Represents the blood pressure value of the previous sampling point of the current sampling point in the discrete sequence, Represents the blood pressure value at the current sampling point in the discrete sequence, Indicates the blood pressure value of the next sampling point after the current sampling point in the discrete sequence. Output value of the current point after filtering; The missing data points are corrected by linear interpolation algorithm, as shown in the following formula: in, is a smooth function: in, is the time constant, is the nearest known blood pressure value to the left of the missing interval, is the nearest known blood pressure value to the right of the missing interval, is the current time point, For the next time point, are the corrected missing data points.

[0011] As a preferred solution of the method for automatic real-time data collection of patients based on the Internet of Things of the present invention, wherein: the physiological parameter association model includes a five-layer deep neural network structure, wherein the number of nodes in the input layer is the sum of the number of sampling points of all physiological parameters, the hidden layer adopts the ReLU activation function, and the output layer gives the abnormality degree score of the physiological parameters; The calculation of the comprehensive abnormality score is shown in the following formula: ;in, is the neural network weight matrix, is the bias vector, For the Real-time physiological parameter values, For the Reference values ​​of physiological parameters, is the standard deviation, is a comprehensive index, is the comprehensive abnormality score.

[0012] As a preferred solution of the method for automatic real-time data collection of patients based on the Internet of Things of the present invention, judging whether the physiological data of the patient is abnormal according to the comprehensive abnormality score obtained from the analysis result includes: When the abnormality score of the preprocessed physiological data exceeds the range of the preset score threshold, the edge computing unit marks the abnormal data, records the time of abnormal occurrence, abnormal parameter type, abnormal value and corresponding abnormal score, and marks it as a first-level abnormal record; when the abnormality score of the preprocessed physiological data exceeds the range of the preset score threshold for three consecutive times, the edge computing unit records the duration of the abnormality, the continuous abnormal value and the cumulative abnormal score, and marks it as a second-level abnormal record.

[0013] In a second aspect, the present invention provides a patient real-time data automatic collection system based on the Internet of Things, which comprises: A data acquisition module, used for detecting the patient's physiological data using multiple sensors; A data transmission module, used to transmit the physiological data to an edge computing unit; an edge computing unit, configured to preprocess the physiological data, construct a physiological parameter association model based on the preprocessed physiological data, determine whether the physiological data of the patient is abnormal according to a comprehensive abnormality score output by the physiological parameter association model, and mark the abnormal data to obtain an abnormal result; The recording module is used to store the pre-processed physiological data and abnormal results in a cloud database according to the data transmission module, and to establish an electronic health record of the patient.

[0014] The beneficial effects of the present invention are: using multiple sensors to detect the patient's vital signs, and combining edge computing and encryption technology to achieve real-time automatic collection and abnormality detection of the patient's physiological data. Specifically, the scheme comprehensively monitors the patient's health status by setting sensors in different parts, including electrocardiogram sensors, blood pressure sensors, blood oxygen sensors, body temperature sensors and respiratory rate sensors. These sensors transmit data to the edge computing unit through low-power Bluetooth technology for data encryption, preprocessing and abnormality analysis. During the data transmission process, the SM2 elliptic curve public key cryptography algorithm and the SM4 block cipher algorithm are used to ensure data security. At the same time, the edge computing unit improves the accuracy and efficiency of data processing through methods such as multi-layer filtering, interpolation processing and wavelet transform, and uses a neural network model to perform comprehensive abnormality scoring on the data, thereby achieving efficient and accurate abnormality detection. Finally, all processed data and abnormal records are stored in the cloud database to establish the patient's electronic health record, which provides an important basis for subsequent medical decision-making and health management. While ensuring data security, this scheme greatly improves the real-time and accuracy of data processing, providing patients with more comprehensive and accurate health monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them: Figure 1 The figure is a flow chart of a method for automatically collecting real-time patient data based on the Internet of Things. DETAILED DESCRIPTION

[0016] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0018] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0019] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0020] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0021] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0022] Example 1: Reference Figure 1 , which is the first embodiment of the present invention, provides a method for automatically collecting real-time patient data based on the Internet of Things, including: S1: Use multiple sensors to detect the patient's physiological data; A first fixing belt is arranged at the middle clavicle line position of the fourth intercostal space on the left side of the patient's chest, the inner surface of the first fixing belt is coated with a conductive gel layer, the first fixing belt is fixed by a Velcro, the first physiological sensor is installed on the first fixing belt by a snap-on structure, the collection electrode of the first physiological sensor is in close contact with the conductive gel layer, the first physiological sensor has a built-in ECG signal amplification circuit and a digital conversion circuit, and the first physiological sensor collects ECG data with a sampling frequency of 250Hz.

[0023] A second fixing belt is set at the radial artery pulsation position of the patient's left wrist. The second fixing belt adopts an adjustable strap structure. The second physiological sensor is installed in the pressure cavity of the second fixing belt. The second physiological sensor measures the systolic pressure and diastolic pressure by air pump pressurization. The second physiological sensor collects blood pressure data every 15 minutes. The blood pressure data includes systolic pressure value, diastolic pressure value and measurement timestamp information.

[0024] A third physiological sensor is set at the end of the patient's right index finger. The third physiological sensor adopts the photoelectric volumetric pulse wave measurement principle to obtain the patient's blood oxygen saturation data and pulse rate data by emitting near-infrared light with a wavelength of 850nm and detecting changes in the intensity of reflected light. The sampling frequency of the blood oxygen saturation data and the pulse rate data is 60Hz.

[0025] A fourth physiological sensor is arranged at the patient's armpit. The fourth physiological sensor adopts a platinum resistance temperature measuring element. The measurement accuracy of the fourth physiological sensor is ±0.1° C., and body temperature data is collected every 5 minutes.

[0026] A fifth physiological sensor is arranged at the chest position of the patient. The fifth physiological sensor detects chest vibration through a piezoelectric sensor element and collects respiratory frequency data. The sampling frequency of the respiratory frequency data is 20 Hz.

[0027] S2: Based on a wireless communication protocol, transmitting the electrocardiogram data and the blood pressure data to an edge computing unit, wherein the edge computing unit includes a preset data abnormality threshold range; The first physiological sensor, the second physiological sensor, the third physiological sensor, the fourth physiological sensor and the fifth physiological sensor are respectively provided with a low-power Bluetooth communication module, and the communication frequency band of the low-power Bluetooth communication module is 2.4 GHz; the low-power Bluetooth communication module has a built-in SM2 elliptic curve public key cryptographic algorithm chip, and the SM2 elliptic curve public key cryptographic algorithm chip stores a device unique identification code and a key pair.

[0028] In the data encryption process at the sensor end, the ECG data, the blood pressure data, the blood oxygen saturation data, the pulse rate data, the temperature data and the respiratory rate data are grouped and timestamped by the unique device identification code to form a data packet to be encrypted. The SM2 elliptic curve public key cryptographic algorithm chip is used to generate a session key, the length of which is 256 bits; the SM3 cryptographic hash algorithm is used to calculate the data packet to be encrypted to obtain a 32-byte message digest; the message digest is combined with the data packet to be encrypted and encrypted by the SM4 block cipher algorithm using the session key to obtain an encrypted data stream.

[0029] The edge computing unit is set within 2 meters of the patient's bed. The edge computing unit has a built-in security chip that supports the national secret algorithm, and the security chip pre-stores the key information of the device's unique identification code; in the decryption process on the edge computing unit end, the identity of the sender of the encrypted data stream is first verified through the SM2 elliptic curve public key cryptographic algorithm chip; the encrypted data stream is decrypted using the SM4 block cipher algorithm using the session key to separate the message digest and the data packet to be encrypted; the message digest of the decrypted data packet to be encrypted is recalculated using the SM3 cryptographic hash algorithm, and compared with the received message digest to verify data integrity.

[0030] A data abnormality threshold range table is pre-stored in the memory of the edge computing unit, and the data abnormality threshold range table includes an ECG ST segment deviation threshold interval of -0.2mV to +0.2mV, a systolic pressure threshold interval of 90mmHg to 140mmHg, a diastolic pressure threshold interval of 60mmHg to 90mmHg, a blood oxygen saturation threshold lower limit of 94%, a pulse rate threshold interval of 60 times / minute to 100 times / minute, a body temperature threshold interval of 36.3℃ to 37.2℃, and a respiratory rate threshold interval of 12 times / minute to 20 times / minute.

[0031] S3: In the edge computing unit, preprocessing the physiological data, and analyzing the preprocessed physiological data; The digital signal processor of the edge computing unit eliminates power frequency interference from the received ECG data through a 50Hz notch filter, and then eliminates baseline drift and high-frequency noise through a 0.05Hz-100Hz bandpass filter to obtain filtered ECG data; the transfer function of the bandpass filter is for: ;in, is the gain coefficient, is the damping ratio, is the natural frequency, is the transfer function of the bandpass filter, is the Laplace transform of complex variables.

[0032] After obtaining the filtered ECG data, the filtered ECG data is subjected to QRS wave group detection using a wavelet transform method, wherein the wavelet transform coefficients The calculation formula is: ;in, is the scale parameter, is the translation parameter, is the wavelet mother function, is the input ECG signal, is the time variable, is the wavelet transform coefficient.

[0033] Furthermore, the peak position of the R wave was determined by the dynamic threshold comparison method. The calculation formula is: ;in, is the smoothing coefficient, the value range is [0.8,0.95], is the dynamic threshold, is the threshold value at the previous moment, is the maximum value of the wavelet coefficient in the current window; Then, the RR interval between adjacent R waves is calculated based on the determined R wave peak position to obtain the heart rate variability index. : ;in, , , is the weight coefficient, is the standard deviation of the RR interval, is the root mean square of the difference between adjacent RR intervals, It is the percentage of adjacent RR intervals with a difference greater than 50ms.

[0034] Furthermore, the blood pressure data is subjected to an improved three-point median filtering method to eliminate pulse interference, and the calculation formula is: ;in, , , is the adaptive weight coefficient, satisfying , Represents the blood pressure value of the previous sampling point of the current sampling point in the discrete sequence, Represents the blood pressure value at the current sampling point in the discrete sequence, Indicates the blood pressure value of the next sampling point after the current sampling point in the discrete sequence. Output value of the current point after filtering.

[0035] The missing data points are corrected by an improved linear interpolation algorithm. for: ;in, is the time constant, is the nearest known blood pressure value to the left of the missing interval, is the nearest known blood pressure value to the right of the missing interval, is the current time point, For the next time point, is the corrected missing data point, is a smooth function: ; The digital signal processor smoothes the blood oxygen saturation data using a moving average filtering method, and the sliding window length is set to 5 sampling points. The digital signal processor processes the respiratory rate data using a Butterworth low-pass filter, and the cutoff frequency is set to 2 Hz.

[0036] Furthermore, the neural network processor of the edge computing unit loads a pre-trained physiological parameter association model, which includes a five-layer deep neural network structure, the number of input layer nodes is the sum of all physiological parameter sampling points, the hidden layer uses a ReLU activation function, and the output layer gives a physiological parameter abnormality score. The pre-processed electrocardiogram data, blood pressure data, blood oxygen saturation data, body temperature data, and respiratory rate data are input into the physiological parameter association model to obtain a comprehensive abnormality score, as shown in the following formula: ;in, is the comprehensive abnormality score, is the neural network weight matrix, is the bias vector, For the Real-time physiological parameter values, For the Reference values ​​of physiological parameters, is the standard deviation, It is a comprehensive index used to evaluate data reliability and is calculated by the following formula: ;in, to is the dynamic weight coefficient, which is calculated in real time by the deep neural network. is the blood pressure change rate, is the blood oxygen saturation, For body temperature, is the respiratory rate; Reference values ​​to be explained The thresholds of various physiological characteristics are preset in the edge computing module.

[0037] S4: The edge computing unit calculates the abnormality score of the preprocessed physiological data and records it in a graded manner: When the abnormality score of the preprocessed physiological data exceeds the threshold range of the preset score threshold, the edge computing unit marks the abnormal data, records the abnormality occurrence time, abnormal parameter type, abnormal value and corresponding abnormal score, and marks it as a first-level abnormal record; when the abnormality score of the preprocessed physiological data exceeds the preset score threshold range for three consecutive times, the edge computing unit records the duration of the abnormality, the continuous abnormal value and the cumulative abnormal score, and marks it as a second-level abnormal record.

[0038] The pre-processed physiological data and abnormal results are stored in a cloud database, and an electronic health record of the patient is established.

[0039] Furthermore, this embodiment also provides a patient real-time data automatic collection system based on the Internet of Things, including: A data acquisition module, used for detecting the patient's physiological data using multiple sensors; A data transmission module, used to transmit the physiological data to an edge computing unit; an edge computing unit, configured to preprocess the physiological data, construct a physiological parameter association model based on the preprocessed physiological data, determine whether the physiological data of the patient is abnormal according to a comprehensive abnormality score output by the physiological parameter association model, and mark the abnormal data to obtain an abnormal result; The recording module is used to store the pre-processed physiological data and abnormal results in a cloud database according to the data transmission module, and to establish an electronic health record of the patient.

[0040] This embodiment also provides a computer device, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method for automatic collection of real-time patient data based on the Internet of Things as proposed in the above embodiment.

[0041] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0042] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for automatically collecting real-time patient data based on the Internet of Things as proposed in the above embodiment is implemented.

[0043] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0044] Example 2: This is the second example of the present invention. This example provides a method for automatically collecting real-time patient data based on the Internet of Things. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0045] A 30-day system test was conducted in the cardiovascular ward of a tertiary hospital, and 60 patients who needed continuous monitoring were selected as test subjects. The test subjects included 20 patients with acute myocardial infarction, 20 patients with heart failure, and 20 patients with arrhythmia, with an age range of 45-75 years old. The patients were randomly divided into a test group (30 people) and a control group (30 people). The test group used the multi-sensor monitoring system of the present invention, and the control group used a traditional monitor for monitoring.

[0046] Patients in the experimental group wore five physiological sensors according to the method described in S1 in Example 1. After the first physiological sensor was fixed at the specified position on the chest, the lead position and signal quality were checked by special software to ensure that the ECG signal was clear and discernible. The strap pressure of the second physiological sensor was calibrated to ensure that the measurement accuracy was within ±3 mmHg. The photoelectric element of the third physiological sensor was calibrated to ensure that the measurement error of blood oxygen saturation was within ±2%. The fourth physiological sensor was temperature calibrated before use to ensure measurement accuracy. The sensitivity of the fifth physiological sensor was adjusted to ensure that the chest vibration signal could be accurately captured.

[0047] All sensor data is transmitted to the edge computing unit via the low-power Bluetooth 4.0 protocol, and the transmission process is encrypted using the SM2 / SM3 / SM4 national encryption algorithm. The edge computing unit uses an ARM Cortex-M4 processor with a main frequency of 120MHz, built-in 2MB Flash memory and 256KB RAM. Texas Instruments' TMS320F28335DSP chip is used for digital signal processing, and Intel Movidius neural computing stick is used for neural network processing.

[0048] In the signal preprocessing stage, the ECG signal is processed by an improved adaptive filtering algorithm, which has a better baseline drift suppression effect than the traditional 50Hz notch filter. The three-point median filter of blood pressure data adopts a dynamic weight coefficient to improve the anti-interference ability. After preprocessing, various physiological parameters are input into the pre-trained deep neural network model. The model adopts a five-layer structure, including two convolutional layers and three fully connected layers. It is trained using the Adam optimizer and has an accuracy of 95.8% on the validation set.

[0049] The system was run continuously for 30 days, and the collected monitoring data are shown in Table 1 below: Table 1 Comparison of monitoring indicators Patient ID ECG abnormality recognition accuracy (%) Blood pressure fluctuation warning lead time (min) Blood oxygen saturation error (%) Accuracy of warning of abnormal vital signs (%) Missing reporting rate (%) False alarm rate (%) Medical response time (min) Test Group 01 98.5 12.3 0.8 96.7 1.2 2.1 3.2 Test Group 02 97.8 11.8 1.1 95.9 1.5 2.3 3.5 Test Group 03 98.2 13.1 0.9 97.2 1.1 1.9 2.8 Control group 01 92.1 5.2 1.8 89.5 3.8 5.6 6.7 Control group 02 91.8 4.8 2.1 88.9 4.2 5.9 7.1 Control group 03 92.4 5.5 1.9 90.1 3.5 5.4 6.4 By comparing the test data, the following analysis conclusions can be drawn: The average accuracy rate of ECG abnormality recognition in the experimental group reached 98.17%, 6.1 percentage points higher than that in the control group. This is mainly due to the improved adaptive filtering algorithm and wavelet transform QRS complex detection method, which significantly improved the quality of ECG signals and the accuracy of feature extraction.

[0050] The experimental group's blood pressure fluctuation warning advance time can detect abnormal blood pressure trends 12.4 minutes in advance on average, while the control group's is only 5.2 minutes. This early warning capability is due to the improved three-point median filtering and dynamic threshold detection algorithm adopted by the system, which buys more intervention time for medical staff.

[0051] The average error of blood oxygen saturation in the test group was controlled within 0.93%, which was better than the 1.93% in the control group. This was due to the 850nm wavelength near-infrared light source and optimized signal processing algorithm used in the system.

[0052] The average accuracy rate of abnormal vital signs warning in the experimental group reached 96.6%, 7.4 percentage points higher than that in the control group. The system uses a deep neural network model to achieve multi-parameter fusion analysis, which greatly improves the accuracy of anomaly detection.

[0053] The average false negative rate of the test group was 1.27%, and the false positive rate was 2.1%, which was significantly better than the 3.83% and 5.63% of the control group, indicating that the system has high reliability and stability.

[0054] The average medical response time in the experimental group was 3.17 minutes, 3.63 minutes shorter than that in the control group. This improvement was due to the system's timely early warning and graded alarm mechanism, which enabled medical staff to identify and respond to potential risks more quickly.

[0055] These data fully prove that the multi-sensor vital signs monitoring system has significant advantages in clinical applications. It not only improves monitoring accuracy and early warning timeliness, but also effectively reduces the workload of medical staff and provides more reliable protection for patient safety. In particular, the system's multi-parameter fusion analysis and early warning mechanism show obvious innovation and practical value.

[0056] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for automatically collecting real-time patient data based on the Internet of Things, characterized in that: include: Use multiple sensors to detect the patient's physiological data; The multi-sensor includes a first physiological sensor, a second physiological sensor, a third physiological sensor, a fourth physiological sensor and a fifth physiological sensor, the first physiological sensor collects the patient's electrocardiogram data, the second physiological sensor collects the patient's blood pressure data, the third physiological sensor collects the patient's blood oxygen saturation data and pulse rate data, the fourth physiological sensor collects the patient's body temperature data, and the fifth physiological sensor collects the patient's respiratory rate data; Transmitting the physiological data to an edge computing unit, preprocessing the physiological data, and building a physiological parameter association model based on the preprocessed physiological data; Determining whether the patient's physiological data is abnormal according to the comprehensive abnormality score output by the physiological parameter association model, and marking the abnormal data to obtain an abnormal result; The pre-processed physiological data and abnormal results are stored in a cloud database, and an electronic health record of the patient is established.

2. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 1, characterized in that: The first physiological sensor is mounted on the first fixing belt through a buckle structure; the first fixing belt is arranged at the middle clavicle line of the fourth intercostal space on the left side of the patient's chest, and the inner surface thereof is coated with a conductive gel layer; the collection electrode of the first physiological sensor is in close contact with the conductive gel layer to collect the electrocardiogram data of the patient; The second physiological sensor is installed in the pressure cavity of the second fixing belt; the second fixing belt is set at the radial artery pulsation position of the patient's left wrist, and adopts an adjustable strap structure; the second physiological sensor measures the blood pressure data of the patient by means of air pump pressurization; The third physiological sensor is arranged at the end of the index finger of the patient's right hand to obtain the patient's blood oxygen saturation data and pulse rate data; The fourth physiological sensor is arranged at the patient's armpit to collect the patient's body temperature data; The fifth physiological sensor is arranged at the chest position of the patient, and detects chest vibration through a piezoelectric sensor element to collect respiratory frequency data.

3. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 2, characterized in that: Preprocessing the physiological data includes: Filtering the ECG data in the received physiological data by the digital signal processor of the edge computing unit to obtain filtered ECG data; Performing QRS wave group detection on the filtered ECG data; The peak position of the R wave was determined by the threshold comparison method, and the RR interval between adjacent R waves was calculated to obtain the heart rate variability index; Using a median filter method to eliminate pulse interference from the blood pressure data and correct missing data points; The blood oxygen saturation data is smoothed using a moving average filtering method; The body temperature data is filtered.

4. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 3, characterized in that: The electrocardiogram data in the physiological data is filtered as shown in the following formula: ;in, is the gain coefficient, is the damping ratio, is the natural frequency, is the transfer function of the bandpass filter, is the Laplace transform complex variable; The QRS wave group detection of the filtered ECG data is performed using a wavelet transform method, wherein the calculation of the wavelet transform coefficient is shown in the following formula: ;in, is the scale parameter, is the translation parameter, is the wavelet mother function, is the input ECG signal, is the time variable, is the wavelet transform coefficient; The calculation of the heart rate variability index is shown in the following formula: ;in, , , is the weight coefficient, is the standard deviation of the RR interval, is the root mean square of the difference between adjacent RR intervals, is the percentage of adjacent RR intervals with a difference greater than 50ms, An indicator of heart rate variability.

5. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 4, characterized in that: The median filtering method is used to eliminate pulse interference on the blood pressure data as shown in the following formula: ;in, , , is the adaptive weight coefficient, Represents the blood pressure value of the previous sampling point of the current sampling point in the discrete sequence, Represents the blood pressure value at the current sampling point in the discrete sequence, Indicates the blood pressure value of the next sampling point after the current sampling point in the discrete sequence. Output value of the current point after filtering; The missing data points are corrected by linear interpolation algorithm, as shown in the following formula: in, is a smooth function: in, is the time constant, is the nearest known blood pressure value to the left of the missing interval, is the nearest known blood pressure value to the right of the missing interval, is the current time point, For the next time point, are the corrected missing data points.

6. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 5, characterized in that: The physiological parameter association model includes a five-layer deep neural network structure, wherein the number of nodes in the input layer is the sum of the number of sampling points of all physiological parameters, the hidden layer uses the ReLU activation function, and the output layer gives the abnormality score of the physiological parameters; The calculation of the comprehensive abnormality score is shown in the following formula: ;in, is the neural network weight matrix, is the bias vector, For the Real-time physiological parameter values, For the Reference values ​​of physiological parameters, is the standard deviation, is a comprehensive index, is the comprehensive abnormality score.

7. The method for automatically collecting real-time patient data based on the Internet of Things as claimed in claim 6, characterized in that: Judging whether the patient's physiological data is abnormal based on the comprehensive abnormality score obtained from the analysis results includes: When the abnormality score of the preprocessed physiological data exceeds the range of the preset score threshold, the edge computing unit marks the abnormal data, records the time of abnormal occurrence, abnormal parameter type, abnormal value and corresponding abnormal score, and marks it as a first-level abnormal record; when the abnormality score of the preprocessed physiological data exceeds the range of the preset score threshold for three consecutive times, the edge computing unit records the duration of the abnormality, the continuous abnormal value and the cumulative abnormal score, and marks it as a second-level abnormal record.

8. A system for automatically collecting patient real-time data based on the Internet of Things, based on the method for automatically collecting patient real-time data based on the Internet of Things according to any one of claims 1 to 7, characterized in that: include: A data acquisition module, used for detecting the patient's physiological data using multiple sensors; A data transmission module, used to transmit the physiological data to an edge computing unit; an edge computing unit, configured to preprocess the physiological data, construct a physiological parameter association model based on the preprocessed physiological data, determine whether the physiological data of the patient is abnormal according to a comprehensive abnormality score output by the physiological parameter association model, and mark the abnormal data to obtain an abnormal result; The recording module is used to store the pre-processed physiological data and abnormal results in a cloud database according to the data transmission module, and to establish an electronic health record of the patient.

Images