Intelligent ward platform based on Internet of Things

Abstract

The invention discloses a smart ward platform based on the Internet of Things, and relates to the technical field of smart wards. A bedside interaction unit, an infusion unit, a nursing electronic whiteboard unit and a ward display screen are arranged in the ward; a patient wears an electrocardiogram monitoring bracelet, and the electrocardiogram monitoring bracelet comprises a millimeter wave blood flow monitoring module which is used for collecting physical sign data and electrocardiogram monitoring data; the physical sign fusion monitoring module is used for collecting real-time heart rate variation coefficient data through the electrocardiogram monitoring bracelet; the infusion optimization system comprises a piezoelectric film sensing array arranged in an infusion device, the identity authentication unit is integrated with a near-infrared structured light projection module and is used for comparing face 3D depth information with HIS system data; and the multi-data coordination unit is used for recommending an execution scheme based on the medical record similarity map and generating a conflict report and a multi-dimensional data comparison thermodynamic diagram. Individualized precise regulation and control of the infusion rate are achieved, the defects of traditional infusion management are overcome, and the safety and effectiveness of the infusion link are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart wards, and in particular to a smart ward platform based on the Internet of Things. Background Art

[0002] As the medical industry continues to evolve, traditional ward management models and equipment are no longer able to meet the demands of modern healthcare. Numerous problems have emerged throughout the patient care process. Patients' access to medical information is severely limited, communication channels between medical staff and patients are poor, infusion safety monitoring suffers from significant vulnerabilities, nursing work is cumbersome and inefficient, patient health education is limited, and vital sign data collection is not only untimely but also inaccurate. These issues significantly impact the patient experience and treatment outcomes, while also placing significant workload pressure on medical staff. The rise of IoT technology offers a new opportunity to address these issues. By incorporating IoT technology into smart ward development, it enables interconnected device connectivity within the ward, real-time data collection and sharing, and thus improves the quality and efficiency of medical services. However, current smart ward solutions on the market still lack the depth and breadth of IoT application, failing to fully leverage the advantages of IoT technology to achieve comprehensive intelligent management in smart wards. Therefore, there is an urgent need for innovative solutions that deeply integrate IoT technology to achieve intelligent ward management. Summary of the Invention

[0003] In response to the shortcomings of the above-mentioned existing technologies, the present invention provides an Internet of Things-based smart ward platform, which focuses on optimizing the infusion process, realizing personalized and precise control of the infusion rate, making up for the defects of traditional infusion management, and improving the safety and effectiveness of the infusion process, thereby comprehensively improving the quality and efficiency of medical services, improving patients' medical experience, and reducing the work pressure of medical staff.

[0004] The specific technical solutions are as follows:

[0005] The IoT-based smart ward platform features a bedside interactive unit, infusion unit, nursing electronic whiteboard unit, and ward display screen. Patients wear an ECG monitoring bracelet with a millimeter-wave blood flow monitoring module to collect vital sign and ECG data. It also includes:

[0006] The vital signs fusion monitoring module collects real-time heart rate variability data through an ECG monitoring bracelet, and integrates and analyzes it with chest micro-motion signals collected by a millimeter-wave non-contact radar installed next to the bed to establish a respiratory-ECG synchronization model. When the desynchronization between respiratory and ECG rhythms lasts for more than 5 minutes, the respiratory induction animation on the display screen is triggered and the respiratory rate is adjusted;

[0007] The infusion optimization system, including a piezoelectric film sensor array built into the infusion device, calculates a personalized infusion rate threshold based on the patient's weight and the vascular elasticity coefficient measured by the millimeter-wave blood flow monitoring module. If the infusion rate threshold deviates by more than ±15%, the bedside interactive unit will be linked to issue an audible and visual alarm.

[0008] The identity authentication unit integrates a near-infrared structured light projection module to achieve contactless medical identity verification by comparing facial 3D depth information with HIS system data;

[0009] The multi-data coordination unit recommends execution plans based on the medical record similarity map, and generates conflict reports and multi-dimensional data comparison heat maps.

[0010] Furthermore, the above scheme transmits HRV analysis data to the infusion optimization system in real time via the SPI bus. When respiratory-ECG desynchronization is detected, the infusion rate recalculation protocol is triggered. The multi-data coordination unit monitors the deviation between the vital sign data and the HIS doctor's order. When the deviation exceeds the preset threshold, a lock instruction is written to the infusion optimization system via the CAN_FD protocol, and the OpenCV library is called to generate a three-dimensional heat map for data comparison. When the infusion optimization system verifies an abnormality, the correction parameters are fed back to the vital sign fusion monitoring module and updated synchronously to the abnormal event log area.

[0011] Furthermore, in the above scheme, the vital sign fusion monitoring module starts the infusion recalculation within 20ms after detecting that the RR interval standard deviation is greater than 50ms; and the multi-data coordination unit completes the comparison and analysis of 50 groups of data points within 200ms.

[0012] Furthermore, in the above solution, the infusion rate threshold calculation formula of the infusion optimization system is:

[0013] Vascular elasticity = (α×BMI+β×C_v)×S_t,

[0014] Among them, α is the drug solution characteristic coefficient, β is the vascular compliance correction factor, and S_t is the average infusion adaptability score in the past 3 days. The clinical data of each coefficient are obtained through the hospital PACS system.

[0015] Furthermore, the above solution generates a dynamic waveform according to the patient's real-time breathing, and the waveform parameters are adjusted synchronously with the target breathing rhythm, and the adjustment effect is fed back in real time through the color gradient of the light band at the edge of the display screen.

[0016] Furthermore, the dynamic waveform of the breathing-induced animation is generated using a spline interpolation algorithm, with the amplitude parameter updated every 50 ms. The target breathing rhythm is dynamically adjusted using the following formula: f_t = f_0 + k·(ΔHRV / 10), where f_0 is the basic breathing guidance frequency, k is the adaptive adjustment coefficient, and ΔHRV is the change in the heart rate variability coefficient; Among them, RMSSD t is the root mean square difference of adjacent RR intervals in real time, RMSSD baseline It is the average value of the patient's resting state in the past 3 days.

[0017] Furthermore, in the above solution, the implementation of the spline interpolation algorithm includes the following parameter constraints:

[0018] (a) Using a cubic uniform B-spline curve 1, the node vectors are set to {u_0, u_1, ..., u_m}, where m = 2k + 3 (k is the number of characteristic points of the respiratory waveform)

[0019] (b) The control point matrix Q_i satisfies: the number of real-time respiratory signal sampling points received every 50ms n∈[15,25], and the deviation threshold δ≤1.2mm;

[0020] (c) Construct collision detection function: Φ(t) = ∑|P_i(t)-Q_i|<0.8σ_max

[0021] Where P_i(t) is the interpolation trajectory point calculated in real time, and σ_max is the maximum chest displacement monitored by the radar;

[0022] (d) The interpolation optimization module is integrated into the FPU unit of the STM32F4 series processor, and it takes ≤30ms to complete a single interpolation.

[0023] Furthermore, in the above solution, the multi-data coordination unit includes three-stage processing:

[0024] 1) Retrieve historical cases in the medical knowledge graph;

[0025] 2) Determine priority based on authority weight: attending physician 50%, nursing team 30%, AI system 20%;

[0026] 3) Display the conflict point through an AR pop-up window for gesture confirmation.

[0027] Furthermore, the steps for establishing the respiratory-ECG synchronization model are as follows:

[0028] Data collection: Heart rate variability data is obtained through an ECG monitoring bracelet, and chest displacement signals captured by millimeter-wave radar are simultaneously received;

[0029] Timing alignment: Dynamic time warping algorithm is used to calibrate the time domain of the two signals to compensate for the equipment acquisition delay;

[0030] Feature extraction: extract the RR interval sequence from the ECG signal and analyze the respiratory cycle of the radar signal through wavelet transform;

[0031] Mutual information is calculated to establish the correlation matrix between the respiratory cycle and the RR interval, and the phase synchronization index between the two is calculated:

[0032] in is the phase difference at time t, and N is the total number of sampling points;

[0033] Model training,Long Short-Term Memory Network is used to train clinical samples and output synchronization / desynchronization state classification.

[0034] Furthermore, in the above solution, the ECG monitoring bracelet can be connected to a wireless network set up in the hospital, and medical staff can communicate with the ECG monitoring bracelet through a mobile nursing terminal to achieve voice dialogue or voice broadcast.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] This invention can improve patients' medical experience, conveniently obtain medical information through the bedside interactive unit, enhance their understanding of the disease and sense of participation in medical treatment; the breathing induction animation uses intuitive dynamic waveforms to display and adjust the breathing rate, innovates health education methods, and helps restore normal breathing rhythm; at the same time, the communication function between the ECG monitoring bracelet and the mobile nursing terminal makes information transmission more timely and accurate, greatly improving the patient's medical experience.

[0037] This invention can enhance medical safety. The infusion optimization system calculates personalized infusion rate thresholds based on the patient's physiological characteristics, and issues timely sound and light alarms in case of abnormalities, effectively preventing and controlling infusion safety risks. The vital sign fusion monitoring module integrates multiple data to establish a respiratory-ECG synchronization model to ensure the timeliness and accuracy of vital sign data collection.

[0038] This invention can improve the efficiency of medical work. Through the multi-data coordination unit, it uses the medical record similarity map to recommend solutions, generate reports and heat maps, and provide strong support for medical decision-making; the identity authentication unit uses advanced technology to achieve non-contact and efficient identity verification, ensuring that medical operations are standardized and orderly, and significantly improving the efficiency and quality of medical work. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the functional framework of the present invention. DETAILED DESCRIPTION

[0040] The following is a further detailed description of the embodiments of the invention in conjunction with the accompanying drawings to make the objectives, technical solutions and technical effects of the invention more clearly presented.

[0041] like Figure 1 As shown, the present invention discloses a smart ward platform based on the Internet of Things, which is equipped with a bedside interaction unit, an infusion unit, a nursing electronic whiteboard unit and a ward display screen in the ward; the patient wears an ECG monitoring bracelet, which includes a millimeter wave blood flow monitoring module for collecting vital sign data and ECG monitoring data; it also includes a vital sign fusion monitoring module, an infusion optimization system, an identity authentication unit and a multi-data coordination unit.

[0042] The vital sign fusion monitoring module collects real-time heart rate variability coefficient data through the ECG monitoring bracelet, and integrates and analyzes it with the chest micro-motion signal collected by the millimeter-wave non-contact radar installed next to the bed to establish a respiratory-ECG synchronization model. When the respiratory and ECG rhythms are out of sync for more than 5 minutes, the respiratory induction animation on the display screen is triggered and the respiratory rate is adjusted.

[0043] The infusion optimization system includes a piezoelectric film sensor array built into the infusion device. Based on the patient's weight and the vascular elasticity coefficient obtained by the millimeter wave blood flow monitoring module, it calculates a personalized infusion rate threshold. If the infusion rate threshold deviation exceeds ±15%, the bedside interactive unit will be linked to issue an audible and visual alarm. The infusion rate threshold calculation formula of the infusion optimization system is:

[0044] Vascular elasticity = (α×BMI+β×C_v)×S_t,

[0045] Among them, α is the drug solution characteristic coefficient, β is the vascular compliance correction factor, and S_t is the average infusion adaptability score in the past 3 days. The clinical data of each coefficient are obtained through the hospital PACS system.

[0046] The identity authentication unit integrates a near-infrared structured light projection module, which compares facial 3D depth information with HIS system data to achieve contactless medical identity verification. The identity authentication unit uses a VCSEL array with a wavelength of 850nm to project a dynamic speckle pattern and uses a TOF sensor to collect depth information. The error rate is ≤0.3% in an environment with a light level of ≤5lux.

[0047] The multi-data coordination unit recommends an action plan based on the medical record similarity graph, generating conflict reports and multi-dimensional data comparison heat maps. This unit includes a three-stage process: 1) searching historical cases within the medical knowledge graph; 2) prioritizing decisions based on authority weights: 50% for the attending physician, 30% for the nursing team, and 20% for the AI ​​system; and 3) displaying conflict points via an AR pop-up window for gesture confirmation. This three-stage process addresses issues such as insufficient information, unclear decision priorities, and inefficient interactions during the medical decision-making process, improving the scientific nature and efficiency of medical decision-making.

[0048] The above solution builds a smart ward platform that connects various unit devices through IoT technology, enabling real-time data collection, transmission, and analysis. Specifically, the vital sign data collected by the vital sign fusion monitoring module is used to determine the patient's respiratory-ECG synchrony. Furthermore, HRV analysis data is transmitted to the infusion optimization system, providing a reference for adjusting the infusion rate. The infusion optimization system calculates infusion rate thresholds based on the patient's physiological characteristics and issues an alarm when an abnormality occurs. It also receives lock commands from the multi-data coordination unit to ensure infusion safety. The identity authentication unit verifies the identity of medical staff, ensuring the safety of medical procedures. The multi-data coordination unit analyzes medical records to provide decision support for medical staff and monitors and analyzes data to identify potential issues. Collaborative efforts among these units address technical challenges inherent in traditional ward management, such as insufficient patient information, loopholes in infusion safety monitoring, and low nursing efficiency, thereby improving the quality and efficiency of medical services.

[0049] Furthermore, the vital sign fusion monitoring module transmits HRV analysis data to the infusion optimization system in real time via the SPI bus. When respiratory-ECG desynchronization is detected, the infusion rate recalculation protocol is triggered. The multi-data coordination unit monitors the deviation between the vital sign data and the HIS doctor's instructions. When the deviation exceeds a preset threshold, it writes a lock instruction to the infusion optimization system via the CAN_FD protocol and calls the OpenCV library to generate a three-dimensional data comparison heat map. If the infusion optimization system detects an abnormality in the verification, it feeds correction parameters back to the vital sign fusion monitoring module and simultaneously updates the correction parameters to the abnormal event log area in the NAND flash memory. Within 20ms of detecting an RR interval standard deviation greater than 50ms, the vital sign fusion monitoring module initiates infusion recalculation; the multi-data coordination unit completes the comparison analysis of 50 data points within 200ms. The infusion system sends heartbeat packets to other modules in real time. If no response is received after a 300ms timeout, a system self-test is triggered. The SPI bus is used to enable rapid transmission of vital sign data. When respiratory-ECG desynchronization is detected, the infusion rate is recalculated to adapt to the patient's physical condition. The multi-data coordination unit monitors the deviation between vital sign data and medical orders in real time, locks the infusion system in time and generates an intuitive data comparison heat map to facilitate medical staff to find problems. The feedback mechanism between the infusion optimization system and the vital sign fusion monitoring module ensures the accuracy of the data and the stability of the system. Among them, setting specific conditions and time limits for triggering infusion recalculation, such as the RR interval standard deviation > 50ms and starting within 20ms, and the time requirements for data comparison and analysis of the multi-data coordination unit, such as completing 50 sets of data points within 200ms, is to ensure the rapid response and efficient operation of the system; through the above operations, the system's response speed to changes in patient vital signs is improved, avoiding medical risks caused by untimely response; ensuring the efficiency of the multi-data coordination unit in processing large amounts of data, and providing timely decision support to medical staff.

[0050] Here, the breathing-induced animation generates a dynamic waveform based on the patient's real-time breathing. The waveform parameters are adjusted synchronously with the target breathing rhythm, and the adjustment effect is fed back in real time through the color gradient of the light band at the edge of the display screen. Specifically, the adjustment effect can be adjusted according to the red-yellow-green feedback. The dynamic waveform generation of the breathing-induced animation uses a spline interpolation algorithm, updating the amplitude parameters every 50ms. The target breathing rhythm is dynamically adjusted using the following formula: f_t = f_0 + k·(ΔHRV / 10), where f_0 is the basic breathing guidance frequency, k is the adaptive adjustment coefficient, and ΔHRV is the change in the heart rate variability coefficient; Among them, RMSSD t is the root mean square difference of adjacent RR intervals in real time, RMSSD baseline It is the average value of the patient's resting state in the past 3 days.

[0051] Here, the patient's real-time breathing data is used to generate a dynamic waveform, which is synchronized with the target breathing rhythm by adjusting the waveform parameters. At the same time, the color change of the light band at the edge of the display screen intuitively feedbacks the effect of the patient's breathing regulation, helping the patient to better master the breathing rhythm.

[0052] Specifically, the implementation of the spline interpolation algorithm includes the following parameter constraints:

[0053] (a) Using a cubic uniform B-spline curve 1, the node vectors are set to {u_0, u_1, ..., u_m}, where m = 2k + 3 (k is the number of characteristic points of the respiratory waveform)

[0054] (b) The control point matrix Q_i satisfies: the number of real-time respiratory signal sampling points received every 50ms n∈[15,25], and the deviation threshold δ≤1.2mm;

[0055] (c) Construct collision detection function: Φ(t) = ∑|P_i(t)-Q_i|<0.8σ_max

[0056] Where P_i(t) is the interpolation trajectory point calculated in real time, and σ_max is the maximum chest displacement monitored by the radar;

[0057] (d) The interpolation optimization module is integrated into the FPU unit of the STM32F4 series processor, and it takes ≤30ms to complete a single interpolation.

[0058] Here, we constrain the parameters of the spline interpolation algorithm to ensure the accuracy and stability of dynamic waveform generation. We limit the number of sampling points and the deviation threshold, and construct a collision detection function to ensure that the waveform matches actual breathing conditions. Integrating the interpolation optimization module into the processor's FPU improves the efficiency of the interpolation calculation.

[0059] Here, the steps for establishing the respiratory-ECG synchronization model are as follows:

[0060] (1) Data acquisition: obtaining heart rate variability data through an ECG monitoring bracelet and simultaneously receiving chest displacement signals captured by millimeter-wave radar;

[0061] (2) Timing alignment: Dynamic time warping algorithm is used to calibrate the time domain of the two signals and compensate for the equipment acquisition delay;

[0062] (3) Feature extraction: extract the RR interval sequence from the ECG signal and analyze the respiratory cycle of the radar signal through wavelet transform;

[0063] (4) Mutual information calculation, establishing the correlation matrix between respiratory cycle and RR interval, and calculating the phase synchronization index between the two:

[0064] in is the phase difference at time t, and N is the total number of sampling points;

[0065] (5) Model training: Long short-term memory network is used to train at least 200 clinical samples and output synchronization / desynchronization status classification.

[0066] Through multi-step processing, from data acquisition and time series alignment to feature extraction and mutual information calculation, and finally using long short-term memory networks for model training, an accurate respiratory-ECG synchronization model is established; it solves the problem of synchronous analysis of respiratory and ECG signals, accurately judges the synchronization status of respiratory and ECG, and provides an important basis for clinical diagnosis and treatment.

[0067] In the above solution, the ECG monitoring bracelet can be connected to the wireless network installed in the hospital. Medical staff can communicate with the ECG monitoring bracelet through the mobile nursing terminal to realize voice dialogue or voice broadcast, which can improve the patient's medical experience and enhance the communication and interaction between medical staff and patients.

[0068] In the implementation of the present invention, the specific functional architecture of the platform can be divided into a perception layer, a network layer, a platform layer and an application layer.

[0069] The perception layer includes an ECG monitoring bracelet for collecting vital signs and ECG monitoring data; a millimeter-wave non-contact radar for collecting chest micro-movement signals; a piezoelectric film sensor array for monitoring infusion rate-related data; and a near-infrared structured light projection module for collecting 3D depth information on the faces of medical staff.

[0070] The network layer includes the hospital wireless network, which is used to realize data transmission and communication between various devices.

[0071] The platform layer includes a vital sign fusion monitoring module, which is used to fuse and analyze vital sign data, establish a respiratory-ECG synchronization model, and transmit HRV analysis data; an infusion optimization system, which is used to calculate infusion rate thresholds, issue abnormal alarms, receive lock instructions, etc.; an identity authentication unit, which is used to verify the identity of medical staff; a multi-data coordination unit, which is used to recommend execution plans, generate conflict reports and heat maps, and monitor data deviations.

[0072] The application layer includes a bedside interaction unit, which is used for interactive operations such as sound and light alarms; a nursing electronic whiteboard unit, which is used to display relevant nursing information; a ward display screen, which is used to display information such as respiratory induction animation; and a mobile nursing terminal, which is used to communicate with the ECG monitoring bracelet to realize voice dialogue or voice broadcast.

[0073] The following is a specific implementation example: During a patient's infusion in the hospital ward, the vital signs fusion monitoring module detected that their breathing and ECG rhythm were out of sync for over five minutes. This triggered a breathing induction animation on the display screen. The animation generated a dynamic waveform based on the patient's real-time breathing, with waveform parameters adjusted synchronously with the target breathing rhythm. The color gradient of the light band at the edge of the display provided real-time feedback on the adjustment effect. Simultaneously, the vital signs fusion monitoring module transmitted HRV analysis data to the infusion optimization system, triggering the infusion rate recalculation protocol. The infusion optimization system recalculates the infusion rate threshold based on the patient's weight, vascular elasticity, and other parameters. If the infusion rate deviates by more than ±15%, the bedside interaction unit triggers an audible and visual alarm. The multi-data coordination unit monitors the deviation between the patient's vital signs data and the HIS doctor's instructions. If an anomaly is detected, it writes a lock command to the infusion optimization system via the CAN_FD protocol and calls the OpenCV library to generate a 3D heat map for data comparison. This heat map allows medical staff to visually understand the data deviation and adjust the treatment plan promptly. In addition, medical staff use mobile nursing terminals to have voice conversations with the patient's ECG monitoring bracelet to understand the patient's feelings and provide patients with more considerate medical services.

[0074] The above description is only a preferred embodiment of the present invention and is not intended to limit the scope of the patent application of the present invention. Any equivalent changes, equivalent replacements or modified changes within the technical spirit and principles suggested by the present invention should be included in the scope of patent protection covered by the present invention.

Claims

1. The IoT-based smart ward platform is equipped with a bedside interactive unit, an infusion unit, a nursing electronic whiteboard unit, and a ward display screen. Patients wear an ECG monitoring bracelet that includes a millimeter-wave blood flow monitoring module to collect vital sign data and ECG monitoring data. Also includes: The vital signs fusion monitoring module collects real-time heart rate variability data through an ECG monitoring bracelet, and integrates and analyzes it with chest micro-motion signals collected by a millimeter-wave non-contact radar installed next to the bed to establish a respiratory-ECG synchronization model. When the desynchronization between respiratory and ECG rhythms lasts for more than 5 minutes, the respiratory induction animation on the display screen is triggered and the respiratory rate is adjusted; The infusion optimization system, including a piezoelectric film sensor array built into the infusion device, calculates a personalized infusion rate threshold based on the patient's weight and the vascular elasticity coefficient measured by the millimeter-wave blood flow monitoring module. If the infusion rate threshold deviates by more than ±15%, the bedside interactive unit will be linked to issue an audible and visual alarm. The identity authentication unit integrates a near-infrared structured light projection module to achieve contactless medical identity verification by comparing facial 3D depth information with HIS system data; The multi-data coordination unit recommends execution plans based on the medical record similarity map, and generates conflict reports and multi-dimensional data comparison heat maps.

2. The IoT-based smart ward platform according to claim 1 is characterized by: The vital sign fusion monitoring module transmits HRV analysis data to the infusion optimization system in real time via the SPI bus. When respiratory-ECG desynchronization is detected, the infusion rate recalculation protocol is triggered; the multi-data coordination unit monitors the deviation value between the vital sign data and the HIS doctor's order. When it exceeds the preset threshold, a lock instruction is written to the infusion optimization system via the CAN_FD protocol, and the OpenCV library is called to generate a data comparison three-dimensional thermal map; when the infusion optimization system verifies an abnormality, it feeds back the correction parameters to the vital sign fusion monitoring module and synchronously updates them to the abnormal event log area.

3. The IoT-based smart ward platform according to claim 2, characterized in that: The vital sign fusion monitoring module starts the infusion recalculation within 20ms after detecting that the RR interval standard deviation is greater than 50ms; the multi-data coordination unit completes the comparison and analysis of 50 groups of data points within 200ms.

4. The IoT-based smart ward platform according to claim 1 is characterized by: The calculation formula for the infusion rate threshold of the infusion optimization system is: Vascular elasticity = (α×BMI+β×C_v)×S_t, Among them, α is the drug solution characteristic coefficient, β is the vascular compliance correction factor, and S_t is the average infusion adaptability score in the past 3 days. The clinical data of each coefficient are obtained through the hospital PACS system.

5. The IoT-based smart ward platform according to claim 1 is characterized by: The breathing induction animation generates a dynamic waveform according to the patient's real-time breathing, and the waveform parameters are adjusted synchronously with the target breathing rhythm, and the adjustment effect is fed back in real time through the color gradient of the light band at the edge of the display screen.

6. The IoT-based smart ward platform according to claim 5, characterized in that: The dynamic waveform of the breathing-induced animation is generated using a spline interpolation algorithm, with the amplitude parameters updated every 50 ms. The target breathing rhythm is dynamically adjusted using the following formula: f_t = f_0 + k·(ΔHRV / 10), where f_0 is the basic breathing guidance frequency, k is the adaptive adjustment coefficient, and ΔHRV is the change in the heart rate variability coefficient; Among them, RMSSD t is the root mean square difference of adjacent RR intervals in real time, RMSSD baseline It is the average value of the patient's resting state in the past 3 days.

7. The IoT-based smart ward platform according to claim 6, characterized in that: The implementation of the spline interpolation algorithm includes the following parameter constraints: (a) Using a cubic uniform B-spline curve 1, the node vectors are set to {u_0, u_1, ..., u_m}, where m = 2k + 3 (k is the number of characteristic points of the respiratory waveform) (b) The control point matrix Q_i satisfies: the number of real-time respiratory signal sampling points received every 50ms n∈[15,25], and the deviation threshold δ≤1.2mm; (c) Construct collision detection function: Φ(t) = ∑|P_i(t)-Q_i|<0.8σ_max Where P_i(t) is the interpolation trajectory point calculated in real time, and σ_max is the maximum chest displacement monitored by the radar; (d) The interpolation optimization module is integrated into the FPU unit of the STM32F4 series processor, and it takes ≤30ms to complete a single interpolation.

8. The IoT-based smart ward platform according to claim 1, characterized in that: The multi-data coordination unit includes three-stage processing: 1) Retrieve historical cases in the medical knowledge graph; 2) Determine priority based on authority weight: attending physician 50%, nursing team 30%, AI system 20%; 3) Display the conflict point through an AR pop-up window for gesture confirmation.

9. The IoT-based smart ward platform according to claim 1, characterized in that: The steps for establishing the respiratory-ECG synchronization model are as follows: Data collection: Heart rate variability data is obtained through an ECG monitoring bracelet, and chest displacement signals captured by millimeter-wave radar are simultaneously received; Timing alignment: Dynamic time warping algorithm is used to calibrate the time domain of the two signals to compensate for the equipment acquisition delay; Feature extraction: extract the RR interval sequence from the ECG signal and analyze the respiratory cycle of the radar signal through wavelet transform; Mutual information is calculated to establish the correlation matrix between the respiratory cycle and the RR interval, and the phase synchronization index between the two is calculated: SI = ∑(Δφ_t<30°) / N, where Δφ_t is the phase difference at time t and N is the total number of sampling points; Model training,Long Short-Term Memory Network is used to train clinical samples and output synchronization / desynchronization state classification.

10. The IoT-based smart ward platform according to claim 1, characterized in that: The ECG monitoring bracelet can be connected to a wireless network set up in the hospital, and medical staff can communicate with the ECG monitoring bracelet through a mobile nursing terminal to achieve voice dialogue or voice broadcast.

Images