Intelligent remote ischemia monitoring system based on multi-sensor fusion

Abstract

The invention discloses an intelligent remote ischemia monitoring system based on multi-sensor fusion, relates to the technical field of intelligent detection, and is used for solving the problem of limitation of monitoring of a single sensor. Comprising a data acquisition module, a region division module, a data fusion and ischemia detection module, a data storage module and a remote monitoring and data transmission module, all the modules are in signal connection, and the region division module is combined with a respiratory gas analysis technology to obtain spatial positions, pathological states and metabolic rate data of physiological organs; dividing a detected area into a plurality of sub-areas, and optimizing sensor arrangement; the data fusion and ischemia monitoring module is combined with a weighted summation method, fusion processing of different physiological parameter data is completed, and evidences of a plurality of sensors are synthesized according to the DS evidence theory. And finally, the remote monitoring and data transmission module performs data transmission through an MQTT protocol. And displaying the physiological data of the patient by using a Web platform.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent detection, and more specifically, to a remote ischemic intelligent monitoring system based on multi-sensor fusion. Background Art

[0002] With the progress of medical technology, physiological monitoring and ischemic detection have become important components in clinical diagnosis and telemedicine. Traditional physiological monitoring methods mainly rely on a single sensor to collect physiological data, such as electrocardiogram, blood oxygen saturation, etc. However, the accuracy and comprehensiveness of single-sensor data are limited, and it is difficult to effectively combine factors such as environmental noise and patient activity status, resulting in insufficient reliability of detection results.

[0003] In recent years, the application of multi-sensor fusion technology in health monitoring has gradually emerged. By combining multiple sensors to obtain different types of physiological and environmental data, the accuracy and real-time performance of monitoring can be effectively improved. However, how to reasonably configure sensors according to the characteristics of physiological organs in different regions, optimize data fusion algorithms, and automatically adjust monitoring strategies according to dynamic changes remains a technical problem in current intelligent health monitoring systems.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a remote ischemic intelligent monitoring system based on multi-sensor fusion to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a preferred embodiment, it includes: a data acquisition module, a region division module, a data fusion and ischemia detection module, a data storage module, and a remote monitoring and data transmission module, and the modules are signal-connected to each other;

[0008] The data acquisition module acquires patient physiological data at the hospital end and the home end;

[0009] The region division module divides the detected region into multiple sub-regions and assigns a dynamically adjustable sensor layout scheme to different sub-regions;

[0010] The data fusion and ischemia detection module uses the data fusion technology SVM to synthesize the outputs of sensors in each sub-region, and finally judges the ischemic symptoms and verifies the rationality and severity of the ischemic symptoms;

[0011] The remote monitoring and data transmission module receives and transmits physiological parameter data and ischemia detection results, and conducts remote monitoring and treatment;

[0012] The data storage module stores all data during the processing.

[0013] In a preferred embodiment, the data acquisition module selects corresponding medical sensors at the hospital end to collect patients' physiological parameter data; and selects portable medical acquisition devices at the home end to collect patients' physiological parameter data.

[0014] In a preferred embodiment, the region division module unevenly divides the detected region into multiple sub-regions;

[0015] The region division module obtains the pathological state data of physiological organs in each sub-region through CT (Computed Tomography) technology; records the oxygen inhalation volume VO2 and carbon dioxide exhalation volume VCO2 of each sub-region in real time, calculates the respiratory quotient RQ in each sub-region, and determines the metabolic rate DL in each sub-region according to the respiratory quotient in each sub-region.

[0016] The performance value Bz of each sub-region is determined by using the weighted average calculation method based on the pathological state of the physiological organ and the quantization value of the metabolic rate data. The performance value Bz of each sub-region is compared with the performance value threshold Yb. When the performance value Bz of each sub-region is greater than or equal to the performance value threshold Yb, it indicates that this sub-region belongs to an important detection region; when the performance value Bz of each sub-region is less than the performance value threshold Yb, it indicates that this sub-region belongs to a secondary detection region.

[0017] In a preferred embodiment, the region division module determines the instantaneous performance score P_inst of the physiological parameter data in the important detection region, and determines the comprehensive performance score P_comp of the physiological parameter data by weighted summing the mean Pavg and standard deviation Pstd of the P_inst data, and sets an abnormal threshold T. The comprehensive performance score P_comp is compared with the abnormal threshold T. When the comprehensive performance score P_comp is greater than or equal to the abnormal threshold T, it indicates that the physiological parameter data is in an abnormal state; when the comprehensive performance score P_comp is less than the abnormal threshold T, it indicates that the physiological parameter data is in a normal state.

[0018] In a preferred embodiment, the region division module calculates the performance score P_te of the physiological parameter data in the abnormal state and the standard performance score Pat of the physiological parameter data in the normal state, determines the compensation value A by using a linear fitting model, and adjusts the performance score P_te of the physiological parameter data in the abnormal state according to the compensation value A to obtain the performance score P of the physiological parameter data in the abnormal state after compensation.

[0019] The current performance score deviation amplitude P1 is calculated through the performance score P of the physiological parameter data in the abnormal state after compensation, and the sensor acquisition frequency adjustment coefficient F of the secondary detection region is calculated according to the current performance score deviation amplitude P1.

[0020] In a preferred embodiment, the region division module collects the motion acceleration and angular velocity data of physiological organs in the important detection region, calculates the magnitude Ato of the acceleration and the magnitude Gto of the angular velocity in the T1 time period according to the mean values Ax, Ay, Az of the motion acceleration of the physiological organs and the mean values Gx, Gy, Gz of the angular velocity data in the T1 time period, and calculates the activity state score AcS according to the quantization values of the magnitude Ato of the acceleration and the magnitude Gto of the angular velocity;

[0021] The region division module detects the current environmental noise, collects the noise decibel value NdB in the important detection region, and calculates the mean value Nm of the noise decibel value in the T2 time period; the abnormal situation score AnS of the physiological organs in the current important detection region is determined by weighted summation of the quantization values of the activity state score AcS and the mean value Nm of the noise decibel value, and an abnormal score threshold Yc is set. When the abnormal situation score AnS is greater than the abnormal score threshold Yc, the number of sensors in the secondary detection region is increased; when the abnormal situation score AnS is less than or equal to the abnormal score threshold Yc, the sensitivity of the sensors in the secondary detection region is increased.

[0022] In a preferred embodiment, a support vector machine (SVM) is used to perform preliminary classification processing on the physiological parameter data and output the posterior probability after classification;

[0023] The correlation between the physiological parameter data of each sensor is calculated through the Pearson correlation coefficient, and a weighted average calculation method is used to process the conflicting evidence.

[0024] In a preferred embodiment, the DS evidence theory is used to convert the posterior probability output by the SVM into the basic assignment probability (BPA) required by the DS evidence theory. Based on the synthesized BPA, an ischemic symptom level set is set, and the weighted BPA value of each ischemic symptom level is calculated through a weighted average decision rule, and the ischemic symptom level with the highest BPA value is selected as the final judgment.

[0025] In a preferred embodiment, when an ischemic symptom is detected, the data fusion and ischemia monitoring module calculates the difference between the waveform signal at the home end and the waveform signal at the hospital end and judges the rationality of the ischemic symptom;

[0026] By considering the changes in physiological parameter data such as electrocardiogram, electroencephalogram, blood oxygen, blood pressure, and body temperature, the severity of the ischemic symptom is determined, and a severity threshold Yz is set. When the severity of the ischemic symptom is greater than the severity threshold Yz, it indicates that the patient has a serious ischemic risk at this time; when the severity of the ischemic symptom is less than or equal to the severity threshold Yz, it indicates that the patient has a non-serious ischemic risk at this time.

[0027] In a preferred embodiment, the remote monitoring and data transmission module uploads physiological parameter data and ischemia detection results, and displays the physiological parameter data and ischemia detection results on the front-end interface; the doctor gives treatment suggestions and conducts remote consultations.

[0028] In a preferred embodiment, the method for obtaining the performance value threshold Yb is as follows:

[0029] Calculate the mean and standard deviation of the pathological state and metabolic rate data, and use K-means clustering to analyze the mean and standard deviation of the pathological state and metabolic rate data to determine the performance value threshold Yb.

[0030] The present invention discloses a remote ischemia intelligent monitoring system based on multi-sensor fusion, which relates to the field of intelligent detection technology and is used to solve the limitation problem of single-sensor monitoring; it includes: a data acquisition module, a region division module, a data fusion and ischemia detection module, a data storage module, and a remote monitoring and data transmission module. The modules are signal-connected to each other. The region division module combines respiratory gas analysis technology to obtain the spatial position, pathological state, and metabolic rate data of physiological organs, divides the detected region into multiple sub-regions, and optimizes the sensor layout; the data fusion and ischemia monitoring module combines the weighted summation method to complete the fusion processing of different physiological parameter data and uses the DS evidence theory to synthesize the evidence of multiple sensors. Finally, the remote monitoring and data transmission module performs data transmission through the MQTT protocol; a Web platform is used to display the physiological data of the patient. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic structural diagram of the remote ischemia intelligent monitoring system based on multi-sensor fusion of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0033] Embodiment

[0034] The present invention discloses a remote ischemia intelligent monitoring system based on multi-sensor fusion, as Figure 1 shown, including: a data acquisition module, a region division module, a data fusion and ischemia detection module, a data storage module, and a remote monitoring and data transmission module. The modules are signal-connected to each other.

[0035] The data storage module is used to store all data during the processing process.

[0036] The data acquisition module is mainly used to connect to the microprocessor STM32 through various sensors, and to monitor and collect the physiological parameter data of the detected area in real time through the SPI interface, including: brain function data, electrocardiogram signals, blood oxygen saturation, blood pressure, and body temperature;

[0037] Among them, the data acquisition module is divided into two major sub-modules: hospital-end physiological parameter data acquisition and home-end physiological parameter data acquisition;

[0038] The specific process of hospital-end physiological parameter data acquisition is as follows:

[0039] Step A1: Select the corresponding medical sensors to collect the patient's physiological parameter data:

[0040] Specifically, select a high-precision electroencephalogram sensor, use multiple electrode patches to collect the patient's electroencephalogram activity signals, and obtain brain function data; select an electrode sensor that meets the ECG standard, collect electrocardiogram signals through the ECG interface, and obtain electrocardiogram data. At the same time, perform cardiac CTA+CTP examinations to collect myocardial blood flow MBF, and obtain myocardial ischemia and vascular stenosis conditions and microcirculation data according to the load / resting MBF ratio and myocardial CT blood flow reserve fraction CT-FFR; through cranial CTA+CTP examinations, collect cerebral blood flow CBF, cerebral blood volume CBV, mean transit time MTT, time to peak TTP, and time Tmax when the contrast agent reaches the maximum residual function of the tissue, and obtain data on cranial cerebrovascular ischemia and vascular stenosis conditions. According to cerebral blood flow monitoring, cerebral blood flow monitoring TCD, collect systolic peak velocity Vs, end-diastolic velocity Vd, mean velocity Vm, pulsatility index PI, resistance index RI, and spectral morphology analysis to obtain vascular stenosis / / occlusion conditions; use a pulse oximeter sensor to detect blood oxygen saturation data through infrared and visible light; use an electronic blood pressure sensor to monitor blood pressure and obtain blood pressure data; use an infrared or contact body temperature sensor to collect body temperature data;

[0041] Step A2: After data acquisition, perform filtering through a low-pass filter to remove high-frequency noise and retain the key features of the signal; the processed physiological parameter data is stored through an SD card or a local storage module and periodically backed up to ensure data security.

[0042] Step A3: Upload the collected and processed physiological parameter data to the data storage module through Wi-Fi or Ethernet.

[0043] The specific process of home-end physiological parameter data acquisition is as follows:

[0044] Step B1: Periodically collect the physiological parameter data of the patient at home using a portable electrocardiogram device, a pulse oximeter, an electronic sphygmomanometer, a thermometer, an electrograph device or an infrared brain function imager, and connect each sensor to the smart device via Bluetooth or Wi-Fi.

[0045] Step B2: Filter and denoise the collected physiological parameter data through the APP or embedded device of the smart device; cache the processed physiological parameter data into the device memory or SD card and perform periodic backups.

[0046] Step B3: Periodically upload the physiological parameter data collected at the home end to the cloud platform or the hospital end via Wi-Fi or Bluetooth, and use AES encryption to ensure data privacy and security.

[0047] The region division module is mainly used to extract the spatial positions of each physiological organ and the joints around the physiological organ in the detected region through medical imaging technologies such as MRI, CTA+CTP or MRA, and use the ICD coding controller to calibrate the positions of each physiological organ and the joints around the physiological organ to ensure the accuracy of region division;

[0048] The region division module uses the region division algorithm based on the Voronoi diagram to unevenly divide the detected region into multiple sub-regions according to the spatial positions of each physiological organ in the detected region and the spatial positions of the joints around the physiological organ. For example: divide the limb region into two sub-regions, the upper limb and the lower limb, according to the shoulder, elbow and knee; divide the chest region into two sub-regions, the heart and the heart and lungs, according to the sternum and ribs; record the region above the neck as the head region; select different sensor layout schemes in different sub-regions, and connect each sensor through a wireless sensor network within each sub-region;

[0049] Furthermore, the region division module scans the physiological organs in each sub-region through CT (Computed Tomography) technology to obtain the CT image data of the physiological organs in each sub-region, uses the medical image processing device Osirix to identify the lesion regions in the CT image data, generates pathological state labels, and obtains the pathological state data of the physiological organs in each sub-region; the region division module connects the respiratory gas analyzer CortexMetalyzer to each sub-region, records the oxygen inhalation volume VO2 and the carbon dioxide excretion volume VCO2 of each sub-region in real time, calculates the respiratory quotient RQ of each sub-region, according to the formula: RQ = VCO2 / VO2, calculates the metabolic rate DL of each sub-region according to the respiratory quotient in each sub-region, according to the formula: DL = (A + B) × RQ × VO2, where A and B represent constant coefficients;

[0050] Further, the region division module normalizes the pathological state and metabolic rate data of physiological organs in each sub-region using mathematical formulas or empirical formulas, converting them into quantitative values. Specifically, according to the formula: In the formula, Normalized_Score is the value after normalization; then, according to the quantitative values of the pathological state and metabolic rate data of physiological organs, the performance value Bz of each sub-region is determined by weighted average calculation. Specifically, according to the formula: Bz = Qt × Lt + Qx × Dx, where Lt represents the quantitative value of pathological state data, Qt represents the weight of pathological state data quantitative value, Jt represents the quantitative value of metabolic rate data, and Qx represents the weight of metabolic rate data quantitative value;

[0051] The region division module uses the data cleaning tool Pandas library to remove missing values and duplicate data from the quantified pathological state and metabolic rate data of physiological organs in each sub-region, generating a clean and complete dataset of influencing parameters; calculates the mean and standard deviation of the pathological state and metabolic rate data, and analyzes the mean and standard deviation of the pathological state and metabolic rate data using K-means clustering to determine the performance value threshold Yb. Compare the performance value Bz of each sub-region with the performance value threshold Yb. When the performance value Bz of each sub-region is greater than or equal to the performance value threshold Yb, it indicates that the sub-region belongs to an important detection region; when the performance value Bz of each sub-region is less than the performance value threshold Yb, it indicates that the sub-region belongs to a secondary detection region;

[0052] Sensor layout scheme for important detection regions: Select high-sensitivity ECG sensors, high-precision pulse oximeter sensors, electroencephalogram EEG or near-infrared brain functional imaging, as well as automatic blood pressure monitors and thermistor temperature sensors, and adopt a hexagonal close-packed layout;

[0053] Sensor layout scheme for secondary detection regions: Select low-power ECG sensors and conventional oximeter sensors, arrange relatively simple pressure sensors and thermal sensors, adopt a grid layout, and the spacing can be appropriately increased to avoid being too dense;

[0054] Further, for the physiological parameter data collected from important detection regions, the region division module sets the sliding window size N, that is, all physiological parameter data in the past N seconds are used for each calculation, and determines the minimum value Xmin and maximum value Xmax of each data in the physiological parameter data. Then, all physiological parameter data in the current time window are normalized to obtain the instantaneous performance score Pi nst of the physiological parameter data in the important detection region; calculate the mean Pavg of the Pi nst data within the window N. Specifically, according to the formula:

[0055]

[0056] Next, calculate the standard deviation Pstd of the Pi_nst data within window N according to the formula:

[0057]

[0058] Determine the comprehensive performance score Pcomp of the physiological parameter data within window N by weighted summing the mean Pavg and standard deviation Pstd of the Pi_nst data. Specifically, according to the formula: Pcomp = Qj × Pavg + Qb × Pstd, where Qj represents the weight of the mean Pavg and Qb represents the weight of the standard deviation Pstd. The region division module sets an abnormal threshold T and compares the comprehensive performance score Pcomp with the abnormal threshold T. When the comprehensive performance score Pcomp is greater than or equal to the abnormal threshold T, it indicates that the physiological parameter data collected in the current important detection region is in an abnormal state. When the comprehensive performance score Pcomp is less than the abnormal threshold T, it indicates that the physiological parameter data collected in the current important detection region is in a normal state. Next, take the data of the past M time windows and calculate the instantaneous change rate dP of the comprehensive performance score Pcomp according to the formula: dP = Pcomp(i) - Pcomp(i - 1), where i represents a natural number. Further, calculate the performance score Pte of the physiological parameter data in the abnormal state according to the formula: Pte = Pcomp(i) + K3 × dP, where K3 represents a regulation coefficient.

[0059] When the physiological parameter data collected in the important detection region is abnormal, it indicates that the activity state of the physiological organ is abnormal. Considering the mutual correlation between physiological organs and the chain reaction during physiological abnormalities, the region division module first calculates the abnormal value of the physiological parameter data in the abnormal situation according to the performance score Pte of the physiological parameter data collected in the important detection region in the abnormal state, in combination with the standard performance score Pat of the physiological parameter data. Among them, the standard performance score of the physiological parameter data represents the performance score of the physiological parameter data in the normal state of the physiological organ. Specifically, according to the formula: ΔP = Pat - Pte, where ΔP represents the abnormal value of the physiological parameter data in the abnormal situation.

[0060] Among them, the region division module sets the size H of the historical data window, calculates the mean and standard deviation of all physiological parameter data in the normal state within H days, and calculates the standard performance score Pat of the physiological parameter data by weighted averaging the mean and standard deviation of the physiological parameter data in the normal state.

[0061] The region division module uses a linear fitting model to construct a compensation value calculation formula: A = K1×(Pte - Pat) + ΔP, where A is the compensation value, K1 is the compensation ratio coefficient. Combining the idea of "A - B", according to the compensation value, the physiological parameter data performance score Pte under abnormal conditions is adjusted to obtain the adjusted physiological parameter data performance score P under abnormal conditions. According to the formula: P = Pte×α + A, where α represents the deviation correction factor and 0 < α ≤ 1; then, the current performance score deviation amplitude P1 is calculated through the physiological parameter data performance score P under compensated abnormal conditions. According to the formula: P1 = |P - Pat|, and the sensor acquisition frequency adjustment coefficient F of the secondary detection region is calculated according to the current performance score deviation amplitude P1. According to the formula: F = F0 + K2×P1, where F0 represents the current sensor acquisition frequency of the secondary detection region, and K2 represents the adjustment ratio coefficient;

[0062] When the sensor in the important detection region fails and cannot operate normally, only adjusting the sensor acquisition frequency in the secondary detection region cannot meet the acquisition requirements of the physiological parameter data in the important detection region. Therefore, the region division module uses the acceleration sensor MPU6050 and the gyroscope sensor to collect the motion acceleration and angular velocity data of the physiological organs in the important detection region, and sets the acquisition period T1 to regularly collect the motion acceleration and angular velocity data of the physiological organs in the important detection region; calculate the mean values Ax, Ay, Az of the motion acceleration and the mean values Gx, Gy, Gz of the angular velocity data of the physiological organs within the T1 time period; then calculate the modulus length Ato of the acceleration within the T1 time period. According to the formula: Ato = (Axmean2 + Aymean2 + Azmean2); calculate the modulus length Gto of the angular velocity within the T1 time period. According to the formula: Gto = (Gxmean2 + Gymean2 + Gzmean2), the normalization method is used to convert the modulus length Ato of the acceleration and the modulus length Gto of the angular velocity into quantization values between 0 and 1, and the normalization maximum values Amax and Gmax are set; the activity state score AcS is calculated according to the quantization values of the modulus length Ato of the acceleration and the modulus length Gto of the angular velocity. Specifically, according to the formula: AcS = W1×Ato + W2×Gto, where W1 represents the weight coefficient of the acceleration quantization value, and W2 represents the weight coefficient of the angular velocity quantization value;

[0063] Further, the region division module uses the MEMS microphone sensor ICS-43434 to detect the current environmental noise, sets the data acquisition period T2, regularly collects the noise decibel value NdB in the important detection region, and calculates the average value Nm of the noise decibel value within the T2 time period; uses the normalization method to convert the average value Nm of the noise decibel value into a quantization value between 0 and 1; then, determines the abnormal situation score AnS of the physiological organ in the current important detection region by weighted summation of the activity state score AcS and the quantization value of the average value Nm of the noise decibel value, sets the abnormal score threshold Yc, when the abnormal situation score AnS is greater than the abnormal score threshold Yc, it indicates that the activity state of the physiological organ is strong and the noise interference is large, and at this time, it is necessary to increase the number of sensors in the secondary detection region; when the abnormal situation score AnS is less than or equal to the abnormal score threshold Yc, it indicates that the activity state of the physiological organ is low and the noise interference is small, and the normal acquisition of the physiological parameter data in the important detection region is ensured by increasing the sensitivity of the sensors in the secondary detection region;

[0064] Further, the region division module normalizes the physiological parameter data collected by the sensors in each sub-region and uploads it to the data storage module, and performs timestamp marking;

[0065] The data fusion and ischemia monitoring module uses the particle swarm optimization algorithm PSO to optimize the penalty factor and kernel parameter of the SVM for the physiological parameter data collected by each sensor, by searching for the optimal penalty factor C and kernel parameter in the parameter space, minimizing the classification error and maximizing the classification accuracy;

[0066] Classify the physiological parameter data through the optimized SVM, and output the posterior probability after classification. The SVM classification formula:

[0067] f(x) = ω T φ(x) + b

[0068] Among them, f(x) represents the posterior probability after classification of the physiological parameter data, ω represents the weight vector of the hyperplane, φ(x) represents the mapped high-dimensional feature space, and b represents the bias.

[0069] It should be noted that when there is evidence conflict in the physiological parameter data from different sensors, calculate the correlation between the physiological parameter data of each sensor through the Pearson correlation coefficient, construct the correlation matrix of the evidence, according to the formula:

[0070]

[0071] Among them, r represents the support degree of the sensor physiological parameter data, Xi and Yi represent the sampling values of the physiological parameter data data set, and represents the average value of the sampling values;

[0072] According to the support degree r of the sensor physiological parameter data in the correlation matrix, a weighted average calculation method is used to process the conflicting evidence, based on the formula:

[0073]

[0074] where w represents the value of the weighted support degree of the physiological parameter data, ri represents the weight of each piece of evidence, and xi represents the evidence value corresponding to each piece of evidence;

[0075] Furthermore, set the basic event set G = {A1, A2, A3...} of each physiological parameter data, representing different levels of ischemic symptoms; map the posterior probability value output by each sensor data to the probability assignment m(G) of a basic event, that is: m(Ai) = f(xi), where i = (1, 2, 3...), m(Ai) represents the basic probability assignment (BPA) of the occurrence probability of the ischemic symptom level Ai, and xi represents the posterior probability of the ith physiological parameter data.

[0076] Use Dempster's combination rule to combine two BPAs, and the formula is as follows:

[0077]

[0078] where m(Ai) and m(Bi) represent the basic probability assignments from different sensors, represents the combined BPA; based on the combined BPA, set the ischemic symptom level set H = {Level A, Level B, Level C...}, and calculate the weighted BPA value of each ischemic symptom level through the weighted average decision rule, based on the formula:

[0079]

[0080] where BPA f (C i ) represents the basic probability assignment value of each sensor to the ischemic symptom level, and Wn represents the weight of each sensor; obtain the weighted BPA value corresponding to each ischemic symptom level (such as Level A, Level B, Level C, etc.), and select the ischemic symptom level with the highest BPA value as the final judgment; for example, the finally calculated weighted BPA value is: BPAf = {Level A: 0.7, Level B: 0.2, Level C: 0.1}, and the result will be output as Level A ischemic symptoms.

[0081] When an ischemic symptom is detected, the data fusion and ischemia monitoring module compares the physiological parameter data collected at the hospital end with the physiological data currently collected at the home end to verify the rationality of the ischemic symptom detection and the severity of the current patient's ischemic symptoms;

[0082] The specific comparison process is as follows:

[0083] The dynamic time warping (DTW) is used to compare the dynamic changes of the physiological parameter data waveforms at the hospital end and the home end. By methods such as the correlation coefficient and Euclidean distance, the differences between the waveform signals at the home end and the hospital end are calculated to determine whether there are abnormal waveforms. Among them, the data of electrocardiogram (ECG) and electroencephalogram (EEG) are combined for comparison. If abnormal waveforms appear in both, it indicates a relatively high rationality of the ischemic symptoms.

[0084] Furthermore, by considering the changes in physiological parameter data such as electrocardiogram, electroencephalogram, blood oxygen, blood pressure, body temperature, etc., the weighted scoring method is used to determine the severity of the ischemic symptoms, and a severity threshold Yz is set. When the severity of the ischemic symptoms is greater than the severity threshold Yz, it indicates that the patient has a serious ischemic risk at this time; when the severity of the ischemic symptoms is less than or equal to the severity threshold Yz, it indicates that the patient has a non-serious ischemic risk at this time.

[0085] After that, the data fusion and ischemia monitoring module uploads the rationality of the ischemic symptom detection and the current ischemic symptom severity data of the patient to the remote monitoring and data transmission module.

[0086] The remote monitoring and data transmission module uses EMQX as the MQTT broker and deploys the MQTT Broker; sets the MQTT topic, and adopts the QoS1 or QoS2 mode to upload physiological data and ischemia detection results to ensure the reliability of data transmission. Then, a Web server is built using Node.js + Express, an MQTT Subscriber is established to parse the received MQTT messages. Furthermore, the Web front-end is developed using React / Vue.js, and the physiological data line chart is drawn using ECharts / D3.js to display the patient's physiological data and ischemia detection results on the front-end interface.

[0087] When the ischemic symptom severity score exceeds the threshold, the remote monitoring and data transmission module will automatically send a warning notice to the doctor end. The doctor views the patient's real-time physiological parameter data through remote monitoring, and gives treatment suggestions through the Web front-end interface according to the physiological data and the severity of the ischemic symptoms for remote consultation.

[0088] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0089] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0090] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application of the technical solution and the constraints of the invention. A person skilled in the art can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0091] In addition, the various functional modules in the embodiments of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0092] As described above, this is only a specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0093] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. Remote ischemia intelligent monitoring system based on multi-sensor fusion, characterized by ; It includes: data acquisition module, area division module, data fusion and ischemia detection module, data storage module and remote monitoring and data transmission module, and signal connection between each module; The data collection module collects the patient's physiological data at the hospital and home end; The area division module divides the detected area into multiple sub-areas and allocates dynamically adjustable sensor layout schemes to different sub-areas; The data fusion and ischemia detection module uses the data fusion technology SVM to integrate the output of sensors in each sub-area, and finally judges the ischemia symptoms and verifies the rationality and severity of the ischemia symptoms; The remote monitoring and data transmission module receives and transmits physiological parameter data and ischemia detection results, and performs remote monitoring and treatment; The data storage module stores all the data in the processing process.

2. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 1 is characterized by: The data acquisition module selects corresponding medical sensors at the hospital end to collect the patient's physiological parameter data; and selects portable medical acquisition equipment at the home end to collect the patient's physiological parameter data.

3. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 2 is characterized by: The region division module divides the detected area into multiple sub-regions unevenly; The regional division module obtains the pathological status data of the physiological organs in each sub-region through CT computer tomography technology; records the oxygen intake VO2 and carbon dioxide discharge VCO2 of each sub-region in real time, and calculates the respiratory quotient RQ in each sub-region, and determines the metabolic rate DL in each sub-region based on the respiratory quotient in each sub-region; According to the pathological state of the physiological organs and the quantitative value of the metabolic rate data, the performance value Bz of each sub-region is determined by weighted average calculation, and the performance value Bz of each sub-region is compared with the performance value threshold Yb. When the performance value Bz of each sub-region is greater than or equal to the performance value threshold Yb, it indicates that the sub-region belongs to an important detection area; when the performance value Bz of each sub-region is less than the performance value threshold Yb, it indicates that the sub-region belongs to a secondary detection area.

4. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 3 is characterized by: The area division module determines the instantaneous performance score Pinst of the physiological parameter data in the important detection area, and determines the comprehensive performance score Pcomp of the physiological parameter data by weighted summing up the mean Pavg and standard deviation Pstd of the Pinst data, and sets the abnormal threshold T, and compares the comprehensive performance score Pcomp with the abnormal threshold T. When the comprehensive performance score Pcomp is greater than or equal to the abnormal threshold T, it indicates that the physiological parameter data is in an abnormal state; When the comprehensive performance score Pcomp is less than the abnormal threshold T, it indicates that the physiological parameter data is in a normal state.

5. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 4, characterized in that; The area division module calculates the performance score Pte of the physiological parameter data under abnormal conditions and the standard performance score Pat of the physiological parameter data under normal conditions, determines the compensation value A by using a linear fitting model, and adjusts the performance score Pte of the physiological parameter data under abnormal conditions according to the compensation value A to obtain the performance score P of the physiological parameter data under abnormal conditions after compensation; The current performance score deviation amplitude P1 is calculated through the compensated physiological parameter data performance score P under abnormal conditions, and the secondary detection area sensor acquisition frequency adjustment coefficient F is calculated according to the current performance score deviation amplitude P1.

6. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 5 is characterized by: The area division module collects the motion acceleration and angular velocity data of the physiological organs in the important detection area, calculates the modulus Ato of the acceleration and the modulus Gto of the angular velocity in the T1 time period according to the mean values ​​Ax, Ay, Az of the motion acceleration of the physiological organs and the mean values ​​Gx, Gy, Gz of the angular velocity data in the T1 time period, and calculates the activity state score AcS according to the quantized values ​​of the modulus Ato of the acceleration and the modulus Gto of the angular velocity; The area division module detects the current environmental noise, collects the noise decibel value NdB in the important detection area, and calculates the average noise decibel value Nm within the T2 time period; determines the abnormal situation score AnS of the physiological organs in the current important detection area according to the weighted sum of the activity state score AcS and the quantized value of the noise decibel value average Nm, and sets the abnormal score threshold Yc. When the abnormal situation score AnS is greater than the abnormal score threshold Yc, the number of sensors in the secondary detection area is increased; when the abnormal situation score AnS is less than or equal to the abnormal score threshold Yc, the sensitivity of the sensors in the secondary detection area is improved.

7. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 6, characterized in that: Use support vector machine (SVM) to perform preliminary classification on physiological parameter data and output the posterior probability after classification; The correlation between the physiological parameter data of each sensor was calculated using the Pearson correlation coefficient, and the weighted average calculation method was used to deal with conflicting evidence.

8. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 7 is characterized by: The posterior probability output by SVM is converted into the basic assignment probability BPA required by DS evidence theory using the DS evidence theory. Based on the synthesized BPA, a set of ischemic symptom levels is set, and the weighted BPA value of each ischemic symptom level is calculated by the weighted average decision rule. The ischemic symptom level with the highest BPA value is selected as the final judgment.

9. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 8, characterized in that: When ischemic symptoms are detected, the data fusion and ischemic monitoring module calculates the difference between the waveform signal at the home end and the waveform signal at the hospital end to determine the rationality of the ischemic symptoms; The severity of ischemic symptoms is determined by considering changes in physiological parameter data such as electrocardiogram, electroencephalogram, infrared brain functional imaging, blood oxygen, blood pressure, body temperature, etc., and a severity threshold Yz is set. When the severity of ischemic symptoms is greater than the severity threshold Yz, it indicates that the patient is at risk of severe ischemia; when the severity of ischemic symptoms is less than or equal to the severity threshold Yz, it indicates that the patient is at risk of non-severe ischemia.

10. The remote ischemia intelligent monitoring system based on multi-sensor fusion according to claim 3 is characterized by: The method for obtaining the performance value threshold Yb is as follows: The mean and standard deviation of the pathological status and metabolic rate data were calculated, and K-means clustering was used to analyze the mean and standard deviation of the pathological status and metabolic rate data to determine the performance value threshold Yb.

Images