Information interaction control method and system for a vital signs monitor

Abstract

This invention relates to the field of vital sign monitoring technology, and more particularly to an information interaction control method and system for a vital sign monitor. The method includes the following steps: synchronously acquiring physiological signals via a sensor array, wherein the physiological signals include electrocardiogram (ECG) waveform data, pulse wave data, motion acceleration data, and body surface temperature data; performing time alignment of the ECG waveform data, pulse wave data, and motion acceleration data based on a clock synchronization protocol, and constructing a matrix to generate a synchronized vital sign matrix; constructing a risk change trajectory for pathological trajectories on filtered data and historical case data to generate a vital sign risk trajectory, and performing out-of-bounds judgment on the vital sign risk trajectory; this invention, through this high-precision synchronization mechanism, ensures that the intrinsic correlation between ECG, pulse wave, motion, and temperature data is accurately captured.

Description

Technical Field

[0001] This invention relates to the field of vital sign monitoring technology, and in particular to an information interaction control method and system for a vital sign monitoring instrument. Background Technology

[0002] Existing vital sign monitoring devices have significant shortcomings in their information interaction and control methods: Multi-source physiological signals (ECG / photoplethysmography / accelerometer) experience timestamp drift due to hardware clock asynchrony, requiring fixed delay compensation and failing to dynamically adapt to crystal oscillator drift, thus rendering motion interference suppression ineffective; the signal processing channel selection mechanism is rigid, statically switching filtering modes based solely on acceleration thresholds, ignoring the dynamic characteristics of motion spectrum entropy (0.5-10Hz) and physiological signal covariance, resulting in over-filtering in static scenarios or artifacts remaining in motion scenarios; risk warnings rely on a single threshold cutoff (e.g., ECG risk >80%), failing to identify early risk trends; and data traceability is weak, storing only the current data fragment when clinical alarms are triggered, lacking full-process node markings, making it impossible to reproduce and analyze false positive events. Summary of the Invention

[0003] Therefore, it is necessary to provide an information interaction control method and system for a vital signs monitor to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, an information interaction control method for a vital signs monitor is provided, the method comprising the following steps:

[0005] Step S1: Synchronously collect physiological signals through a sensor array, including electrocardiogram waveform data, perturbation pulse wave data, motion acceleration data, and body surface temperature data;

[0006] Step S2: Based on the clock synchronization protocol, time-align the ECG waveform data, pulse wave data, and motion acceleration data, and construct a matrix to generate a synchronized vital signs matrix;

[0007] Step S3: Calculate the motion divergence value based on the synchronized vital signs matrix and motion acceleration data, select the processing channel based on the motion divergence value, and perform filtering operation to obtain filtered data;

[0008] Step S4: Perform risk change trajectory construction on the filtered data and historical case data to generate vital sign risk trajectories, and make out-of-bounds judgments on the vital sign risk trajectories; when the vital sign risk trajectory exceeds the threshold, trigger a clinical alarm.

[0009] This specification provides an information interaction control system for a vital signs monitor, used to execute the aforementioned information interaction control method based on the vital signs monitor. The information interaction control system for the vital signs monitor includes:

[0010] The multi-source synchronous acquisition module is used to synchronously acquire physiological signals through a sensor array, including electrocardiogram waveform data, pulse wave data, motion acceleration data, and body surface temperature data.

[0011] The timing alignment and feature matrix construction module is used to perform time alignment on ECG waveform data, capacitive pulse wave data and motion acceleration data based on the clock synchronization protocol, and to construct a matrix to generate a synchronized vital signs matrix.

[0012] The motion interference sensing and filtering channel selection module is used to calculate the motion divergence value based on the synchronized vital signs matrix and motion acceleration data, perform filtering operations, and select the processing channel based on the motion divergence value to obtain filtered data.

[0013] The pathological evolution analysis and clinical early warning module is used to construct risk change trajectories of pathological trajectories from filtered data and historical case data, generate vital sign risk trajectories, and make out-of-bounds judgments on vital sign risk trajectories; when the vital sign risk trajectory exceeds the threshold, a clinical alarm is triggered.

[0014] The beneficial effects of this invention are as follows: on the one hand, by dynamically calculating the motion divergence value through motion acceleration data and intelligently selecting the optimal filtering channel to perform signal processing, the purity of ECG waveforms and pulse wave signals under motion interference is improved; furthermore, the adaptive filtering strategy based on real-time motion state effectively overcomes the limitations of traditional fixed filtering in dynamic scenarios.

[0015] On the other hand, by integrating real-time filtered data with historical case data, a vital sign risk trajectory that evolves over time is constructed; by quantitatively characterizing the dynamic changes in cardiovascular and circulatory system risks, early warning of pathological risks is achieved; in particular, the risk trajectory over-limit judgment mechanism can trigger clinical alarms in real time when the risk exceeds the safety threshold, thus gaining a crucial time window for intervention in critical and severe cases.

[0016] Finally, the synchronous acquisition of the sensor array, combined with the clock synchronization protocol, constructs a vital signs matrix, providing a spatiotemporally aligned data foundation for multimodal signal fusion analysis. This high-precision synchronization mechanism ensures that the intrinsic correlation between ECG, pulse wave, motion, and temperature data is accurately captured. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the steps of an information interaction control method for a vital signs monitor.

[0018] Figure 2 Demonstrates the filtering effect on electrocardiogram waveform data;

[0019] Figure 3 Demonstrates the filtering effect on capacitance pulse wave data;

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] The technical method of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention.

[0022] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0023] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0024] To achieve the above objectives, please refer to Figures 1 to 3 An information interaction control method for a vital signs monitoring device, the method comprising the following steps:

[0025] Step S1: Synchronously collect physiological signals through a sensor array, including electrocardiogram waveform data, perturbation pulse wave data, motion acceleration data, and body surface temperature data;

[0026] In this embodiment of the invention, during the monitoring of vital signs, four types of raw physiological signals are collected in parallel at a fixed sampling frequency (e.g., 250Hz) by a hardware-synchronized sensor array: voltage timing signals generated by electrocardiogram leads (ECG waveform data), light absorption intensity timing signals captured by photoelectric sensors (capacitance pulse wave data), spatial acceleration vectors output by triaxial accelerometers (motion acceleration data), and body surface temperature scalar values ​​measured by thermal sensors (body surface temperature data).

[0027] Step S2: Based on the clock synchronization protocol, time-align the ECG waveform data, pulse wave data, and motion acceleration data, and construct a matrix to generate a synchronized vital signs matrix;

[0028] In this embodiment of the invention, a precise clock synchronization protocol (such as IEEE 1588 PTP) is used to synchronize asynchronously acquired electrocardiogram waveform sequences. }, Capacitance Pulse Wave Sequence { } and motion acceleration sequence { Time axis calibration: First, extract the hardware timestamps of each signal, and then uniformly map them to the reference time axis using a clock skew compensation algorithm. Then, using the highest sampling rate as a baseline (e.g., 250Hz for ECG), cubic spline interpolation is performed on the pulse wave (typically 100Hz) and acceleration (typically 50Hz) data to generate tightly aligned, equally spaced sequences. }; finally, based on timestamps Constructing a synchronized vital sign matrix SVM;

[0029] In one implementation of this invention, a reference time is set. Raw data at 4ms intervals (250Hz):

[0030] Electrocardiogram: Time-collected values ;

[0031] Pulse wave: Time-collected values (No unit light intensity);

[0032] Acceleration: Time-collected values ;

[0033] After clock compensation and interpolation, in Synchronization vectors are generated at specific times, ultimately constructing a 10-second data matrix. (2500 rows × 6 columns).

[0034] Step S3: Calculate the motion divergence value based on the synchronized vital signs matrix and motion acceleration data, select the processing channel based on the motion divergence value, and perform filtering operation to obtain filtered data;

[0035] In this embodiment of the invention, the synchronous vital signs matrix is ​​time-aligned with the triaxial motion acceleration data (sampling frequency of 50Hz). The root mean square (RMS) value of the acceleration vector is calculated as a motion intensity index, and the difference between the motion signal and the vital signs signal is quantified based on covariance analysis to obtain the motion divergence value (range 0-1, the larger the value, the more significant the motion interference). Subsequently, according to a preset threshold (e.g., divergence value > 0.7), a high signal-to-noise ratio processing channel is selected (e.g., channel index optimized by motion compensation algorithm is preferred), and adaptive filtering is performed to suppress high-frequency noise and motion artifacts, and finally the filtered vital signs time series data is output.

[0036] In one implementation of this invention, it is assumed that the synchronized vital signs matrix contains 10 seconds of sampled data (sampling rate 100Hz) of heart rate (70bpm±2bpm), respiratory rate (16bpm±1bpm), and SpO2 (98%±1%), and the motion acceleration data are acceleration vectors along the x-axis (0.8g), y-axis (0.6g), and z-axis (0.7g). After calculation, the motion divergence value is 0.82 (derived from the covariance of acceleration RMS and heart rate signal), which exceeds the threshold of 0.75. Therefore, the second real-time filtering channel is selected for processing, and the noise variance of the original heart rate signal is reduced from 0.15bpm² to 0.05bpm², resulting in a smooth filtered data sequence.

[0037] Step S4: Perform risk change trajectory construction on the filtered data and historical case data to generate vital sign risk trajectories, and make out-of-bounds judgments on the vital sign risk trajectories; when the vital sign risk trajectory exceeds the threshold, trigger a clinical alarm.

[0038] In this embodiment of the invention, the filtered vital sign data output in step S3 is dynamically time-warped (DTW) aligned with the historical case database (pathological data of similar patients), and the transition probability of the current vital sign sequence relative to the historical pathological trajectory is calculated. A continuous risk value sequence (range 0-100, i.e., vital sign risk trajectory) is constructed in a probability-weighted manner. An out-of-bounds judgment is made in combination with a preset clinical threshold (such as static threshold > 85 or dynamic threshold > 3 times the standard deviation). If 5 consecutive sampling points exceed the boundary, a clinical alarm is triggered.

[0039] In one implementation of this invention, it is assumed that the current filtered data is a 10-minute heart rate sequence (mean 72 bpm, fluctuation ±3 bpm) and SpO2 sequence (mean 94%, fluctuation ±1%). The historical case database matches data from 3 patients with similar pneumonia in the worsening stage. According to HMM calculation, the transition probability shows that the current state transitions to high risk with a probability of 0.78. The constructed risk value sequence reaches 89.2 (threshold 85) at t=125 seconds, and the subsequent 4 sampling points (interval 0.1 seconds) are consistently higher than 89.5 (standard deviation 2.1, dynamic threshold 87.3).

[0040] As mentioned above, a level 2 alarm (code ALM-03) is triggered, and the out-of-bounds parameters are recorded as [heart rate risk weight: 0.6, SpO2 risk weight: 0.4, duration: 0.5 seconds].

[0041] Preferably, step S2 includes the following steps:

[0042] The first synchronous time series is extracted by normalizing the electrocardiogram waveform data.

[0043] Perform the same time calibration on the capacitive pulse wave data and extract the second synchronous time series;

[0044] Clock compensation is performed on the raw timestamps of the motion acceleration data to extract the third synchronization time series;

[0045] The first, second, and third synchronized time series are subjected to the same interpolation and resampling to generate the first aligned ECG sequence, the second aligned PPG sequence, and the third aligned acceleration sequence, respectively.

[0046] A synchronized vital signs matrix was constructed using the first aligned ECG sequence, the second aligned PPG sequence, and the third aligned accelerometer sequence.

[0047] In one implementation of this invention, assuming the original sampling point of the electrocardiogram waveform data is 500Hz and the value range is [-1.5mV, 1.5mV], the sequence within the time window of 0-2 seconds is extracted as the first synchronization time series through Z-score normalization (mean is 0, standard deviation is 1). For example, the normalization value is 0.8 at t=1.0 seconds.

[0048] In another implementation of the present invention, time calibration is performed on the photoplethysmography data (original sampling rate 100Hz), linear interpolation is used to compensate for device delay (preset delay 20ms), and a sequence of the same time window 0-2 seconds is extracted as the second synchronization time sequence. For example, the calibration value at t=1.0 seconds is 1024 photoplethysmography units.

[0049] In another implementation of this invention, clock compensation is performed on the motion acceleration data (three-axis original sampling rate 50Hz, timestamp offset 5ms). The time base is adjusted by subtracting a fixed offset, and a third synchronization time series is extracted. For example, the x-axis compensation value is 0.5g at t=1.0 seconds.

[0050] In another implementation of this invention, the first, second, and third synchronized time series are uniformly linearly interpolated and resampled to a common sampling rate of 100Hz to generate a first aligned ECG sequence (e.g., after resampling, the value is 0.75 at t=1.0 seconds), a second aligned PPG sequence (the value is 1010 units at t=1.0 seconds), and a third aligned acceleration sequence (the x-axis value is 0.48g at t=1.0 seconds).

[0051] In another implementation of this invention, the above-mentioned alignment sequence is integrated into a synchronous vital signs matrix according to time points, and the matrix dimension is N×3 (N is the number of sampling points). For example, at t=1.0 seconds, the row vector is [0.75, 1010, 0.48].

[0052] Preferably, the mathematical expression of the synchronized vital sign matrix includes:

[0053]

[0054] in, To synchronize the vital signs matrix, For timestamps, Represented as the first aligned ECG sequence, Represented as the second aligned PPG sequence, Represented as the third aligned acceleration sequence, Expressed as X-axis acceleration, Expressed as Y-axis acceleration, It is expressed as the acceleration along the Z-axis.

[0055] In one implementation of this invention, it is assumed that t = 3.2s. It is -0.35 (original value is -0.4mV). It is 850. It is 920. , , If the amounts are 0.48g, -0.12g, and 1.05g respectively, then the expression of SVM is [3200, -0.35, 850, 920, 0.48, -0.12, 1.05].

[0056] Preferably, step S3 includes the following steps:

[0057] Step S31: Perform triaxial acceleration analysis based on the synchronized vital signs matrix and motion acceleration data, and calculate the motion divergence value;

[0058] Step S32: Select a processing channel based on the motion divergence value; when the motion divergence value is greater than a preset threshold, it is determined to be in motion and the first real-time filtering channel is activated; when the motion divergence value is less than or equal to the preset threshold, it is determined to be in a stationary state and the second real-time filtering channel is activated.

[0059] Step S33: Perform a filtering operation through the first real-time filtering channel or the second real-time filtering channel to obtain filtered data.

[0060] In embodiments of the present invention, the moving root mean square (RMS) of the acceleration vector magnitude is calculated based on the triaxial acceleration data in the synchronized vital signs matrix as an indicator of exercise intensity (window length 0.5 seconds). Simultaneously, the covariance between the electrocardiogram (ECG) sequence and the acceleration magnitude is extracted to generate an exercise divergence value. .Will Compared with the preset threshold θ (θ=0.35), if If the target physiological signal is >θ (determined to be in motion), activate the first real-time filtering channel (using a normalized minimum mean square adaptive filter, step size parameter 0.02, cutoff frequency 0.1-10Hz); otherwise, activate the second real-time filtering channel (static Butterworth low-pass filter, cutoff frequency 5Hz). Step S33 performs filtering operation on the selected channel on the target physiological signal and outputs the noise-suppressed filtered data sequence.

[0061] Preferably, step S31 includes the following steps:

[0062] The triaxial acceleration is calculated based on the synchronized vital signs matrix, and the mean variance within the time window is calculated to obtain the dynamic variance.

[0063] Acceleration vectors were extracted from the synchronized vital signs matrix, and Shannon entropy was calculated in the 0.5-10Hz frequency band of the extracted acceleration vectors to obtain the spectral entropy value.

[0064] The dynamic variance and spectral entropy values ​​are combined to obtain the motion divergence value.

[0065] The specific operation is as follows: Calculate the magnitude of the acceleration vector using a sliding window (window length 0.5 seconds, step size 10 ms). The mean of the variance, that is, for all values ​​within the window. Sample variance Then, take the arithmetic mean of the variances of five consecutive windows to obtain the dynamic variance. Perform a Fast Fourier Transform (FFT, 100Hz sampling rate) on the acceleration vector magnitudes within the same window to extract the spectral energy distribution within the 0.5-10Hz frequency band, and then normalize the energy probability. Substituting into Shannon's entropy formula Obtain the spectral entropy value; calculate the dynamic variance. and spectral entropy Divide by the preset baseline value (assuming) Normalize, then sum by weight. Generate motion divergence values .

[0066] In another implementation of this invention, the weighted summation ratio is 7:3, which is determined based on conventional experience and may sometimes be further determined according to the actual situation.

[0067] Preferably, step S33 includes the following steps:

[0068] Using motion acceleration data as a noise reference signal, the ECG waveform data is input to the first real-time filtering channel to perform reverse elimination of ECG waveform jitter noise, generating the first purified ECG signal; the capacitance pulse wave data is input to the first real-time filtering channel to decompose it into AC and DC components, the AC component is subjected to motion interference suppression, and then it is recombined with the DC component to obtain the first purified pulse wave signal;

[0069] Alternatively, based on the ECG waveform data input to the second real-time filtering channel, a fixed frequency window filter is applied to retain the clinically effective band of 0.5-40Hz, and the second purified ECG signal is directly output; the capacitive pulse wave data input to the second real-time filtering channel is smoothed to generate the second purified pulse wave signal.

[0070] The first purified ECG signal and the first purified pulse wave signal, or the second purified ECG signal and the second purified pulse wave signal, are combined into filtered data.

[0071] In one implementation of this invention, differentiated filtering is performed based on channel selection: if the first real-time filtering channel (motion state) is activated, then...

[0072] 1) Using the magnitude of the triaxial acceleration vector Using a sampling rate of 100Hz as a noise reference signal, a normalized least mean square (NLMS) adaptive filter (step size 0.02, order 32) is applied to the ECG (normalized voltage sequence). Motion artifacts are canceled in real time by adjusting the filter weights, and the first purified ECG signal is output.

[0073] 2) The sequence (in units of light intensity) is bandpass filtered (0.5-8Hz) to separate the AC and DC components. The AC component is scaled by applying a motion suppression coefficient based on the acceleration covariance (e.g., weight 0.85), and then recombined with the DC component to generate the first purified pulse wave signal (e.g., original PPG value 850 → AC=120 / DC=730 → suppressed AC=102 → recombined value 832).

[0074] The recombination formula is as follows: Of course, the weighting formula depends on conventional experience and can sometimes be further determined based on the actual situation.

[0075] If the second real-time filtering channel is activated (in a static state), then

[0076] 1) Perform a 5th-order Butterworth bandpass filter (0.5-40Hz) directly on the ECG sequence to preserve the clinical band and output a second purified ECG signal;

[0077] 2) To Denoising of the sequence (window length 15 points, polynomial order 3) generates a second purified pulse wave signal;

[0078] Finally, the ECG and pulse wave signals from the same channel are merged into filtered data according to the timestamp.

[0079] The specific steps are as follows.

[0080] Input variables are: ECG sequence fragments: [-0.35, -0.41, -0.38]; Sequence fragment: [850,845,848]; Reference noise levels: [0.87g, 0.89g, 0.85g];

[0081] The intermediate steps in the process are:

[0082] ECG output after NLMS filtering: [-0.32, -0.37, -0.35] (motion artifact suppression rate 18%).

[0083] Please see Figure 2 This illustrates the filtering process applied to the ECG sequence. The red curve represents an abnormal valley of -0.41 in the original ECG signal at t=3200ms (exceeding the static threshold of -0.35). After NLMS adaptive filtering, the filtered signal (red dashed line) suppresses the valley to -0.32 (an improvement of 0.09 units), and the original signal of -0.41 at t=3210ms is reduced to -0.37 after filtering. (Motion artifact suppression rate 18%)

[0084] PPG decomposition: DC component = 830, AC component = [20, 15, 18];

[0085] After motor inhibition, AC=[17,12.75,15.3]; recombinant PPG=[847,842.75,845.3].

[0086] Please see Figure 3 The green dashed line represents the original PPG signal, and the green solid line represents the filtered data. The original PPG signal drops to 840 units at t=3210ms (a decrease of 5 units from the previous point). After AC / DC decomposition (DC component 830 units, AC component 15 units) and motion suppression (weight 0.85), the AC component is compressed to 12.75 units, and the recombined output is 842.75 units (error <0.5%). The deviation between the filtered value of 845.3 and the original value of 848 at t=3220ms is due to dynamic compensation of acceleration covariance (compensation coefficient 0.92). Finally, the signal-to-noise ratio of the PPG sequence is improved from 12dB to 28dB.

[0087] The output filtered data is then:

[0088] [t=3200ms:ECG=-0.32,PPG=847];

[0089] [t=3210ms:ECG=-0.37,PPG=842.75];

[0090] [t=3220ms:ECG=-0.35,PPG=845.3].

[0091] Preferably, step S4 includes the following steps:

[0092] Step S41: Perform risk change trajectory construction on the filtered data and historical case data to generate vital sign risk trajectories, which include ECG risk sub-trajectories and blood flow risk sub-trajectories;

[0093] Step S42: Perform boundary judgment based on ECG risk sub-trajectory and blood flow risk sub-trajectory. When the vital sign risk trajectory exceeds the threshold, trigger a clinical alarm to obtain vital sign alarm data.

[0094] Step S43: Perform full-process tracing based on vital sign alarm data and mark key nodes of the detector; generate a report using the key nodes of the detector to obtain a full-process report of the vital sign detector.

[0095] In this embodiment of the invention, the Mahalanobis distance between the QRS complex slope change rate and ST segment offset is calculated between the ECG signal in the filtered data and the ECG template of the same pathological type (such as the trajectory of ventricular premature beat deterioration) in the historical case database; simultaneously, the perfusion index of the blood flow signal (PPG, light intensity sequence) is extracted and correlation analysis is performed with the historical low perfusion risk template to generate a blood flow risk sub-trajectory; in step S42, a dual threshold mechanism is set to trigger a clinical alarm and generate structured alarm data containing timestamp, over-limit type and risk level.

[0096] Preferably, the risk change trajectory construction of pathological trajectories for filtered data and historical case data includes the following steps:

[0097] Acquire historical case data; overlay body surface temperature data onto filtered data to generate a fused vital signs sequence;

[0098] Electrocardiogram (ECG) features and blood flow features were extracted from the fused vital sign sequences.

[0099] ECG features and historical case data are analyzed to perform ECG feature sequence analysis and output cardiovascular risk scores. Temporal analysis of cardiac electrical activity is performed on the cardiovascular risk scores, and the risk change trajectory is constructed based on the timestamp as the horizontal axis and the predicted value of the cardiovascular risk score as the vertical axis to obtain the ECG risk sub-trajectory.

[0100] Blood flow characteristics and historical case data were analyzed for blood pressure, and the SpO2-PWV organ parameters were used to quantify the risk score of circulatory failure. The circulatory failure risk score was analyzed in a time series of the circulatory system, and the risk change trajectory was constructed by using the timestamp as the horizontal axis and the predicted value of the circulatory failure risk score as the vertical axis, thus obtaining the blood flow risk sub-trajectory.

[0101] In one implementation of this invention, historical case data (including time series of ECG waveforms, blood oxygen saturation, pulse wave velocity, etc. of similar patients) are retrieved from a medical database. Simultaneously, filtered data (ECG signals and pulse wave signals) and body surface temperature sensor data (sampling rate 1Hz) are aligned using timestamps and linearly interpolated to a unified frequency of 100Hz to form a six-dimensional fused vital sign sequence [ECG, PPG, Temp]. Then, ECG features are extracted: QRS amplitude (mV), ST segment slope (mV / s), RR interval coefficient of variation (%), and blood flow features: perfusion index PI (%), pulse wave rise time (ms), and SpO2 value (%). Subsequently, risk scoring analysis is performed on the real-time ECG features, constructing an ECG risk sub-trajectory with timestamps as the horizontal axis and the mean of a 30-second sliding window as the vertical axis. Simultaneously, the blood flow features are combined with historical data to calculate the SpO2-PWV organ parameter.

[0102] The specific steps are as follows.

[0103] 1) Assume the input data is:

[0104] Filtered ECG signal: QRS amplitude = 1.1mV, ST slope = 0.6mV / s, RR interval variability = 12%;

[0105] Filtered pulse wave: PI=1.2%, SpO2=96%;

[0106] Body surface temperature: 36.8℃;

[0107] Patient baseline data: Blood pressure BP = 120 / 80 mmHg, Age = 65 years.

[0108] 2) The cardiovascular risk score is:

[0109] Risk score = 0.3 × 1.1 + 0.5 × 0.6 + 0.2 × 12 = 0.33 + 0.3 + 2.4 = 3.03 → Normalized to a percentage: 3.03 × 20 = 60.6 points.

[0110] The risk score for circulatory failure is:

[0111] PWV=0.8× =0.8× 5≈0.8×9.01=7.21m / s (K=0.8)

[0112] Decision tree judgment: PI=1.2%>1.0 (no trigger + score), PWV=7.21<10 (no trigger + score) → basic risk value 40 points.

[0113] 3) Assume that during the monitoring period from 115 seconds to 125 seconds, the ECG risk score shows a continuous upward trend (115 seconds: 58.2 points; 120 seconds: 60.6 points; 125 seconds: 62.1 points), while the blood flow risk score increases synchronously (115 seconds: 38.5%; 120 seconds: 40.0%; 125 seconds: 41.2%), with the ECG risk score exceeding the preset threshold of 60 points for the first time at 120 seconds.

[0114] Key point: An alarm is triggered when the ECG risk score is >60 or the blood flow risk score is >45 for 5 consecutive ECGs. In this case, the ECG risk first exceeded the threshold at 120s (60.6 > 60), but did not form a continuous threshold, so no alarm was triggered.

[0115] In another implementation of this invention, the threshold is determined based on conventional experience. This invention sets the threshold for ECG risk value to 80 and the threshold for blood flow risk value to 75. However, in this embodiment, the threshold is set to ECG risk > 60 or blood flow risk > 45. Sometimes, it can be further determined according to the actual situation.

[0116] Preferably, step S42 includes the following steps:

[0117] ECG risk values ​​and blood flow risk values ​​are generated for the ECG risk sub-trajectory and the blood flow risk sub-trajectory, respectively.

[0118] When the ECG risk value is greater than 80% or the blood flow risk value is greater than 75%, it is determined to be the first boundary, the first level clinical alarm is activated, and local audio and visual warnings are issued.

[0119] When the ECG risk value is greater than 80% and the blood flow risk value is greater than 75%, it is determined to be the second boundary. On the basis of activating the first-level clinical alarm, the second-level clinical alarm is triggered simultaneously, and a risk location report is pushed to the medical staff.

[0120] In one implementation of this invention, risk values ​​within a 30-second sliding window are calculated for both the electrocardiogram risk sub-trajectory (time series) and the blood flow risk sub-trajectory (time series):

[0121] Calculate the risk value within a 30-second sliding window for both the ECG risk sub-trajectory (time series) and the blood flow risk sub-trajectory (time series); Blood flow risk value = the maximum value of the blood flow risk score within the window (e.g., [65,72,78,70] → take 78%).

[0122] Then, a double out-of-bounds check is performed:

[0123] Level 1 alarm: If the ECG risk value is >80% or the blood flow risk value is >75% (e.g., ECG value 82 > 80), immediately activate the local audible and visual alarm (the device triggers a 5Hz buzzer + red LED flashing); Level 2 alarm: If the ECG risk value is >80% and the blood flow risk value is >75% (e.g., ECG value 82 > 80 and blood flow value 78 > 75), in addition to the audible and visual alarm, simultaneously push a JSON report containing key data (including the over-limit timestamp, risk value, and original signal fragment) to the medical staff.

[0124] The specific steps are as follows, and the input data is:

[0125] ECG risk sequence: [78,81,85,82,79]; Blood flow risk sequence: [72,76,78,74,73];

[0126] Risk value calculation: ECG risk value = 90th percentile (after sorting [78,79,81,82,85] → 5th value 85); Blood flow risk value = maximum value (78);

[0127] Transboundary judgment:

[0128] ECG risk value 85% > 80% → First condition met;

[0129] Blood flow risk value 78% > 75% → First condition met;

[0130] Simultaneously meeting the conditions of ECG > 80% and blood flow > 75% → triggering a level 2 alarm.

[0131] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0132] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for information interaction control of a vital sign monitor, characterized in that, The method comprises the following steps: Step S1: synchronously collecting physiological signals through a sensor array, wherein the physiological signals comprise electrocardiogram waveform data, photoplethysmogram waveform data, motion acceleration data, and body surface temperature data; Step S2: time-aligning the electrocardiogram waveform data, the photoplethysmogram waveform data, and the motion acceleration data based on a clock synchronization protocol, and performing matrix construction to generate a synchronous vital sign matrix; Step S3: calculating a motion divergence value based on the synchronous vital sign matrix and the motion acceleration data, selecting a processing channel based on the motion divergence value, and performing a filtering operation to obtain filtered data; Step S3 comprises the following steps: Step S31: performing three-axis acceleration analysis based on the synchronous vital sign matrix and the motion acceleration data, and calculating a motion divergence value; Step S31 comprises the following: calculating three-axis acceleration based on the synchronous vital sign matrix, and calculating a mean variance within a time window to obtain a dynamic variance; extracting an acceleration vector from the synchronous vital sign matrix, and calculating a spectral entropy value of the extracted acceleration vector in a 0.5-10 Hz frequency band; combining the dynamic variance and the spectral entropy value to obtain the motion divergence value; Step S32: selecting a processing channel based on the motion divergence value; when the motion divergence value is greater than a preset threshold value, it is judged that the state is a motion state, and a first real-time filtering channel is activated; when the motion divergence value is less than or equal to the preset threshold value, it is judged that the state is a static state, and a second real-time filtering channel is activated; Step S33: performing a filtering operation through the first real-time filtering channel or the second real-time filtering channel to obtain filtered data; Step S33 specifically comprises: taking the motion acceleration data as a noise reference signal, inputting the electrocardiogram waveform data into the first real-time filtering channel to perform reverse elimination of electrocardiogram waveform jitter noise, and generating a first purified electrocardiogram signal; inputting the photoplethysmogram waveform data into the first real-time filtering channel to decompose into an alternating component and a direct current component, performing motion interference suppression on the alternating component, and then performing component recombination with the direct current component to obtain a first purified pulse wave signal; or, inputting the electrocardiogram waveform data into the second real-time filtering channel to apply a fixed frequency window filter, retaining a 0.5-40 Hz clinical effective wave band, and directly outputting a second purified electrocardiogram signal; inputting the photoplethysmogram waveform data into the second real-time filtering channel to perform smoothing filtering, and generating a second purified pulse wave signal; combining the first purified electrocardiogram signal and the first purified pulse wave signal, or the second purified electrocardiogram signal and the second purified pulse wave signal, into filtered data; Step S4: performing risk change trajectory construction of a pathological trajectory on the filtered data and historical case data to generate a vital sign risk trajectory, and performing out-of-bound judgment on the vital sign risk trajectory; when the vital sign risk trajectory exceeds a threshold value, triggering a clinical alarm.

2. The information interaction control method of the vital sign monitor according to claim 1, wherein, Step S2 comprises: performing normalization processing on the electrocardiogram waveform data to extract a first synchronous time sequence; performing the same time calibration on the photoplethysmogram waveform data to extract a second synchronous time sequence; performing clock compensation on the original time stamp of the motion acceleration data to extract a third synchronous time sequence; The first, second, and third synchronized time series are subjected to the same interpolation and resampling to generate the first aligned ECG sequence, the second aligned PPG sequence, and the third aligned acceleration sequence, respectively. A synchronized vital signs matrix was constructed using the first aligned ECG sequence, the second aligned PPG sequence, and the third aligned accelerometer sequence.

3. The information interaction control method of the vital sign monitor according to claim 2, wherein, The mathematical expression for the synchronized vital signs matrix includes: wherein, is a synchronized vital signs matrix, is a timestamp, is a first aligned electrocardiogram sequence, is a second aligned PPG sequence, is a third aligned acceleration sequence, is an X-axis motion acceleration, is a Y-axis motion acceleration, is a Z-axis motion acceleration.

4. The information interaction control method of vital sign monitors according to claim 1, characterized in that, Step S4 includes the following: Step S41: Perform risk change trajectory construction on the filtered data and historical case data to generate vital sign risk trajectories, which include ECG risk sub-trajectories and blood flow risk sub-trajectories; Step S42: Perform boundary judgment based on ECG risk sub-trajectory and blood flow risk sub-trajectory. When the vital sign risk trajectory exceeds the threshold, trigger a clinical alarm to obtain vital sign alarm data. Step S43: Perform full-process tracing based on vital sign alarm data and mark key nodes of the detector; generate a report using the key nodes of the detector to obtain a full-process report of the vital sign detector.

5. The information interaction control method of vital sign monitor according to claim 4, characterized in that, The risk change trajectory construction for pathological trajectories on filtered data and historical case data includes the following steps: Acquire historical case data; overlay body surface temperature data onto filtered data to generate a fused vital signs sequence; Electrocardiogram (ECG) features and blood flow features were extracted from the fused vital sign sequences. ECG features and historical case data are analyzed to perform ECG feature sequence analysis and output cardiovascular risk scores. Temporal analysis of cardiac electrical activity is performed on the cardiovascular risk scores, and the risk change trajectory is constructed based on the timestamp as the horizontal axis and the predicted value of the cardiovascular risk score as the vertical axis to obtain the ECG risk sub-trajectory. Blood flow characteristics and historical case data were analyzed for blood pressure, and the SpO2-PWV organ parameters were used to quantify the risk score of circulatory failure. The circulatory failure risk score was analyzed in a time series of the circulatory system, and the risk change trajectory was constructed by using the timestamp as the horizontal axis and the predicted value of the circulatory failure risk score as the vertical axis, thus obtaining the blood flow risk sub-trajectory.

6. The information interaction control method of vital sign monitors according to claim 4, characterized in that, Step S42 includes the following steps: ECG risk values ​​and blood flow risk values ​​are generated for the ECG risk sub-trajectory and the blood flow risk sub-trajectory, respectively. When the ECG risk value is greater than 80% or the blood flow risk value is greater than 75%, it is determined to be the first boundary, the first level clinical alarm is activated, and local audio and visual warnings are issued. When the ECG risk value is greater than 80% and the blood flow risk value is greater than 75%, it is determined to be the second boundary. On the basis of activating the first-level clinical alarm, the second-level clinical alarm is triggered simultaneously, and a risk location report is pushed to the medical staff.

7. An information interaction control system of a vital sign monitor, characterized in that, The information interaction control method for executing the vital signs monitor as described in claim 1 includes: The multi-source synchronous acquisition module is used to synchronously acquire physiological signals through a sensor array, including electrocardiogram waveform data, pulse wave data, motion acceleration data, and body surface temperature data. The timing alignment and feature matrix construction module is used to perform time alignment on ECG waveform data, capacitive pulse wave data and motion acceleration data based on the clock synchronization protocol, and to construct a matrix to generate a synchronized vital signs matrix. a motion interference perception and filtering channel selection module, configured to calculate a motion divergence value according to the synchronized vital sign matrix and the motion acceleration data, perform a filtering operation, and select a processing channel based on the motion divergence value to obtain filtered data; a pathological evolution analysis and clinical early warning module, configured to perform risk change trajectory construction of a pathological trajectory on the filtered data and historical case data, generate a vital sign risk trajectory, and perform out-of-bound judgment on the vital sign risk trajectory; when the vital sign risk trajectory exceeds a threshold value, a clinical alarm is triggered.

Images