Body condition monitoring method and wristband emergency monitoring ring

Abstract

The application discloses a body state monitoring method and a wristband type emergency monitoring ring, relates to the technical field of biological signal processing, and the body state monitoring method comprises the following steps: emitting red light of a preset first wavelength to the wrist part of a human body, collecting reflected light signals, obtaining a waveform physiological signal, determining a monitoring index of the human body, inputting the monitoring index into a warning model, obtaining a stress state probability of the human body, judging whether a warning is triggered or not based on the stress state probability and a preset probability threshold, and if the warning is triggered, issuing an alarm and sending corresponding body state data to a preset medical platform. The wristband type emergency monitoring ring can be worn on the wrist of a player to realize convenient detection, and through real-time monitoring of physiological data of the human body and a warning model, stress state detection and prediction are realized, so that the body state of the player can be monitored conveniently, in real time and accurately, and the effect of body monitoring of the player is improved.

Description

Technical Field

[0001] This application relates to the field of biosignal processing technology, and in particular to methods for monitoring body condition and wristband-type emergency monitoring rings. Background Technology

[0002] Currently, in short-term immersive entertainment scenarios such as escape rooms, venues have almost no professional safety measures. If a player suddenly experiences physical discomfort such as palpitations or shortness of breath in the enclosed environment of the escape room, no one may be able to notice in time. Therefore, there is a need for a method and equipment that can monitor the player's physical condition in real time.

[0003] Among current methods for monitoring physical condition, ordinary smart wearable devices cannot meet the requirements of real-time and accurate monitoring of players' physiological data and abnormal warnings in specific scenarios, resulting in low reliability. Other physical condition monitoring devices are bulky and inconvenient, which can affect the range of motion and immersive experience of players in intense games. Therefore, the current methods for monitoring players' physical condition are not very effective. Summary of the Invention

[0004] The main purpose of this application is to provide a method for monitoring physical condition and a wristband-type emergency monitoring ring, aiming to solve the technical problem of poor effectiveness in monitoring the physical condition of players.

[0005] To achieve the above objectives, this application proposes a method for monitoring physical condition, the method comprising: A preset first wavelength of red light is emitted towards the wrist of the human body, and the reflected light signal is collected to obtain a waveform physiological signal; Based on the waveform physiological signal, the monitoring indicators of the human body are determined, and the monitoring indicators are input into the preset early warning model to obtain the probability of the human body's stress state. The early warning model is obtained by training the model to be trained based on training samples labeled with stress state judgment results. Based on the stress state probability and the preset probability threshold, determine whether to trigger an early warning; If an alert is triggered, an alarm will be issued and the corresponding physical condition data will be sent to a pre-set medical platform.

[0006] In one embodiment, the step of determining the monitoring indicators of the human body based on the waveform physiological signal includes: Based on a preset noise cancellation algorithm, noise is removed from the waveform physiological signal to obtain a denoised physiological signal; Based on the denoised physiological signal, the pulsation component, DC component, and waveform feature points are determined. The monitoring index is determined based on the pulsation component, the DC component, and the waveform feature points.

[0007] In one embodiment, the step of removing noise from the waveform physiological signal based on a preset noise cancellation algorithm to obtain a denoised physiological signal includes: Acquire human motion data and use the motion data as reference noise; Based on a preset filter, the effective signal of the target frequency is selected from the waveform physiological signal, and the signals other than the effective signal are removed to obtain the heart rate signal; Based on the reference noise, the noise corresponding to the reference noise is removed from the heart rate signal to obtain the denoised physiological signal.

[0008] In one embodiment, the waveform feature points include systolic peak values, and the monitoring indicators include real-time heart rate, heart rate variability, and perfusion index. The step of determining the monitoring indicators based on the pulsatility component, the DC component, and the waveform feature points includes: Obtain the peak time of each of the contraction peaks, calculate the difference between the peak times, and obtain the peak time difference; The real-time heart rate is determined based on the peak time difference; Calculate the standard deviation of each of the peak time differences to obtain the peak time standard deviation, and determine the heart rate variability based on the peak time standard deviation; The perfusion index is obtained by calculating the ratio of the pulsatile component to the DC component.

[0009] In one embodiment, the monitoring indicator is a monitoring indicator within a sliding window of a preset time length, the stress state probability is the stress state probability corresponding to the sliding window, and the step of determining whether to trigger an alert based on the stress state probability and a preset probability threshold includes: Determine whether the probability of the stress state corresponding to the current sliding window reaches the probability threshold; If this is achieved, it is determined that the current sliding window has detected a stress state; Obtain the detection results of the stress state corresponding to a preset number of historical sliding windows before the current sliding window; If the stress state is detected in all the historical sliding windows, an early warning is triggered.

[0010] In one embodiment, before the step of emitting red light of a preset first wavelength to the wrist of the human body, collecting the reflected light signal, and obtaining the waveform physiological signal, the method further includes: Obtain the training samples, test samples, and multiple preset candidate probability thresholds; The training samples are input into the model to be trained to obtain the prediction results of the model to be trained; Based on the prediction results and the stress state judgment results, the model to be trained is trained to obtain the early warning model; The probability threshold is determined based on the test sample, the candidate probability threshold, and the early warning model.

[0011] In one embodiment, the test sample includes the stress state judgment result corresponding to the test sample, and the step of determining the probability threshold based on the test sample, the candidate probability threshold, and the early warning model includes: The test sample is input into the early warning model to obtain the probability of the stress state; Based on the stress state probability and the candidate probability threshold, the predicted stress judgment result is determined; Based on the predicted stress judgment result and the stress state judgment result, the false positive rate corresponding to each candidate probability threshold is determined; The average false positive rate of each candidate probability threshold tested over multiple rounds; The probability threshold is the candidate probability threshold that has the lowest average false positive rate.

[0012] In one embodiment, after the step of emitting red light of a preset first wavelength to the wrist of the human body, collecting the reflected light signal, and obtaining the waveform physiological signal, the method further includes: A preset second wavelength of infrared light is emitted towards the wrist of the human body, and the reflected light signal is collected to obtain the infrared light physiological signal. Based on the infrared physiological signals, the infrared pulsation component and the infrared DC component are extracted; Calculate the ratio of the pulsating component to the DC component to obtain a first ratio, and calculate the ratio of the infrared light pulsating component to the infrared light DC component to obtain a second ratio; The blood oxygen saturation of the human body is calculated based on the first ratio, the second ratio, and a preset empirical coefficient.

[0013] Furthermore, to achieve the above objectives, this application also proposes a wristband-type emergency monitoring ring, which includes: The red light emitting module is used to emit red light of a preset first wavelength towards the human wrist. Sensors are used to collect reflected light signals to obtain waveform physiological signals; The main control module is used to determine the monitoring indicators of the human body based on the waveform physiological signal, input the monitoring indicators into the preset early warning model to obtain the stress state probability of the human body, and determine whether to trigger an early warning based on the stress state probability and the preset probability threshold. If an early warning is triggered, an alarm is issued and the corresponding physical status data is sent to the preset medical platform.

[0014] In one possible embodiment of this application, the wristband-type emergency monitoring ring further includes: An infrared light emitting module is used to emit a preset second wavelength of red light toward the wrist of the human body; The sensor is also used to collect reflected light signals to obtain infrared light physiological signals; The main control module is also used to extract the infrared pulsation component and the infrared DC component based on the infrared physiological signal, calculate the ratio of the pulsation component to the DC component to obtain a first ratio, calculate the ratio of the infrared pulsation component to the infrared DC component to obtain a second ratio, and calculate the blood oxygen saturation of the human body based on the first ratio, the second ratio and a preset empirical coefficient.

[0015] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the body state monitoring method described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: This application emits a preset first wavelength of red light onto the wrist of a human body, collects the reflected light signal, obtains a waveform physiological signal, determines the human body's monitoring indicators based on the waveform physiological signal, inputs the monitoring indicators into a preset early warning model, and obtains the probability of the human body's stress state. The early warning model is trained based on training samples labeled with stress state judgment results. Based on the stress state probability and a preset probability threshold, it is determined whether an early warning is triggered. If an early warning is triggered, an alarm is issued and the corresponding physical state data is sent to a preset medical platform.

[0017] Compared to current methods that fail to provide real-time, accurate monitoring and anomaly warnings for players' physiological data in specific scenarios, suffer from low reliability, and are bulky and inconvenient, resulting in poor effectiveness of player health monitoring, this application utilizes a wristband-style emergency monitoring ring to collect and analyze waveform physiological signals from the human body in real time. Based on the analysis results, it issues alarms, thus improving the effectiveness of player health monitoring. Specifically, the wristband-style emergency monitoring ring can be conveniently worn on the player's wrist for monitoring. It monitors physiological data in real time by emitting red light and receiving light signals. By inputting the real-time determined monitoring indicators into a trained warning model, it detects the player's stress state and sends an alarm when stress is detected. Therefore, this application can conveniently, in real-time, and accurately monitor the player's physical condition, thereby improving the effectiveness of player health monitoring. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an embodiment of the body condition monitoring method of this application. Figure 2 This is a schematic diagram of the wristband-type emergency monitoring ring structure provided in Embodiment 1 of the body status monitoring method of this application; Figure 3 This is a schematic diagram of feature extraction provided in Embodiment 1 of the body status monitoring method of this application; Figure 4 This is a flowchart illustrating Embodiment 2 of the body condition monitoring method of this application; Figure 5 This is a schematic diagram illustrating the data acquisition consent process involved in the body condition monitoring method in this application embodiment.

[0021] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0022] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0023] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0024] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or body status monitoring device capable of performing the above functions. The following description uses a body status monitoring device as an example to illustrate this embodiment and the subsequent embodiments.

[0025] Currently, in short-term immersive entertainment scenarios such as escape rooms, venues have almost no professional safety measures. If a player suddenly experiences physical discomfort such as palpitations or shortness of breath in the enclosed environment of the escape room, no one may be able to notice in time. Therefore, there is a need for a method and equipment that can monitor the player's physical condition in real time.

[0026] Currently, ordinary smart wearable devices cannot meet the requirements of real-time and accurate monitoring of players' physiological data and abnormal warnings in specific scenarios, resulting in low reliability. Other body condition monitoring devices are bulky and inconvenient, which can affect the range of motion and immersive experience of players in intense games. Therefore, the current methods for monitoring players' bodies are not very effective.

[0027] Furthermore, some users have occasionally come into contact with smart bracelets or watches, but have never used professional medical monitoring devices. When monitoring their physical condition, users would prefer devices that do not require additional learning and can be worn and used immediately upon receipt. Based on this, this application provides a method for monitoring body status, applied to a wristband-type emergency monitoring ring, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the body condition monitoring method of this application.

[0028] In this embodiment, the body condition monitoring method includes steps S10 to S40: Step S10: Emit a preset first wavelength of red light to the wrist of the human body, collect the reflected light signal, and obtain the waveform physiological signal; It should be noted that the preset first wavelength refers to the specific wavelength value of the emitted light signal pre-set during system design. Red light refers to visible light with a wavelength range of approximately 620nm to 750nm, specifically referring to light signals with wavelengths in the red light band. The reflected light signal refers to the light signal emitted by the transmitter, irradiating human tissue (including blood, skin, etc.), and then scattered and reflected back to be detected by the receiver. The waveform physiological signal refers to waveform data that changes over time and reflects physiological activities (such as the periodic change of blood volume in blood vessels with the pulse), typically in the form of photoplethysmography waves.

[0029] It should also be noted that the wristband-type emergency monitoring ring in this embodiment includes a main control module: responsible for processing, calculating, and wirelessly transmitting physiological data; and a blood oxygen sensor module: used to measure blood oxygen based on photoplethysmography, providing accurate raw PPG (Photoplethysmography) signals for calculating heart rate, HRV (Heart Rate Variability), and PI (Perfusion Index). The wristband-type emergency monitoring ring in this embodiment adopts a wristband structure, and the overall design of its components emphasizes low power consumption to meet the battery life requirements for short-term wear. The specific structure can be found in [reference needed]. Figure 2 .

[0030] It is understood that this embodiment uses a method of emitting red light of a preset first wavelength to the wrist of the human body and collecting the reflected light signal. It takes advantage of the fact that red light is sensitive to the absorption characteristics of hemoglobin in the blood, and the skin of the wrist is thin and the blood vessels are shallow, so that the reflected light signal is rich in pulse components. This allows for the acquisition of waveform physiological signals with a high signal-to-noise ratio, providing accurate basic data for subsequent physiological parameter calculations.

[0031] Step S20: Based on the waveform physiological signal, determine the monitoring indicators of the human body, input the monitoring indicators into the preset early warning model, and obtain the probability of the human body's stress state. The early warning model is obtained by training the model to be trained based on training samples labeled with stress state judgment results. It should be noted that monitoring indicators refer to characteristic parameters extracted from waveform physiological signals that can quantitatively reflect the physiological state of the human body, such as heart rate, heart rate variability, and blood oxygen saturation. The preset early warning model refers to a computational model pre-built and stored during system design, used to calculate the output probability based on the input monitoring indicators. The stress state probability refers to the numerical value indicating the likelihood that the human body is currently in a stress state (such as tension, stress, excitement, etc.), usually represented by a value between 0 and 1. Training samples labeled with stress state judgment results refer to sample data used to train the model; each sample contains a set of monitoring indicators and the actual stress state category corresponding to that sample, manually or experimentally labeled. The model to be trained refers to the original model whose initial parameters have not been optimized and which needs to be trained to learn the mapping relationship between input and output.

[0032] It should also be noted that this embodiment first extracts corresponding features based on the waveform physiological signal, and then performs subsequent processing based on the extracted features. The feature extraction based on the waveform physiological signal in this embodiment can be referred to... Figure 3 , Figure 3 The waveform diagram refers to a preprocessed and normalized PPG signal segment. The f-plot is a scatter plot or performance evaluation plot used to demonstrate the performance of the classification model. ACF (Autocorrelation Function), FFT (Fast Fourier Transform), SE (Spectral Entropy), classification, normalized amplitude, Sinus rhythm, and AF (Atrial Fibrillation) are all represented in PPG signals as highly irregular and chaotic waveform amplitudes and pulse intervals.

[0033] It is understood that this embodiment extracts monitoring indicators based on waveform physiological signals and inputs the monitoring indicators into an early warning model that has been trained in advance using labeled samples. This model has learned the statistical laws between the monitoring indicators and the stress state, thereby automatically outputting a quantified stress state probability to improve the objectivity and accuracy of stress state assessment.

[0034] In one feasible implementation, the specific implementation of determining the monitoring indicators of the human body based on the waveform physiological signal can also be: Based on a preset noise cancellation algorithm, noise is removed from the waveform physiological signal to obtain a denoised physiological signal. Based on the denoised physiological signal, the pulsation component, DC component, and waveform feature points are determined. Based on the pulsation component, the DC component, and the waveform feature points, the monitoring index is determined.

[0035] It should be noted that the preset noise cancellation algorithm refers to the mathematical calculation method determined in advance during system design, used to identify and filter out noise components from the original signal, such as filtering algorithms, wavelet denoising, or adaptive filtering. The denoised physiological signal refers to the pure physiological signal obtained after processing the original waveform physiological signal through the noise cancellation algorithm, where noise components are significantly suppressed. The pulsating component refers to the AC component in the waveform physiological signal that reflects the periodic changes in vascular volume caused by heartbeats, usually manifested as a periodic fluctuation. The DC component refers to the portion of the waveform physiological signal that reflects the constant or slow changes in light absorption by non-pulsatile tissues (such as skin, muscles, and bones), representing the baseline level of the signal. Waveform feature points refer to specific physiologically significant locations within a single pulse waveform of the denoised physiological signal, such as peaks, troughs, and dicrotic notches.

[0036] It is understood that this implementation first uses a preset noise reduction algorithm to remove noise from the original waveform physiological signal to obtain a denoised physiological signal. Then, based on the denoised signal, the pulsation component, DC component, and waveform feature points are extracted respectively. These components and feature points directly reflect hemodynamics and tissue optical absorption characteristics. The removal of noise avoids interference factors from contaminating the feature extraction, thereby making the finally determined monitoring indicators more accurate and reliable.

[0037] In one feasible implementation, the specific implementation of removing noise from the waveform physiological signal based on a preset noise cancellation algorithm to obtain a denoised physiological signal can also be: Human motion data is acquired, and the motion data is used as reference noise. Based on a preset filter, the effective signal of the target frequency is filtered out from the waveform physiological signal, and the signals other than the effective signal are removed to obtain the heart rate signal. Based on the reference noise, the noise corresponding to the reference noise is removed from the heart rate signal to obtain the denoised physiological signal.

[0038] It should be noted that human motion data refers to motion state information of the human limbs or torso collected by sensors, such as acceleration and angular velocity. Reference noise refers to the signal used as a reference input in adaptive noise cancellation; here, motion data is used as reference noise to represent the interference components caused by motion. Preset filters refer to signal processing modules with specific frequency response characteristics predetermined during system design, such as bandpass filters and low-pass filters. Effective signals at the target frequency refer to signal components in the waveform physiological signal that fall within a preset target frequency range, corresponding to the possible frequency range of the human heart rate. The heart rate signal refers to the periodic signal obtained after preliminary screening, mainly containing heart rate cycle information; it is an intermediate form of the denoised physiological signal.

[0039] Understandably, this implementation uses human motion data as reference noise. First, a preset filter is used to select the effective signal of the target frequency from the waveform physiological signal to obtain the heart rate signal. Then, based on the reference noise, the motion interference corresponding to it is removed from the heart rate signal. The motion data directly reflects the characteristics of the interference source, making noise elimination adaptive and targeted. This effectively suppresses the contamination of physiological signals by motion artifacts, resulting in a purer denoised physiological signal, which lays the foundation for the accurate extraction of subsequent pulsation components and waveform feature points.

[0040] In one feasible implementation, the waveform feature points include systolic peak values, and the monitoring indicators include real-time heart rate, heart rate variability, and perfusion index. The specific implementation of determining the monitoring indicators based on the pulsating component, the DC component, and the waveform feature points can also be: The peak time of each of the systolic peaks is obtained, the difference between each peak time is calculated to obtain the peak time difference, the real-time heart rate is determined based on the peak time difference, the standard deviation of each peak time difference is calculated to obtain the peak time standard deviation, the heart rate variability is determined based on the peak time standard deviation, and the ratio of the pulsatile component to the DC component is calculated to obtain the perfusion index.

[0041] It should be noted that systolic peak refers to the highest point of the waveform in a complete cardiac cycle, reflecting the moment of maximum intravascular blood volume during cardiac systole. Peak time refers to the time value corresponding to each systolic peak on the time axis. Peak time difference refers to the difference between the peak times of two adjacent systolic peaks, i.e., the time interval between two consecutive heartbeats. Real-time heart rate refers to the number of heartbeats per unit time, reflecting the current heart rate frequency. Heart rate variability refers to the small temporal fluctuations between successive heartbeat cycles, usually quantified by measuring the standard deviation of consecutive peak time differences. The perfusion index is the ratio of the pulsatile component to the direct current component, reflecting the level of peripheral tissue blood flow perfusion.

[0042] It is understood that this implementation calculates the peak time difference based on the peak time of the systolic peak, and then determines the real-time heart rate through the peak time difference. The standard deviation of the peak time difference is used to determine the heart rate variability, and the perfusion index is determined by the ratio of the pulsatile component to the DC component. All these calculations are based on the same set of denoised physiological signals, without the need for additional sensors or data acquisition, thereby achieving efficient and synchronous acquisition of multiple dimensions of monitoring indicators from a single waveform physiological signal.

[0043] Step S30: Based on the stress state probability and the preset probability threshold, determine whether to trigger an early warning; It should be noted that the probability threshold refers to a pre-set probability cutoff value used to convert continuous probability outputs into discrete judgment results. An alert is triggered when the probability of a stress state is greater than or equal to this threshold; otherwise, it is not triggered. An alert is a signal issued by the system when the likelihood of a person being in a stress state exceeds an acceptable range, reminding the user or relevant parties to take appropriate measures.

[0044] It is understood that this embodiment judges based on the comparison between the probability of the stress state and the preset probability threshold. The probability threshold provides an adjustable decision boundary, which enables the system to flexibly control the sensitivity and specificity according to the actual application scenario. This avoids frequent false alarms due to slight fluctuations in probability and can issue timely warnings when the probability reaches a dangerous level, thus realizing a reliable and configurable warning triggering mechanism.

[0045] In step S40, if an alert is triggered, an alarm is issued and the corresponding physical status data is sent to a preset medical platform.

[0046] It should be noted that an alarm refers to a warning signal issued by the system after triggering a warning. It can take the form of sound, flashing light, vibration, screen prompts, etc., to attract the attention of users or supervisors.

[0047] The corresponding physical condition data refers to the physiological monitoring data related to this warning, including at least one or more of the following: waveform physiological signals, monitoring indicators, and stress state probabilities. The pre-defined medical platform refers to a remote server or medical service system pre-designated during system design for receiving and storing user physical condition data.

[0048] It is understood that this embodiment issues an alarm and sends physical status data to a preset medical platform when the warning is triggered. The alarm can remind the user or nearby people to take countermeasures in the first instance, while the remote data upload enables the medical platform to record, analyze or notify medical staff for subsequent intervention, thereby improving the timeliness of response to stress events and the accessibility of medical care.

[0049] In summary, this embodiment emits a preset first wavelength of red light to the wrist of the human body, collects the reflected light signal, obtains a waveform physiological signal, determines the monitoring indicators of the human body based on the waveform physiological signal, and inputs the monitoring indicators into a preset early warning model to obtain the probability of the human body's stress state. The early warning model is trained based on training samples labeled with stress state judgment results. Based on the stress state probability and a preset probability threshold, it is determined whether an early warning is triggered. If an early warning is triggered, an alarm is issued and the corresponding physical state data is sent to a preset medical platform.

[0050] Compared to current methods that fail to provide real-time, accurate monitoring and anomaly warnings for players' physiological data in specific scenarios, suffer from low reliability, and are bulky and inconvenient, resulting in poor effectiveness of player health monitoring, this embodiment utilizes a wristband-style emergency monitoring ring to collect and analyze waveform physiological signals from the human body in real time. Based on the analysis results, it issues alarms, significantly improving the effectiveness of player health monitoring. Specifically, the wristband-style emergency monitoring ring can be conveniently worn on the player's wrist for monitoring. This embodiment uses red light emission and light signal reception to monitor physiological data in real time. By inputting the real-time determined monitoring indicators into a trained warning model, it detects the player's stress state and sends an alarm when stress is detected. Therefore, this embodiment can conveniently, in real-time, and accurately monitor the player's physical condition, thereby improving the effectiveness of player health monitoring.

[0051] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 4 The monitoring index is a monitoring index within a sliding window of a preset time length, and the stress state probability is the stress state probability corresponding to the sliding window. Step S30, the body state monitoring method further includes steps S31~S34: Step S31: Determine whether the probability of the stress state corresponding to the current sliding window reaches the probability threshold; It's important to note that a sliding window refers to a fixed-length time interval defined over continuous time series data. This interval slides forward over time, and after each slide, the window contains the latest data segment. A sliding window with a preset time length refers to a time interval predetermined during system design, such as 30 seconds or 1 minute, used to extract data within that time period for analysis. The current sliding window refers to the latest time interval covered by the sliding window at the current moment.

[0052] It is understood that this embodiment determines whether the probability of the stress state corresponding to the current sliding window has reached the probability threshold, and both the monitoring indicators and the stress state probability are calculated based on a sliding window of fixed time length. This enables the system to track the dynamic changes of human stress state in a continuous and segmented manner. The evaluation results of each window are independent and real-time, thereby avoiding misjudgment caused by fluctuations in single-point data and improving the stability and timeliness of stress state monitoring.

[0053] Step S32: If the condition is met, it is determined that the current sliding window has detected a stress state. It should be noted that detecting a stress state means that within the time interval of the current sliding window, the system determines that the human body is in a stress state, which serves as the basis for triggering warnings or recordings.

[0054] It is understood that in this embodiment, when the probability of a stress state reaches a probability threshold, the current sliding window is determined to be a detected stress state, so that the early warning decision is based on the stress state determination at the sliding window level, thereby avoiding frequent early warnings due to single probability fluctuations and improving the reliability and continuity of stress state identification.

[0055] Step S33: Obtain the detection results of the stress state corresponding to a preset number of historical sliding windows before the current sliding window; It should be noted that the preset window count refers to an integer value predetermined during system design, used to specify the number of historical sliding windows to be acquired. Historical sliding windows refer to one or more sliding windows that have been processed before the current sliding window, arranged in chronological order. The detection result refers to the conclusion obtained after judging each sliding window whether a stress state was detected.

[0056] It is understood that this embodiment obtains the stress state detection results of a preset number of historical sliding windows before the current sliding window. These historical results provide necessary information for judging the continuous pattern or trend of the current stress state, thereby enabling the system to make a more reliable comprehensive judgment based on multiple consecutive detection results in the time series, avoiding misjudgment or omission caused by relying solely on the instantaneous results of a single window.

[0057] Step S34: If all the historical sliding windows detect the stress state, then a warning is triggered.

[0058] It is understood that this embodiment requires all historical sliding windows before the current sliding window to detect a stress state before triggering an alert. This condition ensures that the stress state exists continuously over multiple consecutive time windows rather than fluctuating instantaneously, thereby effectively filtering out false triggers caused by occasional noise interference or brief physiological fluctuations, improving the reliability of alert triggering, and ensuring that the alert is only issued when the stress state has temporal continuity.

[0059] In one feasible implementation, the specific implementation prior to emitting a preset first wavelength of red light towards the wrist of the human body, collecting the reflected light signal, and obtaining the waveform physiological signal can also be: The training samples, test samples, and multiple preset candidate probability thresholds are obtained. The training samples are input into the model to be trained to obtain the prediction results of the model to be trained. Based on the prediction results and the stress state judgment results, the model to be trained is trained to obtain the early warning model. Based on the test samples, the candidate probability thresholds, and the early warning model, the probability thresholds are determined.

[0060] It should be noted that the test samples refer to the set of sample data used to evaluate model performance and help determine the optimal probability threshold, and are independent and non-overlapping with the training samples. Multiple preset candidate probability thresholds refer to a set of candidate critical values ​​pre-set during system design, from which the optimal probability threshold is selected. The prediction result refers to the stress state probability value output by the model to be trained for each sample in the training samples.

[0061] It is understood that, in this implementation method, before signal acquisition, a warning model is trained using training samples and annotation results, and the optimal probability threshold is selected using independent test samples and multiple candidate probability thresholds. Both training and threshold determination are data-driven, enabling the model to learn the true mapping pattern between monitoring indicators and stress states. At the same time, the selection of thresholds also adapts to the output distribution of the model and the actual application requirements, thereby improving the accuracy of stress state probability calculation and the reliability of warning trigger judgment in subsequent actual monitoring.

[0062] In one feasible implementation, the test sample includes the stress state judgment result corresponding to the test sample, and the specific implementation of determining the probability threshold based on the test sample, the candidate probability threshold, and the early warning model can also be: The test sample is input into the early warning model to obtain the stress state probability. Based on the stress state probability and the candidate probability threshold, the predicted stress judgment result is determined. Based on the predicted stress judgment result and the stress state judgment result, the false positive rate corresponding to each candidate probability threshold is determined. The average false positive rate of each candidate probability threshold in multiple rounds of testing is used as the probability threshold, and the candidate probability threshold with the lowest average false positive rate is used as the probability threshold.

[0063] It should be noted that the predicted stress judgment result refers to the binary judgment result obtained by comparing the stress state probability output by the early warning model with a certain candidate probability threshold, and is used to compare with the actual stress state judgment result. The false positive rate refers to the proportion of samples that are incorrectly predicted as stress states by the model among all samples that are actually in a non-stress state; the calculation formula is the number of false positives divided by the total number of true negatives. The average false positive rate refers to the result obtained by arithmetically averaging the multiple false positive rate values ​​calculated from the same candidate probability threshold in multiple rounds of independent testing; it is used to evaluate the average false alarm level of that threshold across multiple tests.

[0064] It is understood that this implementation calculates the average false positive rate of each candidate probability threshold through multiple rounds of testing, and selects the threshold with the lowest average false positive rate as the probability threshold. The false positive rate directly reflects the severity of false alarms. Minimizing the average false positive rate means that the number of false alarms triggered by the warning is minimized in practical applications, thereby effectively reducing the interference to users and the waste of medical resources caused by false alarms, and improving the practicality of the warning system.

[0065] In one feasible implementation, the subsequent step of emitting a preset first wavelength of red light to the wrist of the human body, collecting the reflected light signal, and obtaining the waveform physiological signal can also be: A preset second wavelength of infrared light is emitted towards the wrist of the human body, and the reflected light signal is collected to obtain an infrared physiological signal. Based on the infrared physiological signal, the pulsating component and the DC component of the infrared light are extracted, and the ratio of the pulsating component to the DC component is calculated to obtain a first ratio. The ratio of the pulsating component to the DC component of the infrared light is calculated to obtain a second ratio. Based on the first ratio, the second ratio, and a preset empirical coefficient, the blood oxygen saturation of the human body is calculated.

[0066] It should be noted that the preset second wavelength refers to a different optical signal wavelength value than the first wavelength, which is pre-set during system design. Infrared light refers to invisible light with a wavelength range of approximately 750nm to 1000nm. Here, it specifically refers to optical signals with wavelengths belonging to the infrared light band, which are typically sensitive to blood oxygen saturation detection. Infrared physiological signal refers to the waveform physiological signal obtained by collecting the reflected light signal after emitting infrared light to the human wrist and then converting it through photoelectric conversion.

[0067] The pulsating component of infrared light refers to the AC component extracted from infrared photophysiological signals, reflecting the periodic changes in vascular volume caused by cardiac pulsation. The DC component of infrared light refers to the portion extracted from infrared photophysiological signals reflecting the constant or slow changes in the absorption of infrared light by non-pulsatile tissues. The first ratio refers to the ratio of the pulsating component to the DC component of the red light signal, i.e., the perfusion index calculated earlier. The second ratio refers to the ratio of the pulsating component to the DC component of the infrared light signal, i.e., the perfusion index corresponding to the infrared light.

[0068] The preset empirical coefficient refers to a constant or formula parameter predetermined during system design, used to convert the ratio of two wavelengths into blood oxygen saturation. Blood oxygen saturation refers to the percentage of oxyhemoglobin in the total hemoglobin in the blood, and is an important physiological indicator reflecting the body's respiratory and circulatory functions.

[0069] Understandably, this implementation method, in addition to acquiring red light signals, emits a second wavelength of infrared light and acquires infrared physiological signals. Then, it extracts the pulsation component and DC component of the two wavelengths and calculates their respective ratios. Combined with preset empirical coefficients, it calculates blood oxygen saturation. By utilizing the difference in absorption characteristics of oxyhemoglobin and deoxyhemoglobin to different wavelengths of light, it can simultaneously acquire the key physiological indicator of blood oxygen saturation from the same wrist without adding additional sensors. This enriches the dimensions of the monitoring parameters and provides a more comprehensive basis for the overall assessment of human health.

[0070] In one embodiment, the step of determining to trigger an early warning if all the historical sliding windows detect the stress state includes: Obtain the stress state probability value corresponding to each of the preset number of historical sliding windows before the current sliding window, calculate the amount by which the stress state probability of each historical sliding window exceeds the probability threshold, obtain the exceedance amplitude value of each window, calculate the average amplitude of the exceedance amplitude values ​​of all historical sliding windows, determine whether the average amplitude reaches the preset intensity triggering threshold, and if it does, determine to trigger an early warning.

[0071] Understandably, the excess amount refers to the difference between the probability of the stress state in each historical sliding window and the probability threshold, reflecting the degree to which the stress state in that window exceeds the minimum trigger threshold. The excess amplitude value is the excess amount for each window, which can be positive or zero. The average amplitude is the arithmetic mean obtained by summing the excess amplitude values ​​of all historical sliding windows and dividing by the number of windows, used to characterize the overall intensity level of continuous stress states. The preset intensity trigger threshold refers to a pre-set amplitude threshold value of the system; an alert is only triggered when the average amplitude reaches or exceeds this value.

[0072] Understandably, this embodiment requires all historical sliding windows to detect stress (i.e., the probability of all windows exceeding the threshold), and further calculates the average magnitude of the excess in each window. The warning is only triggered when the average magnitude reaches the intensity trigger threshold. This mechanism not only eliminates single, occasional fluctuations, but also filters out marginal cases where the probability of stress is only slightly higher than the threshold but the intensity is weak and the actual physiological significance is not great. This significantly reduces invalid warnings caused by physiological baseline drift or weak stress response, and improves the clinical reference value of the warning.

[0073] In one embodiment, the step of determining whether the average amplitude reaches a preset intensity trigger threshold, and if so, determining to trigger an early warning, further includes: Obtain the baseline value of the average amplitude of a preset number of historical sliding windows before the current sliding window when the user is in a non-stressed state; calculate the ratio of the currently calculated average amplitude to the baseline value to obtain the relative stress intensity; determine whether the relative stress intensity reaches a preset relative intensity trigger threshold, and simultaneously determine whether the average amplitude of multiple consecutive historical sliding windows shows an increasing trend and the increasing slope reaches a preset rising slope threshold. If both conditions are met, then determine to trigger an alert.

[0074] Understandably, the non-stress state refers to the time period during which no stress state was detected in the historical sliding windows, based on the judgment results; that is, the user is in a calm or non-stressful physiological state. The baseline value refers to the typical value (mean or median) of the average amplitude of multiple consecutive historical sliding windows obtained by the same calculation method when the user is in a non-stressful state, used to characterize the user's individualized normal stress intensity level.

[0075] Relative stress intensity refers to the ratio of the current average amplitude to the individual's baseline value, reflecting the multiple of the current stress intensity relative to the user's normal level. The increasing trend refers to the direction of change in the average amplitude values ​​of consecutive historical sliding windows arranged in chronological order. The rising slope threshold refers to a preset minimum slope value used to quantify whether the rate of increase in average amplitude over time reaches the trigger condition.

[0076] It is understood that this embodiment first uses historical data of users in non-stress states to establish an individualized average amplitude baseline value, and then calculates the ratio of the current average amplitude to the baseline to obtain the relative stress intensity, thereby eliminating the interference of physiological baseline differences between different users on the fixed threshold judgment, and making the warning triggering conditions adaptively adapt to the normal level of each user.

[0077] Furthermore, this embodiment requires that the relative stress intensity reaches a threshold and the average amplitude shows a continuous increasing trend, eliminating occasional interferences that are high in intensity but fluctuate instantaneously rather than continuously deteriorate, thereby significantly improving the personalization level of the warning and thus improving the accuracy of the warning trigger.

[0078] In summary, this embodiment determines whether the probability of the stress state corresponding to the current sliding window reaches a preset probability threshold. If it does, it is determined that the current sliding window has detected a stress state. Subsequently, this embodiment acquires the stress state detection results of each historical sliding window preceding the current sliding window, in a number equal to a preset number of windows. If all acquired historical sliding windows have detected a stress state, it is determined that an alert is triggered; otherwise, no alert is triggered.

[0079] This embodiment first determines whether the probability of the stress state of the current sliding window reaches a threshold. Then, it requires all historical sliding windows with a preset number of windows before the current window to detect the stress state before triggering an alarm. This ensures that the stress state exists continuously within multiple consecutive time windows rather than being a short-lived occurrence. This effectively filters out single false alarms caused by motion artifacts, instantaneous physiological fluctuations, or environmental interference, so that the system only issues an alarm when the stress state has temporal continuity, reducing the interference of invalid alarms on users.

[0080] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the physical condition monitoring method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0081] This application also proposes a wristband-type emergency monitoring ring, the wristband-type emergency monitoring ring comprising: The red light emitting module is used to emit red light of a preset first wavelength towards the human wrist. Sensors are used to collect reflected light signals to obtain waveform physiological signals; The main control module is used to determine the monitoring indicators of the human body based on the waveform physiological signal, input the monitoring indicators into the preset early warning model to obtain the stress state probability of the human body, and determine whether to trigger an early warning based on the stress state probability and the preset probability threshold. If an early warning is triggered, an alarm is issued and the corresponding physical status data is sent to the preset medical platform.

[0082] In one possible embodiment of this application, the wristband-type emergency monitoring ring further includes: An infrared light emitting module is used to emit a preset second wavelength of red light toward the wrist of the human body; The sensor is also used to collect reflected light signals to obtain infrared light physiological signals; The main control module is also used to extract the infrared pulsation component and the infrared DC component based on the infrared physiological signal, calculate the ratio of the pulsation component to the DC component to obtain a first ratio, calculate the ratio of the infrared pulsation component to the infrared DC component to obtain a second ratio, and calculate the blood oxygen saturation of the human body based on the first ratio, the second ratio and a preset empirical coefficient.

[0083] All user-related data involved in this application was obtained with the user's permission or consent, as per [reference]. Figure 5 In other words, when this application is applied to a specific product or technology, user permission is required to acquire and process the relevant data, and the processing of the relevant data must comply with the relevant laws, regulations and regulatory standards of the relevant countries and regions.

[0084] The above description is only a part of the embodiments of this application and does not limit the scope of protection of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for monitoring physical condition, characterized in that, The method, applied to a wristband-type emergency monitoring ring, includes: A preset first wavelength of red light is emitted towards the wrist of the human body, and the reflected light signal is collected to obtain a waveform physiological signal; Based on the waveform physiological signal, the monitoring indicators of the human body are determined, and the monitoring indicators are input into the preset early warning model to obtain the probability of the human body's stress state. The early warning model is obtained by training the model to be trained based on training samples labeled with stress state judgment results. Based on the stress state probability and the preset probability threshold, determine whether to trigger an early warning; If an alert is triggered, an alarm will be issued and the corresponding physical condition data will be sent to a pre-set medical platform.

2. The method as described in claim 1, characterized in that, The step of determining human monitoring indicators based on the waveform physiological signal includes: Based on a preset noise cancellation algorithm, noise is removed from the waveform physiological signal to obtain a denoised physiological signal; Based on the denoised physiological signal, the pulsation component, DC component, and waveform feature points are determined. The monitoring index is determined based on the pulsation component, the DC component, and the waveform feature points.

3. The method as described in claim 2, characterized in that, The step of removing noise from the waveform physiological signal based on the preset noise cancellation algorithm to obtain the denoised physiological signal includes: Acquire human motion data and use the motion data as reference noise; Based on a preset filter, the effective signal of the target frequency is selected from the waveform physiological signal, and the signals other than the effective signal are removed to obtain the heart rate signal; Based on the reference noise, the noise corresponding to the reference noise is removed from the heart rate signal to obtain the denoised physiological signal.

4. The method as described in claim 2, characterized in that, The waveform feature points include systolic peak values, and the monitoring indicators include real-time heart rate, heart rate variability, and perfusion index. The step of determining the monitoring indicators based on the pulsation component, the DC component, and the waveform feature points includes: Obtain the peak time of each of the contraction peaks, calculate the difference between the peak times, and obtain the peak time difference; The real-time heart rate is determined based on the peak time difference; Calculate the standard deviation of each of the peak time differences to obtain the peak time standard deviation, and determine the heart rate variability based on the peak time standard deviation; The perfusion index is obtained by calculating the ratio of the pulsatile component to the DC component.

5. The method as described in claim 1, characterized in that, The monitoring indicator is a monitoring indicator within a sliding window of a preset time length, and the stress state probability is the stress state probability corresponding to the sliding window. The step of determining whether to trigger an alert based on the stress state probability and a preset probability threshold includes: Determine whether the probability of the stress state corresponding to the current sliding window reaches the probability threshold; If this is achieved, it is determined that the current sliding window has detected a stress state; Obtain the detection results of the stress state corresponding to a preset number of historical sliding windows before the current sliding window; If the stress state is detected in all the historical sliding windows, an early warning is triggered.

6. The method as described in claim 1, characterized in that, Before the step of emitting a preset first wavelength of red light to the wrist of the human body, collecting the reflected light signal, and obtaining the waveform physiological signal, the method further includes: Obtain the training samples, test samples, and multiple preset candidate probability thresholds; The training samples are input into the model to be trained to obtain the prediction results of the model to be trained; Based on the prediction results and the stress state judgment results, the model to be trained is trained to obtain the early warning model; The probability threshold is determined based on the test sample, the candidate probability threshold, and the early warning model.

7. The method as described in claim 6, characterized in that, The test sample includes the stress state judgment result corresponding to the test sample, and the step of determining the probability threshold based on the test sample, the candidate probability threshold, and the early warning model includes: The test sample is input into the early warning model to obtain the probability of the stress state; Based on the stress state probability and the candidate probability threshold, the predicted stress judgment result is determined; Based on the predicted stress judgment result and the stress state judgment result, the false positive rate corresponding to each candidate probability threshold is determined; The average false positive rate of each candidate probability threshold tested over multiple rounds; The probability threshold is the candidate probability threshold that has the lowest average false positive rate.

8. The method as described in claim 2, characterized in that, After the step of emitting a preset first wavelength of red light to the wrist of the human body, collecting the reflected light signal, and obtaining a waveform physiological signal, the method further includes: A preset second wavelength of infrared light is emitted towards the wrist of the human body, and the reflected light signal is collected to obtain the infrared light physiological signal. Based on the infrared physiological signals, the infrared pulsation component and the infrared DC component are extracted; Calculate the ratio of the pulsating component to the DC component to obtain a first ratio, and calculate the ratio of the infrared light pulsating component to the infrared light DC component to obtain a second ratio; The blood oxygen saturation of the human body is calculated based on the first ratio, the second ratio, and a preset empirical coefficient.

9. A wristband-type emergency monitoring ring, characterized in that, The wristband-type emergency monitoring ring includes: The red light emitting module is used to emit red light of a preset first wavelength towards the human wrist. Sensors are used to collect reflected light signals to obtain waveform physiological signals; The main control module is used to determine the monitoring indicators of the human body based on the waveform physiological signal, input the monitoring indicators into the preset early warning model to obtain the stress state probability of the human body, and determine whether to trigger an early warning based on the stress state probability and the preset probability threshold. If an early warning is triggered, an alarm is issued and the corresponding physical status data is sent to the preset medical platform.

10. The wristband-type emergency monitoring ring as described in claim 9, characterized in that, The wristband-type emergency monitoring ring also includes: An infrared light emitting module is used to emit a preset second wavelength of red light toward the wrist of the human body; The sensor is also used to collect reflected light signals to obtain infrared light physiological signals; The main control module is also used to extract the infrared pulsation component and the infrared DC component based on the infrared physiological signal, calculate the ratio of the pulsation component to the DC component to obtain a first ratio, calculate the ratio of the infrared pulsation component to the infrared DC component to obtain a second ratio, and calculate the blood oxygen saturation of the human body based on the first ratio, the second ratio and a preset empirical coefficient.

Images