Blood oxygen saturation monitoring and early warning system and method based on oximeter

Abstract

The invention relates to the technical field of blood oxygen monitoring and early warning, in particular to a blood oxygen saturation monitoring and early warning system and method based on an oximeter. The method comprises the following steps: determining physiological information in two states by analyzing historical bracelet data of a user in a motion state and a resting state, screening out a body part which is most suitable for detection of an oximeter based on a measurement position difference for data acquisition, then calculating a recovery speed of blood oxygen fluctuation of the user, simulating a blood oxygen fluctuation range according to the recovery speed, and calculating the recovery speed of the blood oxygen fluctuation of the user. Therefore, a reasonable interval of blood oxygen change is determined, finally, the interval serves as an early warning reference, an abnormal alarm is triggered when the real-time blood oxygen data exceed a set range, and accurate and dynamic blood oxygen monitoring and health management are achieved. According to the method, the blood oxygen measurement position is optimized by analyzing motion and resting data, a dynamic early warning interval is established in combination with a fluctuation rule, and accurate blood oxygen abnormity monitoring and early warning are achieved.

Description

Technical Field

[0001] This invention relates to the field of blood oxygen monitoring and early warning technology, and in particular to a blood oxygen saturation monitoring and early warning system and method based on a pulse oximeter. Background Technology

[0002] Traditional blood oxygen monitoring systems often rely solely on static thresholds from a single device for alerts. This crude approach fails to adapt to individual physiological differences and dynamic environmental changes, resulting in a high false alarm rate. Especially during complex physiological states such as exercise recovery, traditional fixed threshold models cannot capture the true trends in blood oxygen fluctuations, leading to a one-sided health assessment and delayed alerts, missing the optimal intervention window. Meanwhile, when users transition from an active state to a resting state, the existing methods for measuring blood oxygen recovery speed have significant defects. Due to the lack of quantitative analysis of continuous changes in historical states, the system cannot establish a personalized blood oxygen recovery baseline, resulting in warning intervals that are either too broad or too narrow. This can lead to either missing real anomalies or generating excessive alarms. Summary of the Invention

[0003] Therefore, it is necessary to provide a pulse oximeter-based pulse oximeter-based pulse oximeter monitoring and early warning system and method to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a pulse oximeter-based method for monitoring and early warning of blood oxygen saturation includes the following steps: Step S1: Determine the user's first state information when in motion and the user's second state information when at rest based on the acquired historical wristband monitoring data; Step S2: Determine the historical wristband measurement location based on historical wristband monitoring data, and determine the preferred measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations; collect the first pulse oximeter signal at the preferred measurement location of the pulse oximeter; Step S3: Calculate the user's blood oxygen fluctuation recovery speed based on the continuous first and second state information in the historical wristband monitoring data; Step S4: Simulate multiple blood oxygen fluctuation recovery values ​​of the first blood oxygen signal according to the blood oxygen fluctuation recovery rate; use the maximum value among the blood oxygen fluctuation recovery values ​​as the upper limit of the interval and the minimum value among the blood oxygen fluctuation recovery values ​​as the lower limit of the interval, thereby limiting the range of the user's blood oxygen change trend. Step S5: Use the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the upper and lower boundaries of the blood oxygen warning range; when the change value of blood oxygen saturation exceeds the upper and lower boundaries of the blood oxygen warning range, trigger a blood oxygen abnormality warning.

[0005] This specification also provides a pulse oximeter-based blood oxygen saturation monitoring and early warning system for executing the pulse oximeter-based blood oxygen saturation monitoring and early warning method described above. The pulse oximeter-based blood oxygen saturation monitoring and early warning system includes: The state extraction module is used to determine the user's first state information when in motion and the user's second state information when at rest, based on the acquired historical wristband monitoring data. The location inference module is used to determine the historical measurement location of the wristband based on historical wristband monitoring data, and to determine the preferred measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations; and to collect the first pulse oximeter signal at the preferred measurement location of the pulse oximeter. The recovery rate calculation module is used to calculate the user's blood oxygen fluctuation recovery rate based on the continuous first and second state information in the historical wristband monitoring data. The trend simulation module is used to simulate multiple blood oxygen fluctuation recovery values ​​of the first blood oxygen signal according to the blood oxygen fluctuation recovery rate; the maximum value among the blood oxygen fluctuation recovery values ​​is used as the upper limit of the interval, and the minimum value among the blood oxygen fluctuation recovery values ​​is used as the lower limit of the interval, thereby limiting the user's blood oxygen change trend range; The abnormal warning module uses the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the upper and lower boundaries of the blood oxygen warning range; when the change value of blood oxygen saturation exceeds the upper and lower boundaries of the blood oxygen warning range, an abnormal blood oxygen warning is triggered.

[0006] The beneficial effects of this invention are as follows: On the one hand, by classifying and extracting the first and second state information of users in the exercise and rest states from the historical wristband monitoring data, the physiological response characteristics of individuals under different physiological activity levels can be accurately reflected, providing a data basis for subsequent blood oxygen fluctuation modeling and recovery capacity assessment, and effectively avoiding the problem of insufficient adaptability caused by single-state modeling.

[0007] On the other hand, by combining historical wristband measurement positions and preset wristband-pulse oximeter difference conditions to determine the optimal measurement position of the pulse oximeter, the first pulse oximeter signal is collected based on this, which significantly improves the quality and stability of pulse oximeter sampling, while reducing the impact of motion artifacts and external interference on measurement accuracy, ensuring the high fidelity and representativeness of pulse oximeter data.

[0008] On the other hand, by calculating the individual's blood oxygen fluctuation recovery rate through continuous state data and simulating the generation of multiple blood oxygen recovery values, the user's blood oxygen change trend range can be constructed, which can dynamically characterize the individual's blood oxygen regulation ability, achieve highly sensitive and specific capture of physiological state changes, and overcome the problem of poor adaptability of traditional static threshold models to individual differences.

[0009] On the other hand, by using the blood oxygen change trend range as the boundary of the warning range, the blood oxygen abnormality warning mechanism is dynamically triggered, which can promptly alarm when blood oxygen fluctuations exceed the individual threshold range, significantly improving the personalization, real-time nature and accuracy of blood oxygen monitoring. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the steps of a pulse oximeter-based method for monitoring and early warning of blood oxygen saturation. Figure 2 This is a schematic diagram of the blood oxygen fluctuation rate results; 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

[0011] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0012] 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.

[0013] 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.

[0014] To achieve the above objectives, please refer to Figures 1 to 2 A method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter includes the following steps: Step S1: Determine the user's first state information when in motion and the user's second state information when at rest based on the acquired historical wristband monitoring data; In this embodiment of the invention, based on historical monitoring data from a wearable wristband, time series data including triaxial acceleration, heart rate, and photoplethysmogram (PPG) signals are processed to extract the user's state information during both active and resting states. An acceleration magnitude threshold is set as follows. When the continuous acceleration modulus exceeds the threshold and the duration is not less than 8 seconds, the time period is identified as the exercise state segment. The corresponding mean heart rate, mean PPG signal amplitude, and rhythm are extracted from this segment as the first state information. At the same time, the time period with an acceleration modulus below 0.5 m / s² and a duration of not less than 90 seconds is set as the resting state segment. The resting mean heart rate, PPG baseline fluctuation amplitude, and baseline respiratory rate are extracted from this segment as the second state information.

[0015] In one implementation of this invention, assuming that the acceleration modulus of a certain data segment is consistently 2.6 m / s² for 12 seconds, this segment is defined as the motion state segment, corresponding to an average heart rate of 96 bpm, a PPG amplitude of 2.1 mV, and a rhythm cycle of 0.7 seconds. This data segment constitutes a set of first state information. In another segment, the acceleration is consistently below 0.3 m / s² for 120 seconds, the average heart rate is 68 bpm, the PPG amplitude is 1.1 mV, and the respiratory rate is 0.22 Hz. This segment constitutes a set of second state information.

[0016] Step S2: Determine the historical wristband measurement location based on historical wristband monitoring data, and determine the preferred measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations; collect the first pulse oximeter signal at the preferred measurement location of the pulse oximeter; In this embodiment of the invention, while extracting historical wristband monitoring data, the historical wearing position and its corresponding PPG signal quality parameters are identified. The wearing position is limited to three regions: the middle of the back of the wrist, the radial carpal bone, and the ulnar side of the forearm. Preset difference conditions are constructed using the signal-to-noise ratio (SNR) index and tissue light absorption rate. The difference conditions require that the difference in tissue transmittance between the 660nm and 940nm dual-bands should not exceed 8%, and the PPG signal SNR at the current wearing position should not be lower than 20 dB. Among multiple wearing positions, the preferred position that meets the above conditions is selected for the placement of the pulse oximeter.

[0017] In one implementation of this invention, assuming that the PPG signal SNR is 24.3dB at the mid-back of the wrist, the tissue transmittance is 0.68 in the infrared band and 0.63 in the red band, with a difference rate of 7.3%, and the above conditions are met, the mid-back of the wrist is selected as the pulse oximeter measurement point. A reflective pulse oximeter sensor is attached at this location, the light source excitation frequency is set to 500Hz, the sampling period is 15 seconds, and the light intensity data under the 660 nm and 940 nm channels are collected, and the first pulse oxygenation signal is calculated to be 97.2%.

[0018] Step S3: Calculate the user's blood oxygen fluctuation recovery speed based on the continuous first and second state information in the historical wristband monitoring data; In this embodiment of the invention, based on the continuously recorded first and second state information in historical state data, the state pair transitioning from the user's movement to a resting state is located. A state pair with a continuous duration of not less than 150 seconds is set as the effective analysis interval. The lowest point of blood oxygen saturation and the stable value at rest are extracted from the file during this transition process, and their recovery time is measured. The recovery speed parameter is then calculated, defined as the increase in blood oxygen saturation divided by the recovery time. The unit is % / second.

[0019] In one implementation of this invention, assuming that during a certain exercise-resting transition event, the blood oxygen saturation level is as low as 92% after exercise, stabilizes at 97% during the resting phase, and the recovery time is 18 seconds, then... 5%, The recovery time is calculated to be 18 seconds. The recovery rate is 0.2778% / second. This process is repeated for no less than 15 sets of recovery events, and the average of the mode and median is taken as the representative recovery rate value.

[0020] Step S4: Simulate multiple blood oxygen fluctuation recovery values ​​of the first blood oxygen signal according to the blood oxygen fluctuation recovery rate; use the maximum value among the blood oxygen fluctuation recovery values ​​as the upper limit of the interval and the minimum value among the blood oxygen fluctuation recovery values ​​as the lower limit of the interval, thereby limiting the range of the user's blood oxygen change trend. In this embodiment of the invention, a simulated recovery process is constructed using a representative recovery rate value combined with the initial value of the first blood oxygen signal, and the time axis of the recovery process is set from 0 to... A recovery point is constructed every second. The initial value is increased linearly according to the recovery rate or the recovery waveform is simulated using cubic spline interpolation. The blood oxygen fluctuation recovery value per second is calculated. The maximum and minimum blood oxygen values ​​during the entire recovery process are extracted and set as the upper and lower limits of the blood oxygen change trend range, respectively. This range is used to limit the boundary of an individual's blood oxygen fluctuation regulation ability.

[0021] In one implementation of this invention, assuming the initial value of the first blood oxygen signal is 92.5%, the recovery rate is 0.25% / second, and the recovery time is set to 20 seconds, the generated blood oxygen recovery sequence is 92.5%, 92.75%, 93.0%..., and finally recovers to 97.5%. Taking the maximum value of 97.5% and the minimum value of 92.5%, the blood oxygen change trend range of the user is [92.5%, 97.5%].

[0022] Step S5: Use the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the upper and lower boundaries of the blood oxygen warning range; when the change value of blood oxygen saturation exceeds the upper and lower boundaries of the blood oxygen warning range, trigger a blood oxygen abnormality warning.

[0023] In this embodiment of the invention, the established blood oxygen change trend range is used as the dynamic boundary for individual monitoring. The real-time collected blood oxygen saturation value is analyzed using a sliding window mechanism. The window length is set to 6 seconds and the interval is 2 seconds. The average value of the current blood oxygen saturation value is calculated in each window. If the value is lower than the lower limit of the trend range or higher than the upper limit of the range, and the abnormal conditions are met for two consecutive windows, a blood oxygen abnormality warning is triggered.

[0024] In one implementation of this invention, it is assumed that the current window average... If the blood oxygen level is 91.8% and the trend range is [92.5%, 97.5%], and this value is below the lower limit, and the blood oxygen value is still 91.6% in the next window, then the abnormal event will be recorded immediately, an alarm log will be generated, and a prompt will be triggered on the device. At the same time, the raw blood oxygen data, heart rate value, and acceleration data of the current and the two windows before and after will be written into the warning data cache area.

[0025] Preferably, step S1 includes: Extract wristband movement data and heart rate change data from historical wristband monitoring data; Identify the horizontal movement trajectory and vertical undulation trajectory in the wristband's movement data; When the band is confirmed to be moving by the horizontal movement trajectory and the vertical undulation trajectory and the user's heart rate change data is greater than the preset exercise heart rate change threshold, the user is determined to be in an exercise state, and the user state data in the historical band monitoring data is extracted as the first state information. When the band is confirmed to be stationary through horizontal movement and vertical undulation trajectory and the user's heart rate change data is less than the preset exercise heart rate change threshold, the user is determined to be in a resting state, and the user status data in the historical band monitoring data is extracted as the second status information.

[0026] In one implementation of this invention, historical monitoring data recorded by a wearable wristband during continuous wear is acquired. This data includes triaxial acceleration data (X, Y, Z directions), photoplethysmography (PPG) signals, and heart rate time series. The data sampling frequency is set to 50Hz. A medium-range filter is applied to the raw acceleration data for noise reduction, with a window length of 5 sampling points. Then, the vector magnitude of the acceleration values ​​within each 10-second time window is calculated to obtain the total acceleration sequence. The acceleration trend within the corresponding time window is extracted, and the difference between adjacent time windows is extracted using the sliding window method for horizontal movement trajectory analysis and longitudinal fluctuation trajectory analysis. The horizontal movement trajectory is determined based on the acceleration change amplitude along the X and Y axes, with a threshold value set as follows. The longitudinal undulation trajectory is based on the Z-axis acceleration change, and the judgment threshold is set as follows. If the acceleration changes in both directions exceed the threshold, it is determined that there is wristband movement behavior.

[0027] In another implementation of this invention, the heart rate change magnitude is defined by combining synchronously acquired heart rate data sequences and evaluating the heart rate change amplitude within the time period corresponding to each movement event. The heart rate difference between the current time window and the previous time window is used as the preset threshold for heart rate change during exercise, which is 8 beats per minute (bpm). If the wristband detects movement and ΔHR>8bpm, it determines that the user is in an exercise state and extracts the state data within the corresponding time window, including the average acceleration modulus, average heart rate, PPG waveform amplitude, and rhythm cycle, to form the first state information.

[0028] In another implementation of this invention, it is assumed that the average X-axis acceleration over a certain time period is... The average Z-axis acceleration is Heart rate rose from 78 bpm to 91 bpm. If the value is 13 bpm, then the motion judgment condition is met. The state data extracted in this time window is: the mean acceleration is... The average heart rate was 89 bpm, the PPG waveform amplitude was 1.9 mV, and the rhythm cycle was 0.65 seconds. This set of data was recorded as a set of first-state information.

[0029] In another implementation of this invention, if, during the same analysis process, it is identified that the wristband has not exhibited any movement behavior, i.e., the magnitude of the acceleration change in the X and Y directions is less than... The change in acceleration in the Z direction is lower than Meanwhile, the range of heart rate changes If the heart rate is below the preset threshold of 8 bpm, it is determined to be in a resting state. The state data within the corresponding time window is extracted as the second state information. This state data includes the mean static acceleration, the stable heart rate, the standard deviation of the PPG waveform, and the estimated respiratory rate.

[0030] In another implementation of this invention, it is assumed that the X-axis acceleration of the wristband remains constant for 90 consecutive seconds. The following is the Z-axis acceleration: The heart rate changed from 73 bpm to 74 bpm. If the acceleration modulus is 1 bpm, satisfying the resting state criterion, then the mean acceleration modulus value for that time period is extracted as follows: The mean heart rate was 73.5 bpm, the standard deviation of the PPG waveform was 0.15 mV, and the estimated respiratory rate was 0.21 Hz. This set of data was recorded as a set of second-state information.

[0031] Preferably, determining the historical measurement location of the wristband based on historical wristband monitoring data includes: Extract the horizontal movement trajectory and vertical undulation trajectory of the wristband from historical wristband monitoring data; The horizontal movement trajectory marks the coordinates at both ends of the trajectory, and the horizontal swing amplitude of the bracelet is calculated based on the coordinates at both ends of the trajectory. The highest and lowest points of the vertical axis of the wristband are identified based on the vertical undulation trajectory, and the vertical drop of the wristband is determined by the coordinates of the highest and lowest points of the vertical axis. Collect three-dimensional parameters of the human body and construct a three-dimensional human body model; The movement of the bracelet is simulated by traversing multiple bracelet wearing positions in the three-dimensional human body model until the swing amplitude in the bracelet's movement trajectory is consistent with the horizontal swing amplitude and the movement drop in the bracelet's movement trajectory is consistent with the vertical movement drop, thus obtaining the historical bracelet measurement position.

[0032] In one implementation of this invention, triaxial acceleration data is extracted from historical wristband monitoring data. This data includes the wristband's acceleration time series in the X-axis (horizontal forward / backward), Y-axis (horizontal left / right), and Z-axis (vertical up / down) directions. The sampling frequency is set to 50Hz, and a moving average filter is used to smooth the acceleration in each direction. The window length is 10 sampling points. The horizontal movement trajectory of the wristband is identified. The trajectory is estimated by integrating the acceleration in the X-axis and Y-axis directions twice to obtain the displacement value. The continuous movement window time is set to 10 seconds, updated once per second. Within each window, the starting and ending coordinates of the XY plane motion path are selected and denoted as . and Calculate the horizontal swing amplitude accordingly. Setting only when A swinging motion greater than 4 cm is recorded as a valid swinging action.

[0033] In another implementation of this invention, the Z-axis acceleration time series is extracted and peak-valley detection is performed on it, with a detection threshold set as follows. The highest point within each motion cycle is obtained through the local extremum algorithm. and the lowest point Calculate the longitudinal motion drop during this period. Set the minimum effective exercise cycle to 6 seconds. If the value is greater than 3 cm, then it is considered that there is effective longitudinal undulation motion.

[0034] In another implementation of this invention, it is assumed that within a certain 10-second time period, the horizontal start and end coordinates of the wristband are... and Calculate the horizontal swing amplitude Approximately 4.9 cm, identified in the Z-axis acceleration waveform for , for The estimated drop along the Z-axis is approximately 3.3 cm, and this motion is recorded as an effective spatial oscillation event.

[0035] In another implementation of this invention, three-dimensional body parameters of the test user are collected, including forearm length, wrist circumference, vertical distance from elbow to wrist, palm thickness, and coordinates of reference points on the body surface. The measurement method uses a laser rangefinder and a soft ruler for auxiliary recording. All three-dimensional parameters are input into three-dimensional human body modeling software for human body modeling. The modeling adopts the triangular mesh method to construct a standard human body surface model with a mesh accuracy set to 5 mm. Several common wristband wearing positions are preset in the model, including the center of the back of the wrist, the radial carpal bone, the ulnar carpal bone, and the middle of the forearm. The wristband movement is simulated at each position point. During the simulation, the trajectory of the wristband swinging with the forearm is simulated according to biomechanical angles. The simulated swing path and drop data are calculated and output by the software's built-in dynamics module.

[0036] In another implementation of this invention, the horizontal sway amplitude of each simulated path is compared with the actual extracted historical data. Compare the values, and also compare the vertical drop values. The allowable error range is set to no more than 0.5 cm. Only when the simulated swing amplitude and drop at a certain wearing position simultaneously meet the above error tolerance, and the number of simulated motion cycles is consistent with the actual number of tests, is the position determined to be the wearing point of the historical bracelet.

[0037] In another implementation of this invention, assuming that the horizontal swing amplitude of the simulated path at the center of the back of the user's wrist is 4.8 cm and the drop is 3.4 cm, both within the error tolerance range, and the number of simulation cycles at this position is 12, consistent with the actual detection, then the historical wristband measurement position is finally identified as the center of the back of the wrist.

[0038] Preferably, determining the optimal measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations includes: The preset wristband-pulse oximeter measurement difference conditions include the wristband-pulse oximeter selectable target location difference and the wristband-pulse oximeter signal synchronization conditions; The preferred measurement location of the candidate pulse oximeter is determined based on the differences in selectable target locations between the wristband and the pulse oximeter. The optimal pairing position of the wristband and pulse oximeter signals is identified by the synchronization conditions of the wristband and pulse oximeter signals. Based on the preferred pairing position of the wristband-pulse oximeter signal, the preferred pulse oximeter measurement position is selected from the candidate preferred pulse oximeter measurement positions.

[0039] In one implementation of this invention, a preset difference condition for wristband-pulse oximeter measurement is provided. The difference condition includes two types of elements: one is the difference in selectable target locations of wristband-pulse oximeter, which mainly consists of differences in human anatomical regions, soft tissue thickness, and skin blood vessel distribution density. Each target location difference parameter is labeled and encoded in a three-dimensional human body model. The other is the signal synchronization condition of wristband-pulse oximeter, which includes the degree of matching between the two devices in terms of heart pulse frequency and pulse wave phase within the simultaneous detection period. The degree of matching is obtained by calculating the standard correlation coefficient and the maximum cross-correlation value. The signal synchronization threshold is set to a correlation coefficient of not less than 0.85 and a cross-correlation time delay of not more than 80 milliseconds.

[0040] In another implementation of this invention, based on the identified historical wristband measurement locations, a physiological region label is marked on a three-dimensional human body model to which the location belongs. According to a preset wristband-pulse oximeter target location difference list, multiple candidate pulse oximeter target locations adjacent to or anatomically continuous with the labeled region are screened from the model. For example, if the wristband measurement point is located in the center of the back of the wrist, several candidate measurement points are extracted from the ulnar carpal bones, radial carpal bones, the center of the palm, and the inner side of the proximal forearm, based on selectable target location difference parameters. The subcutaneous tissue thickness, epidermal reflectivity, and capillary coverage of each point must meet the allowable deviation range set in the preset parameter table, which is limited to a thickness deviation of no more than ±2.5 mm, a reflectivity deviation of no more than ±0.04, and a capillary distribution density difference of no more than ±12%.

[0041] In another implementation of this invention, a wristband and a pulse oximeter are worn at corresponding locations and 5 minutes of continuous heart rate pulse waveform data are collected simultaneously. The sampling frequency is 100 Hz. The signal preprocessing process includes bandpass filtering (0.5 Hz-10 Hz) and baseline drift removal. The leading edge and peak value of the main pulse are extracted by wavelet transform. The correlation coefficient and maximum cross-correlation value between the pulse waveforms of the wristband and the pulse oximeter are calculated.

[0042] In another implementation of this invention, assuming the candidate pulse oximeter is located on the inner side of the proximal forearm, the correlation coefficient between the detected pulse waveform and the wristband pulse waveform is 0.91, and the maximum cross-correlation delay is 60 milliseconds, both of which meet the signal synchronization threshold condition and are recorded as the preferred signal pairing position.

[0043] In another implementation of this invention, among multiple candidate pulse oximeter locations, the measurement point that simultaneously satisfies the target location difference parameter and the signal synchronization condition is selected as the preferred measurement location of the pulse oximeter.

[0044] In another implementation of this invention, assuming that only the medial proximal forearm and the ulnar carpal bones among the candidate points meet all the above-mentioned index requirements, the medial proximal forearm is selected as the preferred measurement location for the final pulse oximeter.

[0045] Preferably, the calculation of the user's blood oxygen fluctuation recovery speed based on continuous first and second state information from historical wristband monitoring data includes: Query the first and second state information that appear continuously and alternately in the historical wristband monitoring data, and segment them into stage state information; Record the duration of the state change from the first state information to the second state information in the stage state information; Calculate the difference in blood oxygen saturation between states based on the first state information and the second state information; The user's blood oxygen fluctuation recovery speed is determined based on the difference in blood oxygen saturation and the duration of the status change.

[0046] In one implementation of this invention, such as Figure 2 As shown, the system reads the identified first and second state information from the historical wristband monitoring data. The first state information is the wristband data segment recorded when the user is in motion, and the second state information is the wristband data segment recorded when the user is in resting state. By sequentially traversing the wristband data timeline, the system finds a periodic state sequence composed of continuous and alternating first and second states. The system then divides the data into multiple stage state information segments, with each segment recording a timestamp, heart rate, blood oxygen saturation, and exercise amplitude indicators.

[0047] In another implementation of this invention, the start and end times of the state change between the first state and the second state are extracted and denoted as... and Duration of state change The unit is seconds, and the timestamp precision is set to milliseconds to ensure that the state transition detection error does not exceed 0.5 seconds. At the same time, the blood oxygen saturation value at the end of the first state is recorded. Blood oxygen saturation value after stabilization following the start of the second state Blood oxygen saturation values ​​are expressed as percentages (%), and the sampling frequency is no less than 1 Hz.

[0048] In another implementation of this invention, based on the blood oxygen saturation difference... Calculate the recovery rate of blood oxygen fluctuations during each state change process, defining the recovery rate V as the ratio of the change in blood oxygen saturation value to the corresponding state change duration, i.e. The unit is It reflects the extent to which blood oxygen saturation recovers per unit time.

[0049] In another implementation of this invention, assuming that in a certain stage, the blood oxygen saturation value is 92% at the end of the first state information, and after switching to the second state, the blood oxygen saturation reaches 98% after a stabilization process, and the state change duration is 120 seconds, then the blood oxygen recovery rate is: This recovery rate value is recorded in the recovery rate sample set to establish the boundary of the user's individualized blood oxygen change trend range and participate in the construction of the early warning model. The whole process can be repeated to obtain data samples of multiple state recovery segments.

[0050] Preferably, the multiple blood oxygen fluctuation recovery values ​​simulating the first blood oxygen signal according to the blood oxygen fluctuation recovery rate include: Determine the initial blood oxygen saturation based on the first blood oxygen signal; Using the initial blood oxygen saturation as a baseline, blood oxygen recovery simulation was performed according to the blood oxygen fluctuation recovery rate, and multiple recovery-state blood oxygen curves were constructed. The recovery value of blood oxygen fluctuation is derived from the recovered blood oxygen curve and the initial blood oxygen saturation.

[0051] In one implementation of this invention, an initial blood oxygen saturation value is extracted based on a first blood oxygen signal acquired by the pulse oximeter at a preferred measurement location. The initial blood oxygen saturation refers to the blood oxygen value at the moment the exercise state is identified as ending. This blood oxygen value is obtained by a photoplethysmography (PPG) sensor built into the pulse oximeter, and the relative oxygen saturation value is obtained after processing the ratio of red light to infrared light reflection signals. The timestamp of this value is consistent with the end time of the first state, and the unit is a percentage. The measurement time accuracy is controlled at the second level, and the sampling frequency is set to above 1Hz.

[0052] In another implementation of this invention, the initial blood oxygen saturation is used. Using the value as a baseline, the obtained blood oxygen fluctuation recovery rate V (in percentage per second) is retrieved. The blood oxygen recovery process is simulated using discrete-time simulation to construct multiple recovery-state blood oxygen curves. Each blood oxygen curve corresponds to a recovery scenario, where scenario parameters include different resting durations (e.g., 60 seconds, 90 seconds, 120 seconds, 180 seconds) and different upper and lower fluctuation values ​​of the recovery rate (set to a fluctuation range of ±20%). The curves are generated at each time step. Internal iterative update of blood oxygen saturation value Until the assumed resting duration is reached, all blood oxygen values ​​are limited to a physically reasonable range, i.e., below 100% and not less than 80%.

[0053] In another implementation of this invention, the blood oxygen saturation value at the end of each curve is extracted as the blood oxygen fluctuation recovery value of that curve, forming a set of recovery values, denoted as... ,in denoted as blood oxygen recovery value, and n as the number of curves.

[0054] In another implementation of this invention, the initial blood oxygen saturation is set to 92%, the blood oxygen fluctuation recovery rate is 0.06% / s, and three recovery durations are simulated: 60 seconds, 120 seconds, and 180 seconds, respectively, yielding recovery values. Since the blood oxygen level cannot exceed 100%, the excess portion needs to be truncated to 100%, and the final set of recovery values ​​is 95.6%, 99.2%, and 100%.

[0055] Preferably, the upper and lower boundaries of the blood oxygenation warning range are defined by using the blood oxygenation trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the blood oxygenation warning range as follows: Map the upper boundary of the blood oxygen change trend range to the upper limit boundary of blood oxygen saturation. The lower boundary of the blood oxygen change trend range is mapped to a dynamically adjusted lower limit boundary of blood oxygen saturation. When the blood oxygen change trend exceeds the preset trend threshold, the lower limit boundary of blood oxygen saturation is adjusted to 1%. The upper and lower limits of blood oxygen saturation are used to define the upper and lower boundaries of the blood oxygen warning range. The preset threshold for the trend of change is when the change value exceeds 2% of the initial blood oxygen saturation.

[0056] In one implementation of this invention, a blood oxygenation trend interval is constructed based on the set of all recovery values, denoted as... ,in The minimum value among all the endpoint values ​​of the recovery state curves The maximum value represents the reasonable range of blood oxygen fluctuations during the recovery period from exercise to rest. The width of this range reflects the strength of an individual's blood oxygen recovery ability; those with better blood oxygen regulation ability have a narrower fluctuation range, while those with poorer ability have a wider range.

[0057] In another implementation of this invention, the upper boundary of the blood oxygenation trend range is... This is directly mapped to the upper limit of blood oxygen saturation, indicating that an individual's blood oxygen level should not exceed this value under normal physiological conditions; while the lower limit of the interval... The mapping is a dynamically adjusted lower limit boundary for blood oxygen saturation, indicating that if the actual blood oxygen value is lower than this lower limit, it indicates that the recovery ability is abnormal or there is a risk of hypoxia.

[0058] In another implementation of this invention, a preset trend threshold is set as "the change amplitude exceeds 2% of the initial blood oxygen saturation", that is, if the blood oxygen change trend range width When the blood oxygen saturation level is ≥2%, the system determines that the individual's blood oxygen regulation fluctuation is too large and automatically forces the lower limit of blood oxygen to be adjusted to 1%, i.e., the lower limit. =1%, this value is used to trigger the extreme anomaly warning mechanism.

[0059] In another implementation of this invention, the dynamically adjusted upper and lower limits of blood oxygen saturation are used as the upper and lower thresholds of the blood oxygen warning range. The real-time blood oxygen monitoring data is limited by this range. If the real-time blood oxygen signal exceeds either boundary, the corresponding warning logic is triggered, and secondary confirmation is performed by combining the status code and historical data feedback.

[0060] In another implementation of this invention, it is assumed that the set of recovery values ​​obtained through simulation is as follows: The blood oxygen change trend range is [95.6%, 100%], with a range width of 4.4%. Since it exceeds the 2% threshold, the system forcibly corrects the lower limit boundary to 1% while retaining the upper limit boundary at 100%. The final upper and lower boundaries of the blood oxygen warning are 1% and 100%, respectively. This range is used for boundary constraint judgment of subsequent real-time blood oxygen data. If the blood oxygen value is detected at 94.5% at a certain time, it is within the range and the system judges it as normal; if the detected value is 0.8%, it is below 1% and triggers the first-level blood oxygen hypoxia warning.

[0061] Preferably, when the change in blood oxygen saturation exceeds the upper or lower boundary of the blood oxygen warning range, the blood oxygen abnormality warning is triggered, including: A warning trigger mechanism is set at the upper and lower boundaries of the blood oxygen warning range. When the change in blood oxygen saturation exceeds the upper boundary of the blood oxygen warning range, a blood oxygen abnormality warning is triggered. When the change in blood oxygen saturation falls below the lower boundary of the blood oxygen warning range, a blood oxygen abnormality warning is triggered. When the change in blood oxygen saturation reaches the warning range, a blood oxygen warning is triggered.

[0062] In one implementation of this invention, the upper and lower boundaries of the blood oxygen warning interval are obtained based on previous simulations and recovery speed analysis. For example, the upper boundary of blood oxygen is set to 99.2% based on the upper limit of the blood oxygen change trend, and the lower boundary is set to 94.5% based on the recovery model and dynamic adjustment logic. These upper and lower boundaries constitute the blood oxygen warning interval. Subsequently, warning triggering mechanisms are configured at both ends of the boundaries: when the real-time collected blood oxygen saturation change value is higher than the upper boundary of blood oxygen, it indicates that the user's blood oxygen rises too quickly or enters an abnormal oversaturation state, and the system immediately triggers a high blood oxygen abnormality warning; conversely, when the change value is lower than the lower boundary of blood oxygen, the system immediately triggers a low blood oxygen abnormality warning.

[0063] In another implementation of this invention, the system also integrates a prompting and warning mechanism to determine whether the blood oxygen change value has reached the early warning range even though it has not exceeded the limit. For example, the prompt range for a decrease is set to 94.5% to 95.5%, and the prompt range for an increase is set to 98.5% to 99.2%. If the current blood oxygen change value is within this range, the system triggers a blood oxygen prompting and warning, guiding the user to pay attention to ventilation or rest through a mild reminder, thus forming a graded warning system from prompt to abnormal.

[0064] In another implementation of this invention, assuming a user's real-time blood oxygen saturation value is 93.2% during a monitoring session, which is lower than the preset lower boundary of 94.5%, the system determines that a low blood oxygen abnormality warning has been triggered, and immediately vibrates the wristband and displays the message "Warning: Blood oxygen has dropped significantly, please stop your activity and sit down to rest" on the interface; or, if a user's blood oxygen saturation value is 98.7%, which falls within the upper limit of the warning range but has not yet exceeded the upper boundary of 99.2%, the system triggers a blood oxygen warning and displays "Blood oxygen is high, please pay attention to your respiratory rate" on the smart terminal.

[0065] Preferably, when the change in blood oxygen saturation reaches the warning range, triggering a blood oxygen warning includes: Collect a second blood oxygen signal, and determine the stage change value of blood oxygen saturation based on the difference in blood oxygen saturation between the first and second blood oxygen signals; Based on the lower boundary of the blood oxygen warning interval, the range of decline warning is divided inside the lower boundary according to the stage change value of blood oxygen saturation; Based on the upper boundary of the blood oxygen warning interval, the rising indication range is divided inside the upper boundary according to the stage change value of blood oxygen saturation. Integrate the descent and rise alert ranges to obtain the alert and warning range; When the change in blood oxygen saturation reaches the warning range, a blood oxygen warning is triggered.

[0066] In one implementation of this invention, in order to provide a gentle warning before the user's blood oxygen level reaches a dangerous threshold, the system sets a "warning range". This range is located inside the blood oxygen warning interval set by the system and is a range where there are slight changes but may indicate a change in the body's state.

[0067] In another implementation of this invention, after acquiring the first blood oxygen signal, a second blood oxygen signal is acquired after a period of time, and the difference in blood oxygen saturation between the two time points is calculated. This difference is called the stage change value of blood oxygen saturation.

[0068] In another implementation of this invention, a smaller range is further divided inward based on the upper and lower boundaries of the previously set blood oxygen warning range. For example, if the original lower boundary is 94.5%, the system may set a "decreasing warning range" between 94.5% and 95.5%. Similarly, if the upper boundary is 99.2%, an "increasing warning range" will be set between 98.5% and 99.2%.

[0069] In another implementation of this invention, the two warning ranges of "between 94.5% and 95.5%" and "between 98.5% and 99.2%" are integrated to form a complete warning range. Regardless of whether the blood oxygen value shifts upward or downward, as long as it enters this range, it is considered a potential minor abnormality.

[0070] In another implementation of this invention, assuming the user's first blood oxygen signal is 97% and the second blood oxygen signal is 94.6%, with a phase change value of 2.4% between the two signals, the system calculates and finds that the blood oxygen value has entered the preset decline warning range (e.g., between 94.5% and 95.5%). Although it has not yet fallen below the true warning lower limit of 94.5%, the change trend is obvious, so a blood oxygen warning is immediately triggered. The warning message may be "Blood oxygen value has slightly decreased, please rest or supplement oxygen". This kind of warning is usually reminded through APP notification or wristband icon to avoid disturbing the user.

[0071] In another implementation of this invention, if the user's blood oxygen level rises rapidly into the rising warning range, a gentle prompt will also be triggered, such as "Blood oxygen level has risen slightly, please slow down".

[0072] Therefore, the embodiments should be considered as 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.

[0073] 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 monitoring and early warning of blood oxygen saturation based on a pulse oximeter, characterized in that, Includes the following steps: Step S1: Determine the user's first state information when in motion and the user's second state information when at rest based on the acquired historical wristband monitoring data; Step S2: Determine the historical wristband measurement location based on historical wristband monitoring data, and determine the preferred measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations; collect the first pulse oximeter signal at the preferred measurement location of the pulse oximeter; Step S3: Calculate the user's blood oxygen fluctuation recovery speed based on the continuous first and second state information in the historical wristband monitoring data; Step S4: Simulate multiple blood oxygen fluctuation recovery values ​​of the first blood oxygen signal according to the blood oxygen fluctuation recovery rate; The maximum value among the blood oxygen fluctuation recovery values ​​is used as the upper limit of the interval, and the minimum value among the blood oxygen fluctuation recovery values ​​is used as the lower limit of the interval, thereby limiting the range of the user's blood oxygen change trend. Step S5: Use the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the upper and lower boundaries of the blood oxygen warning range; when the change value of blood oxygen saturation exceeds the upper and lower boundaries of the blood oxygen warning range, trigger a blood oxygen abnormality warning.

2. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, Step S1 includes: Extract wristband movement data and heart rate change data from historical wristband monitoring data; Identify the horizontal movement trajectory and vertical undulation trajectory in the wristband's movement data; When the band is confirmed to be moving by the horizontal movement trajectory and the vertical undulation trajectory and the user's heart rate change data is greater than the preset exercise heart rate change threshold, the user is determined to be in an exercise state, and the user state data in the historical band monitoring data is extracted as the first state information. When the band is confirmed to be stationary through horizontal movement and vertical undulation trajectory and the user's heart rate change data is less than the preset exercise heart rate change threshold, the user is determined to be in a resting state, and the user status data in the historical band monitoring data is extracted as the second status information.

3. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, The historical measurement locations of the wristbands were determined based on historical wristband monitoring data, including: Extract the horizontal movement trajectory and vertical undulation trajectory of the wristband from historical wristband monitoring data; The horizontal movement trajectory marks the coordinates at both ends of the trajectory, and the horizontal swing amplitude of the bracelet is calculated based on the coordinates at both ends of the trajectory. The highest and lowest points of the vertical axis of the wristband are identified based on the vertical undulation trajectory, and the vertical drop of the wristband is determined by the coordinates of the highest and lowest points of the vertical axis. Collect three-dimensional parameters of the human body and construct a three-dimensional human body model; The movement of the bracelet is simulated by traversing multiple bracelet wearing positions in the three-dimensional human body model until the swing amplitude in the bracelet's movement trajectory is consistent with the horizontal swing amplitude and the movement drop in the bracelet's movement trajectory is consistent with the vertical movement drop, thus obtaining the historical bracelet measurement position.

4. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, Determining the optimal pulse oximeter measurement location using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations includes: The preset wristband-pulse oximeter measurement difference conditions include the wristband-pulse oximeter selectable target location difference and the wristband-pulse oximeter signal synchronization conditions; The preferred measurement location of the candidate pulse oximeter is determined based on the differences in selectable target locations between the wristband and the pulse oximeter. The optimal pairing position of the wristband and pulse oximeter signals is identified by the synchronization conditions of the wristband and pulse oximeter signals. Based on the preferred pairing position of the wristband-pulse oximeter signal, the preferred pulse oximeter measurement position is selected from the candidate preferred pulse oximeter measurement positions.

5. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, Based on continuous first and second state information from historical wristband monitoring data, the user's blood oxygen fluctuation recovery speed is calculated, including: Query the first and second state information that appear continuously and alternately in the historical wristband monitoring data, and segment them into stage state information; Record the duration of the state change from the first state information to the second state information in the stage state information; Calculate the difference in blood oxygen saturation between states based on the first state information and the second state information; The user's blood oxygen fluctuation recovery speed is determined based on the difference in blood oxygen saturation and the duration of the status change.

6. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, The simulated blood oxygen fluctuation recovery values ​​based on the blood oxygen fluctuation recovery rate of the first blood oxygen signal include: Determine the initial blood oxygen saturation based on the first blood oxygen signal; Using the initial blood oxygen saturation as a baseline, blood oxygen recovery simulation was performed according to the blood oxygen fluctuation recovery rate, and multiple recovery-state blood oxygen curves were constructed. The recovery value of blood oxygen fluctuation is derived from the recovered blood oxygen curve and the initial blood oxygen saturation.

7. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, Using the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, the upper and lower boundaries of the blood oxygen warning range are defined as follows: Map the upper boundary of the blood oxygen change trend range to the upper limit boundary of blood oxygen saturation. The lower boundary of the blood oxygen change trend range is mapped to a dynamically adjusted lower limit boundary of blood oxygen saturation. When the blood oxygen change trend exceeds the preset trend threshold, the lower limit boundary of blood oxygen saturation is adjusted to 1%. The upper and lower limits of blood oxygen saturation are used to define the upper and lower boundaries of the blood oxygen warning range. The preset threshold for the trend of change is when the change value exceeds 2% of the initial blood oxygen saturation.

8. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 1, characterized in that, When the change in blood oxygen saturation exceeds the upper or lower boundary of the blood oxygen warning range, a blood oxygen abnormality warning is triggered, including: A warning trigger mechanism is set at the upper and lower boundaries of the blood oxygen warning range. When the change in blood oxygen saturation exceeds the upper boundary of the blood oxygen warning range, a blood oxygen abnormality warning is triggered. When the change in blood oxygen saturation falls below the lower boundary of the blood oxygen warning range, a blood oxygen abnormality warning is triggered. When the change in blood oxygen saturation reaches the warning range, a blood oxygen warning is triggered.

9. The method for monitoring and early warning of blood oxygen saturation based on a pulse oximeter according to claim 8, characterized in that, When the change in blood oxygen saturation reaches the warning range, the following factors will trigger a blood oxygen warning: Collect a second blood oxygen signal, and determine the stage change value of blood oxygen saturation based on the difference in blood oxygen saturation between the first and second blood oxygen signals; Based on the lower boundary of the blood oxygen warning interval, the range of decline warning is divided inside the lower boundary according to the stage change value of blood oxygen saturation; Based on the upper boundary of the blood oxygen warning interval, the rising indication range is divided inside the upper boundary according to the stage change value of blood oxygen saturation. Integrate the descent and rise alert ranges to obtain the alert and warning range; When the change in blood oxygen saturation reaches the warning range, a blood oxygen warning is triggered.

10. A pulse oximeter-based blood oxygen saturation monitoring and early warning system, characterized in that, For executing the pulse oximeter-based ... The state extraction module is used to determine the user's first state information when in motion and the user's second state information when at rest, based on the acquired historical wristband monitoring data. The location inference module is used to determine the historical measurement location of the wristband based on historical wristband monitoring data, and to determine the preferred measurement location of the pulse oximeter using preset wristband-pulse oximeter measurement difference conditions and historical wristband measurement locations; and to collect the first pulse oximeter signal at the preferred measurement location of the pulse oximeter. The recovery rate calculation module is used to calculate the user's blood oxygen fluctuation recovery rate based on the continuous first and second state information in the historical wristband monitoring data. The trend simulation module is used to simulate multiple blood oxygen fluctuation recovery values ​​of the first blood oxygen signal according to the blood oxygen fluctuation recovery rate; the maximum value among the blood oxygen fluctuation recovery values ​​is used as the upper limit of the interval, and the minimum value among the blood oxygen fluctuation recovery values ​​is used as the lower limit of the interval, thereby limiting the user's blood oxygen change trend range; The abnormal warning module uses the blood oxygen change trend range as the constraint boundary for monitoring blood oxygen saturation, thereby limiting the upper and lower boundaries of the blood oxygen warning range; when the change value of blood oxygen saturation exceeds the upper and lower boundaries of the blood oxygen warning range, an abnormal blood oxygen warning is triggered.

Images