A non-invasive real-time detection method, system and wearable device for blood blockage recovery time

By using a non-invasive detection method based on PPG signals and combining it with the timestamp of blood pressure measurement events, the recovery time after blood flow interruption is automatically extracted. This solves the problems of invasiveness and timeliness of traditional methods, and achieves accurate and robust assessment of blood flow. It is suitable for smart blood pressure monitors and wearable devices.

CN122250946APending Publication Date: 2026-06-23SHANGHAI JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610236872.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-02-28
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing technologies lack non-invasive, rapid, and real-time methods for monitoring blood flow, especially failing to measure the critical indicator of recovery time after blood pressure interruption. Traditional methods are highly invasive, have poor timeliness, and are costly, failing to meet the needs of clinical and home monitoring.

Method used

A non-invasive detection method based on photoplethysmography (PPG) signals is adopted, combined with the blood pressure measurement event timestamp, and the recovery time after blood flow interruption is automatically extracted by fusing multiple feature information. By using bilateral PPG signals and measurement hand information recorded by the blood pressure monitor, noise interference is eliminated to achieve accurate assessment of blood flow.

Benefits of technology

It achieves non-invasive and automated monitoring of blood flow, with more accurate and stable calculation results. It has commercial application potential and can be used in smart blood pressure monitors and wearable devices, expanding the assessment of vascular health and autonomic nervous function.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122250946A_ABST
    Figure CN122250946A_ABST
Patent Text Reader

Abstract

The application discloses a non-invasive real-time evaluation method and system for blood fluidity and a wearable device, and belongs to the technical field of biomedical signal processing and intelligent health monitoring. The method comprises the following steps: acquiring an event timestamp of a blood blocking device and a physiological signal of a blood health measurement module; automatically intercepting a signal segment of a recovery stage based on the event timestamp; parallelly calculating a plurality of independent indexes characterizing the recovery process; performing validity checking on each index, and eliminating invalid or abnormal candidate values; and outputting a final recovery time after blocking by using a dynamic adaptive fusion strategy. The application further discloses a detection system and a wearable device for realizing the method. The application uses recovery kinetics after pressure relief to represent blood fluidity, realizes automatic and accurate detection of the recovery time through a multi-index cooperative judgment and a dynamic fusion mechanism, effectively suppresses noise interference and improves robustness, and can be used for non-invasive evaluation of vascular elasticity, blood viscosity and microcirculation function.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of biomedical signal processing and intelligent health monitoring technology, specifically relating to a non-invasive real-time detection method, system and wearable device for blood flow recovery time after blood blockage. It can be used to non-invasively assess human blood flow, peripheral vascular function and microcirculation status, and is suitable for various application scenarios such as home health monitoring, clinical auxiliary diagnosis and wearable device integration. Background Technology

[0002] In hemodynamic and microcirculation studies, blood fluidity is a core indicator of cardiovascular health. Strong blood fluidity generally indicates low blood viscosity and good microcirculation function, which further suggests cardiovascular health and reflects a good physiological state in the human body.

[0003] Currently, blood flowability assessment mainly relies on traditional blood testing methods, such as blood viscosity testing via ex vivo blood sampling. These methods have the following problems:

[0004] 1) Invasive: Ex vivo blood testing requires drawing venous blood samples, a procedure that may cause discomfort, pain, or even infection risks for patients. For children, the elderly, or patients with chronic diseases, frequent blood draws can increase psychological burden and physical trauma. Furthermore, the complex preprocessing steps required after blood collection, such as centrifugation and stratification, further increase the difficulty and cost of the procedure.

[0005] 2) Poor timeliness: Traditional blood tests typically require samples to be sent to a laboratory for centrifugation, viscometer measurement, and other procedures, with testing cycles lasting from several hours to several days. This lag makes it impossible to provide results in real time, hindering the dynamic monitoring of changes in blood flow, such as in scenarios like post-exercise recovery or evaluation of drug intervention effects. For critically ill patients requiring real-time monitoring (such as those in shock or acute heart failure), the response speed of traditional methods cannot meet clinical needs.

[0006] 3) High cost and limited accessibility: Blood viscosity testing relies on specialized laboratory equipment (such as rotational viscometers and capillary viscometers) and requires certified testing personnel to operate, resulting in high costs per test. This high barrier to entry limits its application in home testing, community screening, and resource-scarce areas.

[0007] Current technologies lack a simple, non-invasive, rapid, and real-time method for monitoring blood flow, particularly failing to measure the crucial indicator of recovery time after blood pressure interruption. This indicator directly reflects vascular elasticity, blood viscosity, and microcirculatory function, and is of significant value in cardiovascular disease risk assessment.

[0008] Cuff-type blood pressure monitors measure blood pressure by applying pressure to the cuff. Inflation of the cuff causes temporary closure of local blood vessels, interrupting blood flow and disrupting microcirculation. After deflation, a recovery process is required to restore normal blood flow. In contrast, non-invasive monitoring technologies (such as PPG) offer advantages such as miniaturization, ease of operation, and widespread availability, but current technologies have not fully realized their potential. In non-invasive cardiovascular monitoring, photoplethysmography (PPG) signals are widely used for measuring heart rate, blood oxygen saturation, and related physiological indicators due to their simplicity, wearability, and ease of acquisition. In cuff-type electronic blood pressure monitors, cuff inflation causes a temporary closure of local blood flow, followed by reperfusion after deflation. The recovery time after this closure reflects peripheral vascular compliance and hemodynamic characteristics, possessing potential clinical and home monitoring value. Current blood pressure measurement devices typically only output blood pressure results and do not provide quantitative indicators of the blood flow recovery process after occlusion. Furthermore, existing PPG monitoring methods are mainly used for steady-state physiological measurements, lacking specific analytical tools for dynamic signal changes during the occlusion-recovery process. In addition, single-channel PPG is susceptible to the effects of residual pressure in the measuring hand, motion noise, or local vascular physiological differences, making it difficult to automatically and stably identify the recovery time after occlusion using traditional methods. The research field still faces a core technological gap: currently, there is no technical solution based on wearable devices and their signals to automate and non-invasively detect the blood flow recovery time after blood pressure measurement, failing to meet the needs of accurate vascular function assessment in clinical and home settings.

[0009] Therefore, there is an urgent need for a technical solution based on PPG signals that can automatically extract effective fragments after blood pressure measurement and integrate multiple feature information to achieve robust detection of recovery time after blockage, so as to realize non-invasive and automated assessment of blood flow. Summary of the Invention

[0010] To address the technical problems of existing electronic blood pressure monitors, such as the inability to quantify the degree of blood flow recovery after cuff deflation, the susceptibility of single-channel PPG to pressure interference, and the inability to objectively determine the recovery time after occlusion, this invention proposes a non-invasive real-time detection method, system, and wearable device for blood flow recovery time after occlusion. Based on photoplethysmography (PPG) signals combined with blood pressure measurement event timestamps, it achieves accurate, robust, and automated monitoring of blood flow, a core health indicator. The blood flow recovery time after occlusion described in this invention refers to the kinetic characteristic parameter characterizing the recovery of peripheral blood perfusion from an occluded state to a steady state after the external blood flow occlusion is removed. This invention focuses not on static blood flow parameters, but on the dynamic recovery characteristics of the blood perfusion process after the removal of externally controllable disturbances.

[0011] The technical solution of the present invention is as follows:

[0012] In a first aspect, the present invention provides a non-invasive real-time blood flow assessment method based on the recovery time after blood stasis, characterized by comprising the following steps:

[0013] Acquire at least one event timestamp output by the blood clogging device, the event timestamp corresponding to the moment when the blood clogging device finishes applying pressure, and a time-stamped physiological signal continuously acquired by at least one blood health measurement module;

[0014] Based on the event timestamp, signal segments corresponding to the recovery phase after blood flow occlusion are automatically extracted from the physiological signals;

[0015] The signal segment is analyzed to calculate at least two independent indicators that characterize the recovery process from different physiological dimensions, and each independent indicator outputs a candidate recovery time after blockage.

[0016] The at least two independent indicators are selected from: indicators based on signal quality recovery characteristics, indicators based on heart rate stability recovery characteristics, and indicators based on bilateral pulse wave synchronicity recovery characteristics.

[0017] The validity of each candidate recovery time after blocking is verified. Based on at least one criterion of signal reliability, noise interference level and cross-index consistency, invalid or abnormal candidate values ​​are eliminated to obtain a set of valid candidate recovery times.

[0018] Based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them, a dynamic fusion strategy is adopted to output the final recovery time after blood blockage, and blood flow is evaluated based on the recovery time after blood blockage.

[0019] Furthermore, when the blood health measurement module is only worn on the blocking side, the index based on signal quality recovery characteristics is calculated in the following way:

[0020] The average signal quality of the first signal during the blood blockage phase and the average signal quality of the second signal during the blood stabilization phase in the signal segment are identified, and the difference between the average signal quality of the second signal and the average signal quality of the first signal is used as a dynamic threshold.

[0021] A sliding window is used to scan the signal quality curve backward from the time point corresponding to the event timestamp. The moment when the average signal quality within the sliding window first reaches or exceeds the sum of the first signal quality average and the dynamic threshold is recorded as the first candidate recovery time t1.

[0022] Furthermore, when the blood health measurement module is worn on both the blocked side and the healthy side, the index based on signal quality recovery characteristics is calculated using the following noise reduction method:

[0023] Calculate the signal quality difference between the blocked side and the healthy side;

[0024] The mean value of the signal quality difference between the channels during the blood blockage period was used as the benchmark;

[0025] A sliding window is used to scan backwards from the time point corresponding to the event timestamp. The moment when the signal quality difference between channels first drops below the set ratio of the benchmark within the sliding window is recorded as the first candidate recovery time t1.

[0026] Furthermore, when the blood health measurement module is worn only on the blocked side, the index based on heart rate stability recovery characteristics is calculated in the following way:

[0027] The first mean heart rate during the blood blockage period and the second mean heart rate during the blood stabilization period are identified in the signal segment, and the absolute value of the difference between the second mean heart rate and the first mean heart rate is used as a dynamic threshold.

[0028] A sliding window is used to scan the heart rate curve backward from the time point corresponding to the event timestamp. The moment when the average heart rate within the sliding window first enters the dynamic threshold interval centered on the second average heart rate and remains stable for at least three consecutive sliding windows is recorded as the second candidate recovery time t2.

[0029] Furthermore, when the blood health measurement module is worn on both the blocked side and the healthy side, the index based on heart rate stability recovery characteristics is calculated using the following noise reduction method:

[0030] Calculate the bilateral heart rate error between the blocked side and the healthy side to eliminate systemic heart rate fluctuation interference;

[0031] The mean of the bilateral heart rate errors during the blood stasis period was used as the benchmark;

[0032] A sliding window is used to scan backwards from the time point corresponding to the event timestamp. The moment when the bilateral heart rate error first drops below the set proportion of the benchmark within the sliding window is recorded as the second candidate recovery time t2.

[0033] Furthermore, when the blood health measurement module is worn on both the blocked side and the healthy side respectively, the index based on the bilateral pulse wave synchronicity recovery characteristics is calculated in the following way:

[0034] The positions of all peaks in the PPG signals on the blocked side and the healthy side were detected separately, generating two peak time series.

[0035] The nearest neighbor time matching algorithm is used to pair the peaks on both sides. Peaks with a time difference of less than a preset threshold are identified as matching peak pairs. The preset threshold is 50-100 milliseconds.

[0036] Starting from the time point corresponding to the event timestamp, scan backwards and record the moment when N consecutive matching peak pairs first appear as the third candidate recovery time t3, where N is 3-5.

[0037] Furthermore, the validity verification of the recovery time after blocking for each candidate includes a combination of at least two of the following verification methods:

[0038] Value range verification: Remove candidate values ​​that exceed the preset physiological reasonable range of 0.5 seconds to 30 seconds;

[0039] Waveform stability verification: Remove candidate values ​​whose average signal quality index of the signal segment on which the calculation is based is lower than 0.5;

[0040] Peak detection rate verification: Remove candidate values ​​with peak detection rates lower than a preset threshold;

[0041] Cross-indicator consistency pre-verification: Remove suspicious values ​​that differ from other valid candidate values ​​in the same batch by more than a preset threshold.

[0042] Furthermore, the step of using a dynamic adaptive fusion strategy to output the final recovery time after blood stasis, based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them, specifically includes:

[0043] Case A: If the set of valid candidate recovery times is empty, the detection is deemed to have failed, an invalid result is output, and the user is prompted.

[0044] Case B: If there is only one valid value in the set of valid candidate recovery times, then that valid value will be used as the final result.

[0045] Case C: If there are two or more valid values ​​in the set of valid candidate recovery times, then calculate the maximum difference between the valid values, D = max(T). valid )-min(T valid );

[0046] Case C1: If D ≤ preset difference threshold θ, then it is determined that the indicators are highly consistent, and the arithmetic mean of each effective value or the weighted average based on the dynamic weighting of signal quality is output.

[0047] Case C2: If D > preset difference threshold θ, it is determined that the index is disturbed and drifts. The two most similar valid values ​​are selected from all valid values, and the arithmetic mean of the two valid values ​​is output. The preset difference threshold θ is 2 seconds.

[0048] Furthermore, the step of using a dynamic fusion strategy to output the final recovery time after blood stasis, based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them, specifically includes:

[0049] If the set of valid candidate recovery times is empty, then output an invalid result;

[0050] If there is only one valid value in the set of valid candidate recovery times, then that valid value will be used as the final result.

[0051] If there are two or more valid values ​​in the set of valid candidate recovery times, the maximum difference between the valid values ​​is calculated.

[0052] If the maximum difference is less than or equal to a preset threshold, the arithmetic mean or weighted average of each valid value is output.

[0053] If the maximum difference is greater than a preset threshold, then the two most similar valid values ​​are selected from all valid values, and the arithmetic mean of the two valid values ​​is output.

[0054] Secondly, the present invention provides a non-invasive real-time detection system for recovery time after blood stasis, characterized in that it includes:

[0055] The blood flow blocking module is used to apply external pressure to the target limb to form a blood flow blocking state, and to release the external pressure at a predetermined time and output event information with a precise timestamp.

[0056] The blood health measurement module is used to continuously collect physiological signals characterizing the peripheral blood perfusion status during the process of blood flow occlusion and release. It includes at least one wearable sensor with physiological signal acquisition functions such as photoplethysmography pulse wave, which is used to continuously collect and output physiological signals with precise timestamps.

[0057] The data processing terminal is communicatively connected to the blood blocking module and the blood health measurement module. It has a built-in signal processing unit and the detection method described above, and is used to receive data, perform time alignment, signal analysis and result output.

[0058] Furthermore, the blood health measurement module and the blocking component of the blood blocking module are deployed on the same side of the limb as the heart, with the blood health measurement module located at the distal end of the limb and the blocking component located at the proximal end of the limb.

[0059] Furthermore, the blood health measurement module includes two PPG acquisition units, which are worn on the blocked limb and the healthy limb respectively, forming a bilateral control acquisition layout.

[0060] Furthermore, based on the measurement hand information output by the blood clogging module, the data processing terminal automatically identifies and marks the clogging side channel and the healthy side channel, and uses the healthy side channel signal as a dynamic reference baseline for recovery time analysis.

[0061] Furthermore, the blood health measurement module is selected from one or more of the following wearable devices: smart ring, smartwatch, smart bracelet, finger clip pulse oximeter, smart earphone, clothing or medical clip with PPG acquisition function; the physiological signal includes at least one signal that can characterize the peripheral blood perfusion state or has a definite correlation with its changes, the signal includes photoplethysmography (PPG) signal, blood oxygen signal, skin temperature signal or a combination of multiple signals; the data processing terminal is selected from one of the following devices: smartphone, tablet computer, personal computer, cloud server, edge computing node, embedded processor of the wearable sensor body.

[0062] Thirdly, the present invention provides a wearable device as a blood health measurement module in the above-mentioned system, characterized in that it includes:

[0063] Physiological signal sensor, used to collect physiological signals such as raw photoplethysmography pulse waves;

[0064] Analog-to-digital converter module, used to convert analog signals into digital signals;

[0065] The timestamp calibration module is used to ensure that each sampling point has an accurate timestamp that is consistent with the time reference of the blood blocking device;

[0066] The signal preprocessing module is used to filter, denoise, and perform baseline correction on the raw signal;

[0067] The data transmission module is used to transmit time-stamped physiological signals to the data processing terminal.

[0068] The technical solution of this invention mainly includes the following improvements:

[0069] (1) Propose a non-invasive, real-time measurement method for blood flow based on PPG: This invention utilizes PPG signals for the first time and places the blood blocking device and the blood measurement device on the same side of the heart in sequence. Candidate values ​​of blood recovery time after blocking are extracted by specified rules, and a reliable blood recovery time after blocking is calculated, and the blood flow analysis is completed.

[0070] (2) The invention proposes the use of bilateral PPG signals combined with hand information from blood pressure measurement to automatically determine the effective channel: This invention is the first to utilize bilateral PPG signals worn on two fingers, combined with hand information recorded by the blood pressure monitor, to automatically identify the sensor of the hand being pressurized. Channel allocation and validity verification are performed before data enters the analysis, avoiding analysis failures caused by incorrect channels. Acquiring bilateral data can minimize the impact of common noise on experimental results.

[0071] (3) Constructing an automatic PPG fragment extraction mechanism based on blood pressure measurement event timestamps: This invention utilizes the event timestamps such as "start measurement" and "complete measurement" provided by the blood pressure monitor to automatically extract PPG fragments after each blood pressure measurement, ensuring that the time period analyzed by the algorithm accurately corresponds to the blood flow recovery stage after the cuff is deflated, fundamentally solving the problem that traditional manual cutting cannot be batch processed.

[0072] (4) A multi-indicator recovery judgment mechanism is proposed: Quality threshold method: The average quality of the signal at non-measurement locations is not less than a specified threshold, and the first time that meets the quality constraint is found by using the moving average signal quality. Heart rate average method: The average heart rate before the end of the measurement is used as the baseline, and the heart rate is detected to be stable through a sliding window. Peak pairing method: The nearest neighbor pairing of the peak times on both sides is used to determine the time of synchronous recovery of the bilateral pulses, reflecting the stability of blood flow. This invention is the first to simultaneously integrate multiple recovery judgment indicators.

[0073] The present invention has the following significant technical effects:

[0074] (1) The recovery time calculation after blockade is more accurate and stable: The present invention uses two-sided PPG quality matching, heart rate stability and peak matching synchronization to jointly evaluate the recovery process. Compared with the traditional method that only relies on a single channel or a single indicator, its recovery time result after blockade is more robust and can effectively eliminate the influence of residual pressure and noise in the measurement hand.

[0075] (2) Significantly improved quantifiability, applicable to large-scale batch processing and clinical evaluation: This invention achieves automated recovery point identification, generating a clear recovery time in seconds after blockage without manual waveform observation. This makes the recovery time after blockage a standardized, engineerable, and objective indicator, facilitating clinical research, home health monitoring, and integration by device manufacturers. This invention relies on the timestamps automatically recorded by the blood pressure monitor and existing PPG sensor data to operate, possessing strong commercialization capabilities and can be directly deployed in smart blood pressure monitors and wearable devices.

[0076] (3) Strong anti-spurious recovery ability, avoiding misjudgment by traditional methods: The heart rate mean method combined with the quality judgment method can avoid transient false recovery caused by noise; the peak pairing method can ensure synchronous recovery of bilateral pulses and avoid misjudgment caused by premature recovery of single channel. It can serve as a potential new indicator of vascular health and autonomic nerve function, and has medical expansion value: The recovery time after blockage provided by this invention can be used to reflect the return capacity of peripheral blood vessels, vascular elasticity and neural regulation ability, and can expand the clinical additional functions of electronic blood pressure monitors. Attached Figure Description

[0077] Figure 1 This is a schematic diagram of the data acquisition process of the present invention.

[0078] Figure 2 Single signal acquisition procedure (2.5-minute rest + approximately 1-minute blood pressure measurement)

[0079] Figure 3 Components of a blood clotting device (such as a cuff blood pressure monitor)

[0080] Figure 4 Components of blood measurement devices (such as smart rings)

[0081] Figure 5 Photoplethysmography (PPG)

[0082] Figure 6 Calculation process for recovery time after blood flow interruption

[0083] Figure 7 Single-channel threshold calculation rules

[0084] Figure 8 Threshold calculation rules for dual channels

[0085] Figure 9 Calculation scheme for recovery time after blood flow interruption

[0086] Figure 10 This is a schematic diagram of the composition and structure of the recovery time detection system after blood stasis according to the present invention. Detailed Implementation

[0087] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should be noted that these embodiments are merely illustrative of specific implementations of the present invention and do not constitute a limitation on the scope of protection of the present invention. All equivalent substitutions or modifications made based on the core ideas of the present invention fall within the scope of protection of the present invention.

[0088] Please refer to Figure 10 This invention provides a blood stasis recovery time detection system, comprising three parts: a blood stasis module, a blood health measurement module, and a data processing terminal. It consists of the following components:

[0089] 1. A blood occlusion module is used to apply controllable pressure to a local area of ​​the limb to temporarily occlude blood vessels and output event information with precise timestamps. In this embodiment, the module uses a commercially available blood pressure monitor, which can be an upper arm or wrist electronic blood pressure monitor, and can transmit measurement event recording data to a data processing terminal via Bluetooth or a data cable. Its internal components are as follows: Figure 3 As shown, it mainly includes:

[0090] Cuff compression module: used to apply controllable pressure to a local area of ​​the limb to achieve temporary vascular occlusion;

[0091] Pressure sensor: monitors changes in cuff pressure in real time;

[0092] Event timestamp generation module: Generates high-precision timestamps (accuracy better than 10 milliseconds) at key event points (such as "start pressurization", "reach maximum pressure", "start venting", "measurement completed").

[0093] Data transmission module: Supports Bluetooth, Wi-Fi or wired data cable, used to transmit event timestamps and blood pressure measurement values ​​to the data processing terminal;

[0094] Power management module: Provides stable power to all modules.

[0095] 2. A blood health measurement module, used to continuously collect physiological signals reflecting blood flow dynamics. In a preferred embodiment of the invention, this module employs a wearable device with PPG signal acquisition capabilities, such as a smart ring, smartwatch, or fingertip pulse oximeter, and its internal components are as follows: Figure 4 As shown, it mainly includes:

[0096] PPG sensor: Includes at least one light source (green light, red light or infrared light) and a corresponding photodetector, used to sense changes in blood volume;

[0097] Analog-to-digital conversion module: converts analog photoplethysmography signals into digital signals, with a sampling frequency of not less than 100Hz, preferably above 250Hz;

[0098] Timestamp calibration module: Ensures that each sampling point has an accurate timestamp with the same time reference as the blood blocking module, and the time synchronization accuracy is better than 10 milliseconds (such as through Bluetooth synchronization or NTP protocol calibration).

[0099] Signal preprocessing module: performs filtering (such as 0.5Hz-10Hz bandpass filtering), noise reduction (such as motion artifact suppression), baseline correction, and other processing on the raw PPG signal;

[0100] Feature extraction module (optional): can extract signal features in real time or near real time, such as peak position, instantaneous heart rate, signal quality index (SQI), etc.

[0101] Data transmission module: Transmits the original signal or pre-extracted feature data with timestamps to the data processing terminal via Bluetooth or data cable.

[0102] 3. A data processing terminal, used to receive, synchronize, and analyze data from the above two modules, and incorporates the core detection algorithm of this invention. This terminal can be: a local consumer electronic device, such as a smartphone, tablet, or personal computer; a cloud server or edge computing node; or an embedded processor within a wearable device (i.e., the algorithm runs directly on the ring or watch). The functional modules of the data processing terminal include: a data receiving and parsing module, a time synchronization and signal segment extraction module, a multi-indicator calculation module, a dynamic fusion and decision-making module, and a result output and visualization module.

[0103] Reference Figure 1 and Figure 2 The data acquisition process of this invention is as follows:

[0104] S1. Device Wearing: Subjects wear the device according to the system layout requirements. In a preferred embodiment, subjects wear a smart ring on each finger of both sides, and an electronic blood pressure cuff on one upper arm, forming a "pressure side-healthy side" control layout.

[0105] S2. Resting period: Subjects remain at rest for approximately 2.5 minutes, during which time baseline physiological signals are continuously acquired by bilateral PPG devices.

[0106] S3. Blood Pressure Measurement and Blood Flow Interruption Phase: Activate the electronic blood pressure monitor and execute the standard blood pressure measurement procedure. The cuff is inflated to a predetermined pressure (usually higher than the systolic pressure) and maintained for a short period, causing blood flow interruption to the distal end of the inflated limb; subsequently, the cuff is gradually deflated. The blood pressure monitor generates precise timestamps at key event points such as "begin inflation," "reach maximum pressure," "begin deflation," and "measurement complete," and transmits these timestamps in real time to the data processing terminal via Bluetooth.

[0107] S4. Blood Flow Recovery Phase: After the cuff is fully deflated, blood flow on the pressurized side begins to recover, and the PPG signal gradually returns to normal from a significantly depressed or absent state. Bilateral PPG devices continue to acquire signals until the next measurement begins or the preset monitoring duration ends.

[0108] S5. Data Upload and Synchronization: All timestamped PPG signal data and blood pressure monitor event data are aggregated to the data processing terminal, and the terminal completes data alignment based on a unified time reference.

[0109] Reference Figure 6 The specific implementation steps of the method of the present invention are as follows:

[0110] Step 1: Automatic segmentation of signal based on event-driven mechanism

[0111] The data processing terminal receives the "measurement complete" event timestamp T from the blood pressure monitor. complete Then, automatically extract from the continuous PPG signal stream the signal from T completeThe signal segment S from the start of the first measurement to the start of the next measurement (or a preset duration, such as 30 seconds). analysis This segment precisely corresponds to the blood flow recovery phase.

[0112] Step 2: Calculation of multi-dimensional physiological indicators and generation of candidate recovery times

[0113] For the extracted signal segments, sanalysis was performed, and the following three independent indicators characterizing the recovery process from different physiological mechanisms were calculated in parallel to generate candidate recovery times t1, t2, and t3:

[0114] (1) Indicator t1 based on signal quality recovery characteristics

[0115] Single-channel configuration (e.g.) Figure 7 ): Identify the first mean signal quality Q during the blood occlusion phase of the signal segment (near the lowest signal quality point during cuff compression). occlusion The mean Q signal quality of the second signal during the stable blood phase (end of the fragment, when the signal has been fully recovered) stable The difference between the two is ΔQ = Q. stable - Q occlusion As a dynamic threshold, a sliding window (window width 2 seconds, step size 0.2 seconds) is used to select the threshold from T. complete Begin scanning the signal quality curve backwards; when the average signal quality within the sliding window first reaches or exceeds Q... occlusion + α • When ΔQ is reached (α is an adjustable coefficient, preferably 0.7-0.9), the starting time of this window is recorded as t1. The value of α can be adaptively adjusted according to individual physiological characteristics, such as dynamically setting parameters based on user age, gender, blood pressure level, etc. (see step 4 for details).

[0116] Dual-channel configuration (such as) Figure 8 Using the signal from the healthy side (not the measuring hand) as a reference, calculate the signal quality difference ΔQ between the blocked side and the healthy side. inter = Q healthy - Q measured This difference can effectively eliminate common noise interference such as ambient light and motion artifacts. The mean value of ΔQ_inter during blood flow cessation is Q. diff_occ Based on T. A sliding window is used from T... complete Begin scanning backwards, when ΔQ_inter first drops to β·Q within the sliding window. diff_occ When β is an adjustable coefficient, preferably 0.2-0.4, the starting time of this window is denoted as t1. The value of β can also be adaptively adjusted according to individual physiological characteristics.

[0117] (2) t2 index based on heart rate stability recovery characteristics

[0118] Single-channel configuration: Identifies the first mean heart rate (HR) during blood occlusion in the signal segment. occlusion Mean heart rate (HR) during blood stabilization period stable The difference between the two is ΔHR = |HR stable - HR occlusion | As a dynamic threshold. A sliding window is used from T complete Begin scanning the heart rate curve backwards; when the average heart rate within the sliding window first enters [HR]... stable -γ·ΔHR,HR stable When the value remains stable within the interval [+γ·ΔHR] for at least three consecutive sliding windows, the starting time of that window is recorded as t2. γ is an adjustable coefficient, preferably 0.3-0.5, which can be adaptively adjusted according to individual characteristics.

[0119] Dual-channel configuration: Using the heart rate of the healthy side as a reference, calculate the bilateral heart rate error ΔHR between the blocked side and the healthy side. inter = |HR measured - HR healthy |, to eliminate interference from systemic heart rate fluctuations. Using ΔHR inter Mean HR during blood stasis diff_occ Based on T. A sliding window is used from T... complete Begin scanning backwards, when ΔHR_inter first drops to δ·HR within the sliding window. diff_occ The following time (δ is an adjustable coefficient, preferably 0.2-0.4) is used to denote the starting time of the window as t2. δ can be adaptively adjusted according to individual characteristics.

[0120] (3) Index t3 based on bilateral pulse wave synchronicity recovery characteristics (dual-channel configuration only)

[0121] The nearest neighbor time matching algorithm is used to pair the peak positions of the two-sided PPG signals. The specific implementation is as follows:

[0122] The positions of all peaks in the PPG signals on the blocked and healthy sides were detected separately, generating two peak time series P. measured and Phealthy;

[0123] For each peak time p of the healthy side healthy_i Find the closest peak time p on the blocking side. measured_j If the time difference |p measured_j - p healthy_i If the time is less than the preset threshold ε (preferably 50-100 milliseconds), it is determined to be a pair of matching peaks;

[0124] From T completeBegin scanning backwards, and record the moment when N consecutive pairs (preferably 3-5 pairs) of matching peaks first appear as t3. This indicator reflects the restoration of synchronicity in bilateral pulse propagation, signifying that the hemodynamic function on the compressed side has essentially returned to normal.

[0125] Step 3: Validation of candidate recovery times

[0126] The validity of each candidate recovery time t1, t2, and t3 is verified using a combination of at least two of the following verification methods, and invalid or abnormal candidate values ​​are automatically eliminated:

[0127] Value range verification: Remove candidate values ​​that exceed the preset physiological reasonable range (e.g., 0.5 seconds to 30 seconds);

[0128] Waveform stability verification: Remove candidate values ​​whose average signal quality index of the signal segment on which the calculation is based is lower than 0.5;

[0129] Peak detection rate verification: Remove candidate values ​​with peak detection rates lower than a preset threshold (e.g., less than 30 peaks per minute);

[0130] Cross-indicator consistency pre-verification: Remove suspicious values ​​that differ from other valid candidate values ​​in the same batch by more than a preset threshold (e.g., 5 seconds).

[0131] After validity verification, the set of valid candidate recovery times T is obtained. valid = {t i | i=1..k}, where k is the number of valid indicators (0≤k≤3).

[0132] Step 4: Adaptive adjustment of dynamic threshold and blocking parameters

[0133] To further enhance the algorithm's personalized adaptability, this invention introduces an adaptive adjustment mechanism:

[0134] (1) Adaptive adjustment of threshold coefficients: The coefficients α, β, γ, and δ in step 2 above can be dynamically set according to the individual physiological characteristics of the user. For example, a user profile database can be established, and appropriate coefficient values ​​can be predicted by looking up tables or regression models based on parameters such as age, gender, blood pressure level, and BMI. For young and healthy people, smaller α and β values ​​(such as 0.7 and 0.2) can be selected to capture rapid signal recovery; for the elderly or people with hypertension, larger α and β values ​​(such as 0.9 and 0.4) can be selected to avoid premature judgment caused by noise.

[0135] (2) Adaptive adjustment of blocking parameters: Before each measurement, the blocking pressure and blocking duration of the blood blocking device are adaptively adjusted based on the user's historical blood pressure level or real-time PPG signal characteristics. For example, if the user's historical systolic blood pressure is high, the blocking pressure is appropriately increased to ensure sufficient blocking; if the real-time PPG signal shows a fast heart rate, the blocking duration can be appropriately shortened to reduce discomfort. This adaptive adjustment can be achieved through closed-loop control: the data processing terminal sends adjustment instructions to the blood pressure monitor based on the user's profile or real-time analysis results, and the blood pressure monitor executes the personalized blocking scheme accordingly.

[0136] Step 5: Dynamic Adaptive Fusion Strategy

[0137] Based on the effective candidate recovery time set T valid The number of valid values ​​k and their differences are used to output the final recovery time T after blood stasis using hierarchical decision logic. final :

[0138] Case A (k=0): All indicators are invalid, the detection is judged as a failure, an invalid result is output, and possible causes (such as motion interference, poor signal quality, etc.) are prompted through the user interface, suggesting that the measurement be repeated.

[0139] Case B (k=1): There is only one valid value, indicating that the index has reliable feature recognition capability under the current signal conditions. This valid value is directly used as the final result T. final Output.

[0140] Case C (k≥2): There are multiple valid values. Calculate the maximum difference between the valid values, D = max(T). valid )- min(T valid The preset difference threshold θ is preferably 2 seconds (adjustable according to clinical application needs):

[0141] Case C1 (D ≤ θ): All indicators are highly consistent, the recovery process is stable and reliable, and the output is the arithmetic mean of each effective value, or the weighted average of the values ​​dynamically weighted based on the signal quality on which each indicator is calculated.

[0142] Case C2 (D > θ): The index is drifting due to interference, requiring further analysis. Iterate through all pairwise combinations from all valid values, calculate the difference between each pair, and select the pair with the smallest difference (T). a , T b The arithmetic mean of these arithmetic values ​​is used as the final result T. final This strategy can automatically exclude abnormal drift indicators affected by sudden noise.

[0143] Step 6: Multimodal sensor fusion

[0144] To further improve the accuracy and robustness of detection, this invention supports multimodal sensor fusion. In addition to the PPG signal, at least two of the following signals can be acquired simultaneously:

[0145] Electrocardiogram (ECG) signals: acquired via wearable electrodes for accurate detection of heart rate and heart rate variability;

[0146] BCG signal: acquired by a piezoelectric sensor or accelerometer, reflecting the mechanical activity of the heart;

[0147] Heart sound signal (PCG): Acquired through an electronic stethoscope and used to identify heart sound characteristics;

[0148] Tonometry: Acquired by a pressure sensor, reflecting the arterial pressure waveform;

[0149] Ultrasonic Doppler blood flow signal: acquired through a miniature ultrasound probe, directly measuring blood flow velocity.

[0150] Multimodal signal fusion methods include:

[0151] Feature-level fusion: Extract recovery-related features (such as the recovery of R-wave amplitude in ECG and the recovery of the second heart sound intensity in PCG) from each modal signal, generate candidate recovery times, and then perform joint verification and fusion according to steps 3-5.

[0152] Decision-level fusion: After each modality independently calculates its recovery time, a comprehensive decision is made using a voting or weighted average method.

[0153] Signal-level fusion: After time alignment of multimodal signals, the signals are input into a deep learning model to directly predict the recovery time end-to-end.

[0154] based on Figure 6 As can be seen from the process, when outputting the final post-blockage recovery time, in the blood post-blockage recovery time detection method proposed in this invention, the post-blockage recovery time is jointly determined by multiple independent physiologically relevant indicators, including three post-blockage recovery times obtained based on the recovery trend of the mean heart rate, based on bilateral peak pairing, and based on signal quality changes. The quality threshold method and the heart rate method use the first point that reaches the specified quality value / heart rate value, while the bilateral peak pairing method finds the first paired peak of the two channels, the latter only affecting the two channels.

[0155] Down Figure 7 and Figure 8 This invention demonstrates the calculation rules for signal quality and heart rate error thresholds used in this invention. When using only one channel, the stabilized result is selected as the reference segment; when using two channels, the non-blocking hand is selected as the reference segment. The threshold method is used to locate heart rate and signal quality during the blood blockage phase and the blood stabilization phase.

[0156] superior Figure 7 and Figure 8 This paper demonstrates methods for calculating heart rate and signal quality thresholds from the blood clotting phase to the blood recovery phase using both single-channel and dual-channel methods. A potential post-clotting recovery time is defined as the time point at which this condition is met. The post-clotting recovery time calculated based on signal quality is denoted as t1, and the post-clotting recovery time calculated based on heart rate is denoted as t2.

[0157] This invention proposes a strategy for dynamically fusing three types of post-blocking recovery times: when the difference (i.e., dispersion) between the three post-blocking recovery times exceeds a preset threshold, unreliable indicators are automatically eliminated, and only the two most reliable and consistent post-blocking recovery times are fused. The calculation process is as follows: Figure 9 As shown.

[0158] Furthermore, since the physiological mechanisms, signal characteristics, and calculation processes underlying the three types of indicators are not entirely the same, in actual data acquisition environments, the three time points may exhibit three types of results: consistent, approximately consistent, or significantly dispersed. To ensure that the determination of the final recovery time after the blockade is both physiologically logical and highly robust, this rule is proposed based on a comprehensive consideration of physiological dynamics and signal processing theory. After blood pressure measurement, the blockade effect caused by cuff deflation gradually subsides, and the local blood perfusion, pulse wave morphology, and heart rate regulation of the measured hand will all return to the individual's original physiological state. Since the recovery process is continuous, monotonous, and predictable, the actual recovery time after the blockade should have a high degree of consistency among the calculated results of different indicators.

[0159] If no valid candidate values ​​are found, the core reason is that the quality of the physiological signal does not meet the analysis requirements. This could be due to user limb movement during the detection process, causing motion artifacts that interfere with the waveform integrity of signals such as PPG; it could also be due to factors such as fluctuations in ambient light intensity or poor contact between the sensor and the skin, resulting in a low signal-to-noise ratio; in addition, extreme physiological conditions or temporary equipment malfunctions may also cause signal feature extraction failure, ultimately making it impossible to locate the effective recovery time after blood puncture. If only one valid candidate value is found, it indicates that the indicator has reliable feature recognition capabilities under the current signal conditions (interference and abnormal situations have been excluded), and its output recovery time candidate value can truly reflect the recovery status of blood perfusion. Therefore, this valid value is directly defined as the final recovery time after blood puncture. If two valid candidate values ​​are found, since the two indicators characterize the blood recovery process from different dimensions (such as signal quality recovery, heart rate stability, pulse wave feature matching, etc.), each candidate value has a certain degree of reliability. The feature information of the two dimensions can be combined to offset the random errors that may exist in a single indicator, improving the stability and accuracy of the results.

[0160] If the three post-block recovery times are very close to each other, it indicates a stable recovery process and that all three indicators are reliable. In this case, the average of the three can be directly used as the final post-block recovery time. However, in actual measurements, one indicator may experience abnormal drift, false detection, or absence, causing its corresponding post-block recovery time to deviate significantly from the other two indicators. If this is not eliminated and the average or minimum value of the three is directly used, it may lead to incorrect judgment. Therefore, this invention introduces a post-block recovery time dispersion judgment mechanism: when the difference between the maximum and minimum values ​​of the three post-block recovery times is greater than a preset threshold (e.g., 2 seconds), it indicates that the three no longer represent a stable and consistent physiological recovery trend. In this case, further analysis of the relative closeness between the three time points is required. Specifically, this invention automatically identifies the two closest post-block recovery times by calculating the difference between each pair of time points. Therefore, this invention considers the two closest post-block recovery times as the most reliable indicators and uses their average as the final post-block recovery time. This retains the advantages of multi-indicator fusion while effectively eliminating misjudgments caused by sudden noise.

[0161] Example 2: Subgroup Performance Study

[0162] This embodiment verifies the performance of the blood pressure measurement recovery time detection method proposed in this invention in different populations. A grouped study approach is used to evaluate the stability, repeatability, and anti-interference ability of the calculated recovery time. This embodiment further illustrates the feasibility and clinical application potential of the system and algorithm of this invention. This embodiment analyzes multiple rounds of blood pressure measurement data collected from multiple subjects. Each subject needs to wear an electronic blood pressure monitor and two smart rings. The system has accurate blood pressure event timestamps (e.g., "Start Measurement" and "Complete Measurement"). Table 1 below shows the blood pressure recovery time measurement results for users under different conditions, including t1, t2, and t3 calculated sequentially by three methods, and the final estimated index:

[0163] Table 1. Recovery time after blocking multiple users (unit: seconds)

[0164]

[0165] Table 1 above shows the calculated recovery time after blood stasis in various scenarios. For the first user, all three methods return an indicator, so the output is the average of the three. For the second and fourth users, only one valid candidate value is available, so this value is output as the recovery time after blood stasis. For the third and fifth users, two subsequent values ​​are available, and the average of the two is output as an estimate. Based on the characteristics of the peak pairing method, it will generally output a valid value, unless a single peak cannot be identified. Table 2 below shows the recovery time after blood stasis for different age groups, genders, and BMI levels.

[0166] Table 2. Statistical results of recovery time after blockade in different subgroups

[0167]

[0168] In this embodiment, statistical analysis was conducted on the recovery time after blood flow interruption in different subject populations to verify the universality and robustness of the method of the present invention in multiple individuals. Recovery time was grouped and statistically analyzed based on basic characteristics such as age, gender, and body mass index (BMI). First, regarding age grouping, the results show that younger groups generally exhibited shorter and more stable recovery times after blood flow interruption, while individual differences increased with age, but the overall trend remained consistent. Regarding gender grouping, the absolute value and dispersion of recovery time after blood flow interruption were slightly greater in the male group than in the female group, indicating that there may be some differences in recovery time after blood flow interruption at the gender level. From the perspective of BMI, overall, changes in BMI only showed a certain degree of difference, but did not cause significant systemic effects. Regarding blood pressure classification, blood pressure classification showed some differences in recovery time after blood flow interruption, but the recovery time detection method of the present invention could obtain stable outputs in all blood pressure classification groups. Among the groups with blood pressure grades 1 and 2, the recovery time after blockade was similar. The group with blood pressure grade 3 showed a more concentrated recovery time with the smallest fluctuation range, indicating that vascular compliance and reperfusion processes were relatively consistent in this group. In higher blood pressure grades, the average recovery time after blockade increased, and the standard deviation was larger, suggesting that individual differences in the reperfusion process after blockade were more significant in hypertensive individuals, possibly influenced by physiological factors such as decreased vascular elasticity and weakened microcirculation regulation. However, even in this group, the method of this invention could still stably detect t1, t2, or t3 and complete reliable fusion, demonstrating the algorithm's tolerance to physiological differences.

[0169] In summary, the subgroup statistical results of this embodiment demonstrate that the post-blocking recovery time detection method of the present invention can be applied to subject groups of different ages, genders, and blood pressure levels, and can effectively adapt to users with different physiological characteristics. This embodiment further proves the universality of the present invention under multiple physiological conditions and different vascular states, providing strong support for its application in long-term monitoring, smart wearable devices, and hemodynamic assessment.

[0170] Example 3: Study on the Influence of Posture on a Postural Recovery Time Detection System Based on Post-Blockage

[0171] This embodiment aims to verify whether there are regular differences in the blood recovery process of PPG signals after blood pressure measurement under different body postures (sitting, standing and lying), and to further illustrate the applicability, stability and robustness of the post-blocking recovery time detection method proposed in this invention under different posture conditions. In order to study the influence of posture changes on blood flow recovery dynamics, this embodiment records the blood pressure measurement and recovery process under the following three typical postures on the same subject: (1) Sitting posture: The subject sits on a stable chair with the backrest relaxed and both feet flat on the ground; the measuring hand is naturally flat and kept at the same height as the heart, and the first PPG acquisition device is worn; the non-measuring hand wears the second PPG acquisition device. (2) Standing posture: The subject stands with both feet flat and maintains an upright position; the measuring hand hangs naturally and is raised to the level of the heart for blood pressure measurement; the PPG device wearing procedure is the same as that for the sitting posture. (3) Lying posture: The subject lies flat on the bed; both upper limbs are naturally flat on both sides of the body; the PPG device wearing procedure is the same as that for the sitting posture.

[0172] Table 3. Statistical results of recovery time after trepanation in different postures.

[0173]

[0174] Table 3 includes multiple repeated measurements for each of the three body positions to ensure data stability. Furthermore, when exploring the effect of different postures on recovery time after blood stasis, only the posture factor changed; the rest of the data collection and alignment procedures remained completely consistent, ensuring the controllability of experimental conditions. Combining the experimental results in Table 3, differences exist between different postures. The standing posture resulted in a longer recovery time after blood stasis, with greater variation in its distribution. In contrast, the sitting and lying postures showed relatively similar recovery times after blood stasis, with a more uniform distribution. This experimental result verifies that there are regular differences in the recovery time after blood stasis under different body positions (sitting, standing, and lying), and further illustrates the applicability, stability, and robustness of the method of this invention under different posture conditions.

[0175] Based on the above embodiments, the technical solution of the present invention is not limited to the specifically described steps, device combinations, or algorithm structures. To enhance the scope of protection of the present invention and prevent others from circumventing the core objective of the invention in different forms, the following alternative or equivalent implementation methods are proposed:

[0176] (1) Sensor System Alternatives: In the post-blocking recovery time detection technology proposed in this invention, the embodiment uses a smart ring as the main data acquisition device to acquire PPG signals. However, the method of this invention is not limited to a specific form or wearing method of the device. As long as the device used can acquire PPG signals with sufficient signal-to-noise ratio, sampling rate, and stability, it can support the post-blocking recovery time detection process described in this invention. Therefore, this invention is also applicable to a variety of wearable or portable devices with PPG or blood flow-related physiological signal acquisition capabilities, including but not limited to the following alternative forms: smartwatches, smart earphones (earrings or ear hooks), and pulse oximeters (finger clips or wrist clips). In addition, this invention can be extended to various new or special forms of PPG acquisition terminals, including but not limited to: new non-contact sensing terminals such as remote PPG (rPPG) detection devices, and flexible devices such as clothing, mattresses, and medical clips with physiological optical sensing capabilities.

[0177] (2) Substitution of Sensor Location and Type: The embodiments mainly use PPG sensors worn on the fingers as a means of blood health detection, but the post-blocking recovery time detection method of the present invention is not limited to this location and optical band. In terms of sensor location, as long as it can stably reflect hemodynamic changes, the present invention is also applicable to sensors worn on skin areas with good blood flow signals, such as earlobes, wrists, ankles, toes, forearms, and backs of hands. In terms of sensor type, the present invention is not only applicable to green light, but also to any optical, electrical, or mechanical sensor system that can reflect blood perfusion, pulse propagation, or blood flow recovery trends, such as: electrocardiogram, cardiac impaction, heart sound signal, pulse pressure wave, ultrasound, and other cardiovascular health measurement methods. For wearable devices with multi-sensor fusion capabilities, the present invention can also utilize their combined signals to better assist in improving the accuracy and robustness of post-blocking recovery time detection. Therefore, the post-blocking recovery time detection method proposed in the present invention can be implemented in various sensor locations and on various types of hardware platforms, without being limited by specific sensor types, wearing positions, or physical measurement methods, and has high adaptability and scalability.

[0178] (3) Alternative solutions to the recovery feature system: This invention is based on three independent indicators: signal quality, heart rate, and paired peaks. However, alternative forms are also allowed: any one or two indicators can be used to determine the recovery time after blood blockage. Based on this invention, the number of indicators can be expanded to more than four, for example, by adding: pulse amplitude recovery point, first or second derivative recovery point of waveform, and heart rate variability (HRV) recovery indicator. Alternatively, the indicator fusion method can be replaced by: weighted average method, minimum reliable indicator method, multi-indicator voting method, or fusion strategy based on Bayesian confidence interval. Therefore, regardless of the number of indicators or the fusion method, as long as the aim is to obtain the recovery time after blood blockage, it falls within the protection scope of this invention.

[0179] (4) Alternatives to the blocking method and blocking parameters: Although the embodiments use an electronic blood pressure cuff blocking mode, this method is also applicable to devices with blood blocking functions, such as manual blood pressure monitors, ambulatory blood pressure monitoring devices (ABPM), and fingertip optical volume compression devices. Therefore, the blocking method does not affect the realization of the core idea of ​​this invention. The embodiments of this invention may use fixed blocking parameters (such as blocking pressure and blocking duration), but adaptive adjustment methods of the blocking parameters are equivalent alternatives, specifically including adaptive blocking pressure and adaptive blocking duration adjustment. The above-mentioned alternatives are consistent with this invention in terms of technical concept, implementation principle, and technical effect.

[0180] (5) Equivalent substitution of signal preprocessing and denoising methods: In this embodiment of the invention, a specific signal preprocessing process (such as filtering, denoising, and baseline correction) is adopted, but the core detection logic is not limited by the preprocessing method. As long as the preprocessed data can meet the accuracy requirements of the recovery time analysis after blocking, any reasonable signal optimization method is an equivalent substitution. Specifically, this includes the following: substitution of filtering algorithms, and substitution of signal alignment and standardization. The difference in preprocessing methods only affects the signal quality optimization path and does not change the core concept.

[0181] (6) Expansion and substitution of suitable populations and application scenarios: The core technical solution of this invention can be extended to different populations and scenarios, as long as the application purpose is still to extend the recovery time after blood blockage. This is an equivalent substitution, specifically including the expansion of suitable populations: from ordinary healthy people to special populations, such as critically ill children, the elderly, pregnant women, and patients with vascular diseases (such as arteriosclerosis, thrombosis risk groups), etc., only requiring adjustment of algorithm parameters (such as signal quality threshold, heart rate threshold) according to the physiological characteristics of the population; and the expansion of application scenarios: from the scenario of this embodiment to clinical monitoring scenarios, sports and health scenarios, occupational health scenarios, etc.

[0182] In summary, this invention focuses on the automated and precise detection of recovery time after blood flow interruption to obtain health information on blood flow. It proposes a systematic technical solution comprised of hardware collaboration, time synchronization, multi-indicator detection, and dynamic fusion, including the following five aspects:

[0183] (1) Construction of a non-invasive blood flow detection system: This invention constructs a post-blockage recovery time detection system that works in concert with a blood blocking device and a blood measurement device, realizing a complete closed loop of blood blocking-blood flow detection-post-blockage recovery time calculation-blood flow analysis. In Examples 1 and 2, the cuff-type blood pressure monitor is responsible for outputting blood blocking and measurement events, the smart ring collects blood flow change signals in real time, and the data terminal performs time synchronization, signal processing, and post-blockage recovery time calculation.

[0184] (2) PPG-based recovery time detection technology after blood occlusion: Utilizing the high sensitivity of PPG signals to changes in blood perfusion, the dynamic process of peripheral blood vessels from "occlusion" to "reperfusion" after occlusion is captured. Through signal quality assessment, heart rate extraction, and peak feature recognition of PPG signals, the recovery process is quantitatively characterized, breaking through the traditional application limitation of PPG signals only for steady-state physiological quantity measurement.

[0185] (3) Precise timestamp alignment technology for blood flow interruption events: The electronic blood pressure monitor outputs the precise timestamp of the measurement event in real time, and the PPG sensor records the original signal with timestamp at a sampling frequency of not less than 100Hz; the data processing terminal automatically extracts the PPG segment from the end of the measurement to the start of the next measurement through timestamp matching, accurately locks the blood flow recovery stage, and realizes automated analysis of batch data.

[0186] (4) Multi-index recovery time determination algorithm system: In view of the one-sidedness of single index determination, three independent physiological indicators are designed to realize the multi-dimensional characterization of recovery time after blockage, forming a complementary verification determination system: ① Signal quality threshold method: The average channel difference of signal quality at the time of blockage and the time of stabilization is used as the dynamic threshold to locate the starting point of signal quality recovery (t1); ② Heart rate average recovery method: The average channel difference of heart rate at the time of blockage and the time of stabilization is used as the dynamic threshold to locate the heart rate average recovery point (t2); ③ Bilateral peak pairing method: The "nearest neighbor time matching algorithm" is used to determine the synchronicity of the appearance time of bilateral pulse wave peaks and locate the recovery point (t3). The three indicators characterize the recovery process from different physiological dimensions and improve the reliability of the results.

[0187] (5) Multi-indicator dynamic fusion strategy: To address the potential discrepancies in the results of multiple indicators, a dynamic rule of effective indicator screening, difference determination, and effective value fusion is introduced: If there is no effective indicator, it may indicate that the blood has not recovered for a long time or that the signal-to-noise ratio is extremely low; if there is only one effective indicator, the value is directly adopted; if there are multiple effective indicators, the difference determination is made and the effective values ​​are fused to output the final result, which can avoid misjudgment caused by a single indicator to a certain extent and improve the accuracy and reliability of the recovery time estimation after blood blockage.

[0188] (6) Dynamic management and validity verification technology of dual-sided PPG channels: Overcoming the shortcomings of existing single-channel PPGs that are susceptible to compression and noise interference, an innovative solution of simultaneous acquisition of dual-sided fingers and linkage of measurement hand information is proposed: Through the measurement hand information output by the electronic blood pressure monitor, the corresponding PPG channel on the pressurized side is automatically identified, and the intelligent allocation of pressurized side-reference side is completed; at the same time, based on the PPG signal of the non-measured hand, common noises such as motion artifacts and ambient light are eliminated through signal quality comparison and heart rate error calculation, solving the problem of analysis failure caused by misjudgment of single-channel signals, and providing reliable data input for the calculation of recovery time after blockage.

[0189] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention. Where there is no conflict, the above embodiments and features described therein can be combined with each other.

Claims

1. A non-invasive real-time blood flow assessment method based on recovery time after blood stasis, characterized in that, Includes the following steps: Acquire at least one event timestamp output by the blood clogging device, the event timestamp corresponding to the moment when the blood clogging device finishes applying pressure, and a time-stamped physiological signal continuously acquired by at least one blood health measurement module; Based on the event timestamp, signal segments corresponding to the recovery phase after blood flow occlusion are automatically extracted from the physiological signals. The signal segment is analyzed to calculate at least two independent indicators that characterize the recovery process from different physiological dimensions, and each independent indicator outputs a candidate recovery time after blockage. The at least two independent indicators are selected from: indicators based on signal quality recovery characteristics, indicators based on heart rate stability recovery characteristics, and indicators based on bilateral pulse wave synchronicity recovery characteristics. The validity of each candidate recovery time after blocking is verified. Based on at least one criterion of signal reliability, noise interference level and cross-index consistency, invalid or abnormal candidate values ​​are eliminated to obtain a set of valid candidate recovery times. Based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them, a dynamic fusion strategy is adopted to output the final recovery time after blood blockage, and blood flow is evaluated based on the recovery time after blood blockage.

2. The non-invasive real-time detection method for recovery time after blood flow interruption according to claim 1, characterized in that, When the blood health measurement module is worn only on the blocked side, the index based on signal quality recovery characteristics is calculated in the following way: The average signal quality of the first signal during the blood blockage phase and the average signal quality of the second signal during the blood stabilization phase in the signal segment are identified, and the difference between the average signal quality of the second signal and the average signal quality of the first signal is used as a dynamic threshold. A sliding window is used to scan the signal quality curve backward from the time point corresponding to the event timestamp. The moment when the average signal quality within the sliding window first reaches or exceeds the sum of the first signal quality average and the dynamic threshold is recorded as the first candidate recovery time t1.

3. The non-invasive real-time detection method for recovery time after blood flow interruption according to claim 1, characterized in that, When the blood health measurement module is worn on both the blocked side and the healthy side, the index based on signal quality recovery characteristics is calculated using the following noise reduction method: Calculate the signal quality difference between the blocked side and the healthy side; The mean value of the signal quality difference between the channels during the blood blockage period was used as the benchmark; A sliding window is used to scan backwards from the time point corresponding to the event timestamp. The moment when the signal quality difference between channels first drops below the set ratio of the benchmark within the sliding window is recorded as the first candidate recovery time t1.

4. The non-invasive real-time detection method for recovery time after blood flow interruption according to claim 1, characterized in that, When the blood health measurement module is worn only on the blocked side, the index based on heart rate stability recovery characteristics is calculated in the following way: The first mean heart rate during the blood blockage period and the second mean heart rate during the blood stabilization period are identified in the signal segment, and the absolute value of the difference between the second mean heart rate and the first mean heart rate is used as a dynamic threshold. A sliding window is used to scan the heart rate curve backward from the time point corresponding to the event timestamp. The moment when the average heart rate within the sliding window first enters the dynamic threshold interval centered on the second average heart rate and remains stable for at least three consecutive sliding windows is recorded as the second candidate recovery time t2.

5. The non-invasive real-time detection method for recovery time after blood stasis according to claim 1, characterized in that, When the blood health measurement module is worn on both the blocked side and the healthy side, the index based on heart rate stability recovery characteristics is calculated using the following noise reduction method: Calculate the bilateral heart rate error between the blocked side and the healthy side to eliminate systemic heart rate fluctuation interference; The mean of the bilateral heart rate errors during the blood stasis period was used as the benchmark; A sliding window is used to scan backwards from the time point corresponding to the event timestamp. The moment when the bilateral heart rate error first drops below the set proportion of the benchmark within the sliding window is recorded as the second candidate recovery time t2.

6. The non-invasive real-time detection method for recovery time after blood flow interruption according to claim 1, characterized in that, When the blood health measurement module is worn on the blocked side and the healthy side respectively, the index based on the bilateral pulse wave synchronicity recovery characteristics is calculated in the following way: The positions of all peaks in the PPG signals on the blocked side and the healthy side were detected separately, generating two peak time series. The nearest neighbor time matching algorithm is used to pair the peaks on both sides. Peaks with a time difference of less than a preset threshold are identified as matching peak pairs. The preset threshold is 50-100 milliseconds. Starting from the time point corresponding to the event timestamp, scan backwards and record the moment when N consecutive matching peak pairs first appear as the third candidate recovery time t3, where N is 3-5.

7. The non-invasive real-time detection method for recovery time after blood stasis according to claim 1, characterized in that, The validity verification of the recovery time after blocking for each candidate includes a combination of at least two of the following verification methods: Value range verification: Remove candidate values ​​that exceed the preset physiological reasonable range of 0.5 seconds to 30 seconds; Waveform stability verification: Remove candidate values ​​whose average signal quality index of the signal segment on which the calculation is based is lower than 0.5; Peak detection rate verification: Remove candidate values ​​with peak detection rates lower than a preset threshold; Cross-indicator consistency pre-verification: Remove suspicious values ​​that differ from other valid candidate values ​​in the same batch by more than a preset threshold.

8. The non-invasive real-time detection method for recovery time after blood stasis according to claim 1, characterized in that, The step of outputting the final recovery time after blood stasis using a dynamic adaptive fusion strategy based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them specifically includes: Case A: If the set of valid candidate recovery times is empty, the detection is deemed to have failed, an invalid result is output, and the user is prompted. Case B: If there is only one valid value in the set of valid candidate recovery times, then that valid value will be used as the final result. Case C: If there are two or more valid values ​​in the set of valid candidate recovery times, then calculate the maximum difference between the valid values ​​D = max(T) valid )-min(T valid ); Case C1: If D ≤ preset difference threshold θ, then it is determined that the indicators are highly consistent, and the arithmetic mean of each effective value or the weighted average based on the dynamic weighting of signal quality is output. Case C2: If D > preset difference threshold θ, it is determined that the index is disturbed and drifts. The two most similar valid values ​​are selected from all valid values, and the arithmetic mean of the two valid values ​​is output. The preset difference threshold θ is 2 seconds.

9. The non-invasive real-time detection method for recovery time after blood stasis according to claim 1, characterized in that, The step of outputting the final recovery time after blood stasis using a dynamic fusion strategy based on the number of valid values ​​in the set of valid candidate recovery times and the degree of difference between them specifically includes: If the set of valid candidate recovery times is empty, then output an invalid result; If there is only one valid value in the set of valid candidate recovery times, then that valid value will be used as the final result. If there are two or more valid values ​​in the set of valid candidate recovery times, the maximum difference between the valid values ​​is calculated. If the maximum difference is less than or equal to a preset threshold, the arithmetic mean or weighted average of each valid value is output. If the maximum difference is greater than a preset threshold, then the two most similar valid values ​​are selected from all valid values, and the arithmetic mean of the two valid values ​​is output.

10. A non-invasive real-time detection system for recovery time after blood flow interruption, characterized in that, include: The blood flow blocking module is used to apply external pressure to the target limb to form a blood flow blocking state, and to release the external pressure at a predetermined time and output event information with a precise timestamp. The blood health measurement module is used to continuously collect physiological signals characterizing the peripheral blood perfusion status during the process of blood flow occlusion and release. It includes at least one wearable sensor with physiological signal acquisition functions such as photoplethysmography pulse wave, which is used to continuously collect and output physiological signals with precise timestamps. The data processing terminal is communicatively connected to the blood blocking module and the blood health measurement module, and has a built-in signal processing unit and the detection method as described in any one of claims 1-9, for receiving data, performing time alignment, signal analysis and result output.

11. The non-invasive real-time detection system for recovery time after blood stasis according to claim 10, characterized in that, The blood health measurement module and the blocking component of the blood blocking module are deployed on the same side of the limb as the heart, with the blood health measurement module located at the distal end of the limb and the blocking component located at the proximal end of the limb.

12. The non-invasive real-time detection system for recovery time after blood stasis according to claim 10, characterized in that, The blood health measurement module includes two PPG collection units, which are worn on the blocked limb and the healthy limb respectively, forming a bilateral control collection layout.

13. The non-invasive real-time detection system for recovery time after blood stasis according to claim 10, characterized in that, The data processing terminal automatically identifies and marks the blocked side channel and the healthy side channel based on the measurement hand information output by the blood blocking module, and uses the healthy side channel signal as a dynamic reference baseline for recovery time analysis.

14. The non-invasive real-time detection system for recovery time after blood stasis according to claim 10, characterized in that, The blood health measurement module is selected from one or more of the following wearable devices: smart ring, smartwatch, smart bracelet, finger clip pulse oximeter, smart earphone, clothing or medical clip with PPG acquisition function; the physiological signal includes at least one signal that can characterize the peripheral blood perfusion state or has a definite correlation with its changes, the signal includes photoplethysmography (PPG) signal, blood oxygen signal, skin temperature signal or a combination of multiple signals; the data processing terminal is selected from one of the following devices: smartphone, tablet computer, personal computer, cloud server, edge computing node, embedded processor of the wearable sensor body.

15. A wearable device, as a blood health measurement module in the system of any one of claims 10-15, characterized in that, include: Physiological signal sensor, used to collect physiological signals such as raw photoplethysmography pulse waves; Analog-to-digital converter module, used to convert analog signals into digital signals; The timestamp calibration module is used to ensure that each sampling point has an accurate timestamp that is consistent with the time reference of the blood blocking device; The signal preprocessing module is used to filter, denoise, and perform baseline correction on the raw signal; The data transmission module is used to transmit time-stamped physiological signals to the data processing terminal.