CN-122004806-A - Wearable vital sign monitoring system capable of synchronously collecting multiple parameters

CN122004806ACN 122004806 ACN122004806 ACN 122004806ACN-122004806-A

Abstract

The invention relates to the technical field of wearable vital sign monitoring systems with synchronously acquired multiple parameters, and particularly discloses a wearable vital sign monitoring system with synchronously acquired multiple parameters. The system comprises a multi-mode physiological sensing array, an edge calculation and fusion module, a psychological state quantification engine and a closed-loop nerve regulation module. The invention introduces a feedback learning and model updating mechanism based on reinforcement learning, so that the system has the capability of continuous self-optimization. The system can gradually adapt to the physiological and psychological characteristic differences of different users by recording the parameters and effects of each intervention and dynamically optimizing the stimulation parameter mapping table by using a strategy gradient method. The intervention strategy is evolved from static and universal setting to dynamic and personalized schemes, so that the intervention accuracy and the comfort and acceptance of users are continuously improved in long-term use, and technical support is provided for realizing personalized mental health daily maintenance.

Inventors

XU XUYAO
CHEN LIXIONG
LIN JINYANG
ZHU ZHIJIE

Assignees

常州三水救生装备有限公司

Dates

Publication Date: 20260512
Application Date: 20260306

Claims (10)

1. Wearable vital sign monitoring system of synchronous collection of many parameters, its characterized in that includes: the multi-mode physiological sensing array is used for parallelly collecting multi-dimensional physiological original signals of a user with millisecond time synchronization precision; The edge calculation and fusion module is used for preprocessing and extracting characteristics of the multidimensional physiological original signals acquired by the multi-mode physiological sensing array, and performing space-time alignment and characteristic level fusion of multi-source heterogeneous data so as to generate fusion characteristic vectors; the psychological state quantization engine is used for receiving the fusion characteristic vector generated by the edge calculation and fusion module, calculating through a preset psychological state quantization model and outputting a quantization index representing the current psychological state of the user; And the closed-loop nerve regulation module is used for receiving the psychological state quantization index output by the psychological state quantization engine, and automatically generating and executing non-invasive nerve regulation stimulation matched with the deviation direction and amplitude of the psychological state quantization index when the psychological state quantization index exceeds a preset steady-state interval so as to guide the physiological index to return to the steady-state interval.
2. The multi-parameter synchronous acquisition wearable vital sign monitoring system of claim 1, wherein the multi-modal physiological sensing array integrates at least 4 types of physiological sensors including a photoplethysmography sensor, a galvanic skin activity sensor, a tri-axial accelerometer and gyroscope combination sensor, and an infrared thermal radiation sensor; the photoelectric volume pulse wave sensor takes a light emitting diode as a light source, and a photoelectric detector receives light intensity signals reflected or transmitted by human tissues and is used for extracting time domain and frequency domain characteristics of heart rate and heart rate variability; The skin electric activity sensor adopts a constant voltage type measuring circuit, contacts with the skin part of a user through a pair of silver-silver chloride electrodes and is used for measuring the skin conductivity level and the fluctuation frequency thereof; The triaxial accelerometer and gyroscope combined sensor is used for continuously monitoring limb movement acceleration and angular velocity of a user, eliminating interference of movement artifacts on photo-capacitive product pulse wave signals through a built-in gesture resolving algorithm, and extracting movement characteristics representing activity intensity and behavior mode of the user; the infrared thermal radiation sensor is aligned to the body surface area of the user in a non-contact mode, and the body surface temperature and the dynamic change trend thereof are inverted by measuring the intensity of the received infrared radiation.
3. The wearable vital sign monitoring system with multi-parameter synchronous acquisition according to claim 2, wherein the edge computing and fusion module comprises a signal preprocessing unit, a feature extraction unit and a data fusion unit; The signal preprocessing unit is used for respectively executing band-pass filtering, power frequency interference notch and baseline drift correction on each path of physiological original signal; the characteristic extraction unit is used for extracting at least 5 heart rate variability characteristics including the standard deviation of adjacent heart beat intervals and the ratio of low-frequency power to high-frequency power from the preprocessed photoplethysmography pulse wave signals, extracting at least 3 characteristics including the average value of skin conductance and the number of nonspecific fluctuation from the skin electric activity signals, extracting at least 2 characteristics including the activity count and the resting duration ratio from the motion signals, and extracting at least 2 characteristics including the average value of temperature and the gradient of temperature change from the body surface temperature signals; The data fusion unit is used for establishing a unified time stamp system, aligning characteristic data from different sensors to the same time reference, calculating nonlinear association weights among different modal characteristics by adopting a typical correlation analysis method based on a kernel function, and carrying out weighted summation on all the characteristics according to the calculated association weights to generate a high-dimensional fusion characteristic vector.
4. The wearable vital sign monitoring system with multi-parameter synchronous acquisition according to claim 3, wherein the exemplary correlation analysis method based on the kernel function comprises the steps of mapping physiological feature vectors from different modalities to a high-dimensional regenerated kernel hilbert space respectively, solving projection direction vectors in the high-dimensional regenerated kernel hilbert space, which maximize correlation between projections of the feature vectors of different modalities, and calculating contribution weights of the features of each modality in a fusion space according to the solved projection direction vectors.
5. The wearable vital sign monitoring system of claim 4, wherein the mental state quantization model built in the mental state quantization engine is a deep neural network model, and the deep neural network model takes the fusion feature vector as an input; the deep neural network model comprises 3 hidden layers, the number of neurons of each layer is 128, 64 and 32, and an activation function adopts a correction linear unit; The output layer of the deep neural network model is 2 neurons, and the two dimensions of the psychological state quantization index, namely the arousal degree and the titer, respectively correspond to each other, wherein the arousal degree represents the psychological activation level, and the titer represents the positive and negative orientation of emotion.
6. The multi-parameter synchronous acquisition wearable vital sign monitoring system of claim 5, wherein the closed-loop neuromodulation module comprises a stimulation decision unit and a transcranial alternating current stimulation unit; a two-dimensional psychological state stable state interval is preset in the stimulation decision unit, and the two-dimensional psychological state stable state interval is defined as an elliptical area on a two-dimensional plane formed by the awakening degree and the titer; The stimulation decision unit is used for comparing the position relation between the quantized index point output by the psychological state quantization engine and the boundary of the steady-state interval in real time, generating no output when the quantized index point is positioned in the steady-state interval, calculating the vector from the quantized index point to the central point of the steady-state interval when the quantized index point is positioned outside the steady-state interval, wherein the direction and the length of the vector respectively define the target direction and the regulation intensity requirement to be regulated; the transcranial alternating current stimulation unit is used for receiving the stimulation parameter instruction from the stimulation decision unit and generating and executing corresponding non-invasive nerve regulation stimulation.
7. The multi-parameter synchronous acquisition wearable vital sign monitoring system of claim 6, wherein the transcranial alternating current stimulation unit comprises a waveform generator, a constant current source circuit, and a pair of wearable stimulation electrodes; The waveform generator is used for receiving the stimulation parameter instruction from the stimulation decision unit and generating an alternating current signal with corresponding frequency and waveform; the constant current source circuit is used for ensuring that the intensity of the current output to the human body is accurately stabilized at a numerical value required by an instruction, and the maximum output current is not more than 2 milliamperes; The pair of wearable stimulating electrodes are made of conductive silica gel, are positioned and attached to the scalp region of a user according to the international 10-20 brain electrode placement system, and are used for safely applying weak alternating current signals to the cerebral cortex functional network.
8. The multi-parameter synchronous acquisition wearable vital sign monitoring system of claim 7, further comprising a feedback learning and model update module; The feedback learning and model updating module is used for continuously recording the parameters of each execution of the stimulus, the psychological state quantification indexes before the execution and the psychological state quantification index change track in a preset time window after the execution of the closed-loop neural adjusting module, calculating the effectiveness score of the stimulus intervention through a built-in stimulus effect evaluation algorithm, storing all the intervention records and the effectiveness score in a local intervention log database, and performing fine adjustment optimization on the stimulus parameter mapping table in the stimulus decision unit by a strategy gradient method in a reinforcement learning algorithm at regular intervals.
9. The wearable vital sign monitoring system with multi-parameter synchronous acquisition according to claim 8, wherein the calculation process of the stimulation effect evaluation algorithm is as follows, an expected adjustment direction vector formed by a psychological state quantization index point and a steady-state interval center point before stimulation execution is obtained, an actual movement vector formed by a psychological state quantization index mean value and a steady-state interval center point in a preset time window after stimulation execution is obtained, cosine similarity between the actual movement vector and the expected adjustment direction vector is calculated, and the cosine similarity is multiplied by a ratio of the actual movement vector module length to the expected adjustment vector module length to obtain a validity score between-1 and 1.
10. The multi-parameter synchronous acquisition wearable vital sign monitoring system of claim 9, wherein the operation of the system follows a hierarchical decision and control architecture; The layered decision and control architecture comprises a bottom sensing control layer, a middle state evaluation and decision layer and a high-level intervention execution and learning layer; the bottom sensing control layer operates at a sampling rate of 100 Hz and is responsible for driving the multi-mode physiological sensing array, collecting original signals and performing primary filtering; the middle layer state evaluation and decision layer operates at the frequency of 1 Hz, and performs feature extraction, data fusion, mental state quantitative calculation, steady-state interval comparison and stimulation decision; The high-level intervention execution and learning layer operates in an event-driven mode, activates the closed-loop nerve regulation module only when a stimulation instruction is received, and operates the updating flow of the feedback learning and model updating module in a period of days or weeks.

Description

Wearable vital sign monitoring system capable of synchronously collecting multiple parameters Technical Field The invention belongs to the technical field of medical monitoring, and particularly relates to a wearable vital sign monitoring system capable of synchronously collecting multiple parameters. Background Vital sign monitoring is an important basis in the fields of medical health, sports science and daily health management, and the health condition of individuals is assessed by continuously and accurately collecting physiological parameters of the human body. With the rapid development of sensor technology, microelectronic technology and wireless communication technology, wearable equipment has become a mainstream technology carrier for realizing noninvasive and continuous vital sign monitoring, and is widely applied to daily tracking of basic physiological indexes such as heart rate, blood oxygen, body temperature, activity and the like. Wearable monitoring systems oriented to mental health assessment and intervention are hotspot directions of current research and application. The system aims at indirectly deducing the psychological state of the user, such as the stress level, the emotion fluctuation or the anxiety degree, through analyzing the physiological signals collected by the wearing equipment, so as to provide objective basis for early identification and active intervention of psychological health. The prior art mainly carries out psychological state assessment through single or few physiological parameters, and the method has limitations. The evaluation of the psychological state by the existing wearable equipment usually depends on retrospective questionnaires or alarms based on simple threshold rules, the subjective performance and hysteresis of the evaluation process are strong, and the real-time and objective quantification of the psychological fluctuation cannot be realized. The existing system functions are stopped at the steps of 'monitoring' and 'early warning', and an automatic adjusting and intervening mechanism linked with an evaluation result is lacked, so that an open loop of 'diagnosis only and no treatment' is formed. The monitoring and intervention link is split, so that when the system cannot detect abnormal psychological states of the user, an effective adjusting means can be provided immediately to relieve negative states, and therefore complete closed-loop management from state perception to active adjustment is difficult to achieve, and practical utility of the system in preventing psychological problem deterioration and improving daily psychological health level is limited. Disclosure of Invention The invention aims to provide a wearable vital sign monitoring system for synchronously collecting multiple parameters, which aims to solve the problem that the psychological state cannot be objectively quantized and closed-loop regulated in real time due to the fact that single parameters are relied on, evaluation is delayed, and monitoring and intervention are performed on fracture in the prior art. The invention provides a wearable vital sign monitoring system with multiple parameters synchronously acquired, which comprises: the multi-mode physiological sensing array is used for parallelly collecting multi-dimensional physiological original signals of a user with millisecond time synchronization precision; The edge calculation and fusion module is used for preprocessing and extracting characteristics of the multidimensional physiological original signals acquired by the multi-mode physiological sensing array, and performing space-time alignment and characteristic level fusion of multi-source heterogeneous data so as to generate fusion characteristic vectors; the psychological state quantization engine is used for receiving the fusion characteristic vector generated by the edge calculation and fusion module, calculating through a preset psychological state quantization model and outputting a quantization index representing the current psychological state of the user; And the closed-loop nerve regulation module is used for receiving the psychological state quantization index output by the psychological state quantization engine, and automatically generating and executing non-invasive nerve regulation stimulation matched with the deviation direction and amplitude of the psychological state quantization index when the psychological state quantization index exceeds a preset steady-state interval so as to guide the physiological index to return to the steady-state interval. Preferably, the multi-modal physiological sensing array integrates at least 4 types of physiological sensors, including a photoplethysmography sensor, a galvanic skin activity sensor, a tri-axial accelerometer and gyroscope combination sensor, and an infrared thermal radiation sensor; the photoelectric volume pulse wave sensor takes a light emitting diode as a light source, and a photoelectric detector receiv