Wrist Signal Acquisition and Processing Method and System Based on Biological Tissue Analysis

Abstract

The present invention relates to the technical field of medical monitoring, and particularly to a wrist signal acquisition and processing method and system based on biological tissue analysis. The system is used for: designing a wrist-worn device integrated with a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor, and optimizing the position and structure design to generate an optimized device for wrist multi-signal acquisition; using the optimized device for wrist multi-signal acquisition to collect PPG signals, ECG signals, and GSR signals corresponding to wrist biological tissues in real time, and performing high-frequency drift preprocessing and biological tissue feature analysis to obtain a wrist biological tissue signal feature set; performing feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set, and performing wrist physiological state evaluation and analysis to obtain a wrist biological tissue state evaluation result. The present invention can achieve accurate acquisition and analysis of wrist signals, thereby providing reliable data support for personalized analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of medical monitoring, and particularly to a wrist signal acquisition and processing method and system based on biological tissue analysis. Background Art

[0002] Wrist signal acquisition technology has shown great potential in the field of health monitoring due to its convenience and non-invasiveness. Wrist biological tissues contain rich physiological information. For example, PPG signals can reflect heart rate and blood oxygen saturation, ECG signals can reflect cardiac electrical activity, GSR signals can reflect skin electrical responses, etc. However, during the wrist signal acquisition process, it is vulnerable to various interferences, such as motion artifacts, environmental noise, and signal crosstalk, which lead to a decline in signal quality and seriously affect the accuracy of subsequent analysis. However, traditional wrist signal processing systems mostly focus on the filtering and noise reduction of single signals and lack consideration of the collaborative processing of multi-source signals, making it difficult to effectively solve the problems of signal interference and drift. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide a wrist signal acquisition and processing method and system based on biological tissue analysis to solve at least one of the above technical problems.

[0004] To achieve the above object, a wrist signal acquisition and processing system based on biological tissue analysis includes the following modules:

[0005] A wrist device sensing optimization design module, which is used to design a wrist wearable device integrating a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist wearable device based on the distribution characteristics of wrist biological tissues to generate an optimized wrist multi-signal acquisition device;

[0006] A wrist multi-signal preprocessing module, which is used to use the optimized wrist multi-signal acquisition device to collect PPG signals, ECG signals, and GSR signals corresponding to wrist biological tissues in real time to generate PPG signals of wrist biological tissues, ECG signals of wrist biological tissues, and GSR signals of wrist biological tissues; perform high-frequency drift preprocessing on the PPG signals of wrist biological tissues, ECG signals of wrist biological tissues, and GSR signals of wrist biological tissues to generate clear PPG signals, clear ECG signals, and clear GSR signals;

[0007] A biological tissue feature analysis module, which is used to perform biological tissue feature analysis on the clear PPG signals, clear ECG signals, and clear GSR signals to obtain a wrist biological tissue signal feature set;

[0008] A wrist biometric status evaluation module is used to perform feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set to generate a wrist biological tissue signal feature vector; and use a convolutional neural network to perform wrist physiological status evaluation and analysis on the wrist biological tissue signal feature vector to obtain a wrist biological tissue status evaluation result.

[0009] Furthermore, the wrist device sensing optimization design module includes the following functions:

[0010] Design a wrist wearable device integrating a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor;

[0011] Obtain the distribution characteristics of wrist biological tissues, including the distribution position of the radial artery in the wrist, the distribution position of the wrist muscles, and the distribution position of the wrist nerves;

[0012] Based on the distribution position of the radial artery in the wrist, perform PPG sensing optimization design on the photoplethysmogram sensor in the wrist wearable device to accurately align the photoplethysmogram sensor with the distribution position of the radial artery in the wrist, and ensure close contact between the sensor and the wrist skin through a flexible fitting technology to generate an optimized design scheme for the pulse wave sensing position;

[0013] Based on the distribution position of the wrist muscles, perform ECG electrode layout optimization on the electrocardiogram sensor in the wrist wearable device to generate an optimized design scheme for the ECG sensing electrode layout; based on the distribution position of the wrist nerves, perform skin interference optimization design on the galvanic skin response sensor in the wrist wearable device to generate an optimized design scheme for the galvanic skin response sensing interference;

[0014] Based on the optimized design scheme for the pulse wave sensing position, the optimized design scheme for the ECG sensing electrode layout, and the optimized design scheme for the galvanic skin response sensing interference, perform position and layout structure optimization design on the corresponding positions and layout structures of the photoplethysmogram sensor, electrocardiogram sensor, and galvanic skin response sensor in the wrist wearable device to generate an optimized device for wrist multi-signal acquisition.

[0015] Furthermore, the ECG electrode layout optimization for the electrocardiogram sensor in the wrist wearable device based on the distribution position of the wrist muscles includes:

[0016] Based on the distribution position of the wrist muscles, use the muscle electrical interference impact calculation formula to calculate the electrical interference impact on the electrode layout corresponding to the electrocardiogram sensor in the wrist wearable device to obtain the ECG electrode layout electrical interference impact factor;

[0017] Based on the ECG electrode layout electrical interference impact factor, perform ECG electrode layout optimization on the electrocardiogram sensor in the wrist wearable device to generate an optimized design scheme for the ECG sensing electrode layout.

[0018] Further, the calculation formula for the influence of muscle electrical interference is specifically as follows:

[0019] ;

[0020] In the formula, is the influence factor of electrical interference in the ECG electrode layout, is the spatial area where the wrist-worn device is located, is the layout position of the electrocardiogram sensing electrode, is the number of position points of muscle electrical activity distribution, is the th muscle activity point at the position where the muscle electrical activity intensity, is the maximum value of muscle electrical activity intensity, is the th influence weight of the muscle activity point on electrical interference, is the exponential function, is the th layout position corresponding to the muscle activity point, is the th standard deviation of the layout position corresponding to the muscle activity point, is the maximum frequency of muscle electrical activity, is the interference adjustment coefficient of muscle electrical activity, is the correction coefficient of the influence factor of electrical interference in the ECG electrode layout.

[0021] Further, the wrist multi-signal preprocessing module includes the following functions:

[0022] Using the photoplethysmogram sensor in the wrist multi-signal acquisition optimization device to collect the PPG signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue PPG signal;

[0023] Using the electrocardiogram sensor in the wrist multi-signal acquisition optimization device to collect the ECG signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue ECG signal;

[0024] Using the galvanic skin response sensor in the wrist multi-signal acquisition optimization device to collect the GSR signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue GSR signal;

[0025] Performing high-frequency drift preprocessing on the wrist biological tissue PPG signal, the wrist biological tissue ECG signal, and the wrist biological tissue GSR signal to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal.

[0026] Further, the high-frequency drift preprocessing of the wrist biological tissue PPG signal, the wrist biological tissue ECG signal, and the wrist biological tissue GSR signal includes:

[0027] Perform timestamp alignment and synchronization processing on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal, compensate for the time deviation corresponding to the signal based on the interpolation algorithm, and ensure that the wrist biological tissue PPG signal, ECG signal, and GSR signal are completely aligned on the time axis to obtain a set of multi-source synchronized and aligned signals of the wrist biological tissue;

[0028] Perform wavelet transform decomposition on the set of multi-source synchronized and aligned signals of the wrist biological tissue to separate the high-frequency and low-frequency components in the wrist biological tissue signal, and obtain the multi-source high-frequency signal of the wrist biological tissue and the multi-source low-frequency signal of the wrist biological tissue;

[0029] Perform high-frequency drift dynamic compensation on the multi-source high-frequency signal of the wrist biological tissue to obtain the high-frequency drift compensation signal of the wrist biological tissue;

[0030] Perform low-frequency noise elimination processing on the multi-source low-frequency signal of the wrist biological tissue to further eliminate the residual noise in the corresponding low-frequency component of the wrist biological tissue signal, and obtain the low-frequency noise removal signal of the wrist biological tissue;

[0031] Perform multi-modal signal fusion on the high-frequency drift compensation signal of the wrist biological tissue and the low-frequency noise removal signal of the wrist biological tissue to generate a multi-modal clear signal map of the wrist biological tissue; Based on the multi-modal clear signal map of the wrist biological tissue and combined with the signal reconstruction method, perform clear signal reconstruction output on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to generate a PPG clear signal, an ECG clear signal, and a GSR clear signal.

[0032] Further, the high-frequency drift dynamic compensation for the multi-source high-frequency signal of the wrist biological tissue includes:

[0033] Perform high-frequency drift component analysis on the multi-source high-frequency signal of the wrist biological tissue to generate a high-frequency drift component map of the wrist biological tissue;

[0034] Perform high-frequency baseline drift fitting processing on the high-frequency drift component map of the wrist biological tissue to obtain a high-frequency baseline drift signal tensor of the wrist biological tissue;

[0035] Perform high-frequency drift dynamic compensation on the multi-source high-frequency signal of the wrist biological tissue based on the high-frequency baseline drift signal tensor of the wrist biological tissue to obtain the high-frequency drift compensation signal of the wrist biological tissue.

[0036] Further, the biological tissue feature analysis module includes the following functions:

[0037] Obtain the corresponding pulse wave rise time and pulse wave fall time from the clear PPG signal, and analyze the peak amplitude and dicrotic wave characteristics of the clear PPG signal based on the pulse wave rise time and pulse wave fall time to obtain the PPG signal fluctuation characteristics of the pulse wave, including the fluctuation amplitude and fluctuation phase corresponding to the pulse wave peak and dicrotic wave;

[0038] Analyze the electrocardiogram signal characteristics of the corresponding P wave group, QRS wave group, and T wave group in the clear ECG signal to obtain the ECG signal fluctuation characteristics of the electrocardiogram, including the amplitude, width, and morphology corresponding to the P wave, QRS wave, and T wave;

[0039] Analyze the skin conductance characteristics of the clear GSR signal to obtain the GSR signal fluctuation characteristics of the skin conductance, including the skin conductance change rate and the average skin conductance fluctuation;

[0040] Merge the PPG signal fluctuation characteristics of the pulse wave, the ECG signal fluctuation characteristics of the electrocardiogram, and the GSR signal fluctuation characteristics of the skin conductance to obtain the wrist biological tissue signal feature set.

[0041] Furthermore, the wrist biometric state evaluation module includes the following functions:

[0042] Remove feature redundancy from the wrist biological tissue signal feature set to obtain the wrist biological tissue signal redundancy-removed feature set;

[0043] Perform principal component feature dimensionality reduction processing on the wrist biological tissue signal redundancy-removed feature set and combine them to form a comprehensive feature vector to generate the wrist biological tissue signal feature vector;

[0044] Analyze the topological distribution of the wrist biological tissue signal feature vector in the feature space to obtain the distribution density of the wrist biological tissue signal feature space;

[0045] Based on the distribution density of the wrist biological tissue signal feature space and combined with a convolutional neural network, design an adaptive convolutional kernel, so as to use a 5x5 convolutional kernel to capture the global features of the wrist biological tissue signal in the signal feature region corresponding to dense features and gentle changes, and use a 1x1 convolutional kernel to capture the local detail features of the wrist biological tissue signal in the signal feature region corresponding to sparse features and drastic changes, so as to design and generate a wrist feature-enhanced adaptive convolutional kernel model;

[0046] The wrist feature-enhanced adaptive convolution kernel model is used to perform multi-scale convolution operations on the wrist biological tissue signal feature vectors to generate wrist multi-scale convolution enhanced feature vectors; based on the wrist multi-scale convolution enhanced feature vectors and combined with the fuzzy logic inference method, wrist state evaluation is carried out to obtain the wrist physiological tissue state evaluation index; based on the wrist physiological tissue state evaluation index and combined with the multi-objective optimization algorithm, physiological evaluation calibration is carried out to obtain the wrist biological tissue state evaluation result.

[0047] Furthermore, the present invention also provides a wrist signal acquisition and processing method based on biological tissue analysis for implementing the wrist signal acquisition and processing system based on biological tissue analysis as described above. The wrist signal acquisition and processing method based on biological tissue analysis includes the following steps:

[0048] Step S1: Design a wrist wearable device corresponding to an integrated photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist wearable device based on the distribution characteristics of wrist biological tissues to generate an optimized multi-signal acquisition device for the wrist.

[0049] Step S2: Use the optimized multi-signal acquisition device for the wrist to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissues in real time to generate the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal; perform high-frequency drift preprocessing on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to generate clear PPG signals, clear ECG signals, and clear GSR signals.

[0050] Step S3: Perform biological tissue feature analysis on the clear PPG signals, clear ECG signals, and clear GSR signals to obtain a wrist biological tissue signal feature set.

[0051] Step S4: Perform feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set to generate wrist biological tissue signal feature vectors; use a convolutional neural network to perform wrist physiological state evaluation analysis on the wrist biological tissue signal feature vectors to obtain the wrist biological tissue state evaluation result.

[0052] The beneficial effects of the present invention:

[0053] 1. The wrist signal acquisition and processing system based on biological tissue analysis proposed by the present invention, compared with the prior art, the beneficial effects of the present application are as follows: by combining a photoplethysmogram (PPG) sensor, an electrocardiogram (ECG) sensor, and a galvanic skin response (GSR) sensor, an integrated wrist wearable device is designed. Different sensors obtain relevant data of wrist biological tissues from different physiological signal perspectives. The PPG signal reflects the changes in blood flow and can be used to monitor cardiovascular health; the ECG signal reflects the electrical activity of the heart and helps in the assessment of heart rate and rhythm; the GSR signal provides information on skin reactions and can be used to evaluate an individual's emotional response or stress level. The design of the wrist wearable device must be optimized based on the distribution characteristics of wrist biological tissues to ensure that the sensors can accurately capture various signals. Through the study of the structures of different levels of wrist biological tissues (such as skin, fat, muscle, etc.), designers can select appropriate positions and structures to optimize the acquisition effect of the sensors and maximize the intensity and accuracy of the signals. Through the integration and optimization of these sensors, the device is equipped with efficient physiological data acquisition capabilities. Secondly, by using the wrist multi-signal acquisition and optimization device to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissue in real time, the main purpose of this step is to obtain multiple biological signals of the human wrist in real time through the wrist multi-signal acquisition device, and capture the PPG signal, ECG signal, and GSR signal in real time through high-precision sensors. During the acquisition process, due to the complexity of the signals, there will be certain high-frequency drift noises (such as motion artifacts, environmental light interference, muscle contractions, etc.). Therefore, by performing high-frequency drift preprocessing on these signals, unnecessary noises and artifacts can be removed, thereby enhancing the clarity of the signals and making subsequent analysis more accurate and stable. The preprocessed signals are closer to the real physiological responses of the human body and can provide more accurate data support for the health status assessment of biological tissues. The processed signals can not only effectively reduce interference but also improve the quality of the data, making the details of the signals clearer and more reliable, and can also fully consider the collaborative processing of multi-source signals, thus effectively solving the problems of signal interference and high-frequency drift. Then, by performing biological tissue feature analysis on the clear PPG signal, clear ECG signal, and clear GSR signal, through in-depth feature extraction of these signals, physiological characteristics in aspects such as cardiovascular health and mental state can be obtained, thereby providing a basis for subsequent physiological state assessment. Finally, by performing feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set, representative features are extracted to improve the efficiency and accuracy of data processing. Biological signals usually have high dimensionality and complexity, which requires dimensionality reduction processing of the feature set for more efficient analysis and modeling. During the dimensionality reduction process, those features that contribute less or are redundant to the evaluation results are removed, and only the most distinguishable features are retained, thereby improving the effect of the analysis model.After dimensionality reduction, deep learning models such as convolutional neural networks (CNNs) are used to analyze the biological signal feature vectors. These deep learning models can automatically extract useful patterns from the data and identify features related to physiological states such as health status, mood changes, and stress levels. The CNN model can effectively process complex signal data, identify deep features in the signals, and perform state evaluation based on this. The state evaluation result of the wrist biological tissue can provide real-time health feedback for the wearer, such as heart health, emotional state, stress level, etc., helping users understand their physical conditions in a timely manner and take corresponding measures for adjustment, thus providing strong support for personalized health management.

[0054] 2. The wrist signal acquisition and processing method based on biological tissue analysis proposed by the present invention, implementing any wrist signal acquisition and processing system based on biological tissue analysis of the present invention, is used to execute operations between computer programs running on each module to achieve biological tissue analysis-based, enabling the internal structure of the system to cooperate with each other and intelligently execute each process of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, objects, and advantages of the present invention will become more apparent:

[0056] Figure 1 It is a schematic diagram of the modules of the wrist signal acquisition and processing system based on biological tissue analysis of the present invention;

[0057] Figure 2 is Figure 1 a schematic diagram of the functional flow of the wrist device sensing optimization design module in DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] The following clearly and completely describes the technical method of the present invention with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work fall within the scope of protection of the present invention.

[0059] In addition, the drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0060] It should be understood that although terms such as "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0061] To achieve the above object, please refer to Figures 1 to 2 , the present invention provides a wrist signal acquisition and processing system based on biological tissue analysis, and the system includes the following modules:

[0062] Wrist device sensing optimization design module, which is used to design a wrist wearable device integrating a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist wearable device based on the distribution characteristics of wrist biological tissues to generate an optimized device for collecting multiple wrist signals;

[0063] Wrist multi-signal preprocessing module, which is used to use the optimized device for collecting multiple wrist signals to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissue in real time to generate the PPG signal of the wrist biological tissue, the ECG signal of the wrist biological tissue, and the GSR signal of the wrist biological tissue; perform high-frequency drift preprocessing on the PPG signal of the wrist biological tissue, the ECG signal of the wrist biological tissue, and the GSR signal of the wrist biological tissue to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal;

[0064] Biological tissue feature analysis module, which is used to perform biological tissue feature analysis on the clear PPG signal, clear ECG signal, and clear GSR signal to obtain a wrist biological tissue signal feature set;

[0065] Wrist biometric status evaluation module, which is used to perform feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set to generate a wrist biological tissue signal feature vector; use a convolutional neural network to perform wrist physiological state evaluation analysis on the wrist biological tissue signal feature vector to obtain a wrist biological tissue state evaluation result.

[0066] In an embodiment of the present invention, please refer to Figure 1 As shown, it is a schematic diagram of the modules of the wrist signal acquisition and processing system based on biological tissue analysis of the present invention. In this example, the wrist signal acquisition and processing system based on biological tissue analysis includes the following modules:

[0067] S1: Wrist device sensing optimization design module, which is used to design a wrist-worn device integrated with a photoplethysmogram (PPG) sensor, an electrocardiogram (ECG) sensor, and a galvanic skin response (GSR) sensor; obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist-worn device based on the distribution characteristics of wrist biological tissues to generate an optimized device for multi-signal acquisition on the wrist.

[0068] In the embodiment of the present invention, by designing a wrist-worn device composed of an integrated photoplethysmogram (PPG) sensor, an electrocardiogram (ECG) sensor, and a galvanic skin response (GSR) sensor, it is first necessary to clarify the distribution characteristics of wrist biological tissues. By modeling the anatomical analysis of the wrist and the thickness of biological tissues, the influence of wrist tissues on different sensor signals is determined. For example, a three-dimensional imaging of blood vessels, muscle layers, etc. on and under the skin surface is performed using an optical scanner to analyze its absorption and reflection characteristics of the PPG signal, and further investigate the interference of the ECG signal by the wrist fat layer and muscles. According to these analyses, the sensor position and structural layout are optimized to ensure the maximum acquisition efficiency of PPG, ECG, and GSR signals. During the design process, the comfort and wearing stability of the wrist device should also be considered to ensure that the sensor can be accurately fixed on the target biological tissue layer, and finally an optimized device for multi-signal acquisition on the wrist is generated.

[0069] S2: Wrist multi-signal preprocessing module, which is used to use the optimized device for multi-signal acquisition on the wrist to real-time collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissue to generate the PPG signal of the wrist biological tissue, the ECG signal of the wrist biological tissue, and the GSR signal of the wrist biological tissue; perform high-frequency drift preprocessing on the PPG signal of the wrist biological tissue, the ECG signal of the wrist biological tissue, and the GSR signal of the wrist biological tissue to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal.

[0070] In an embodiment of the present invention, a wrist multi-signal acquisition and optimization device is used to collect corresponding PPG signals, ECG signals, and GSR signals in biological tissues in real time. The PPG signal is collected by bringing the integrated optoelectronic sensor into contact with the skin surface, and the reflected light of blood flow under the skin is measured using a light source and a photodiode to obtain a pulse waveform. The ECG signal is collected by detecting the weak electric field changes generated by the electrical activity of the heart through an electrode sensor attached to the skin surface. The GSR signal reflects the emotional changes or sweat gland activities through the conductance changes on the skin surface. The collected signals are usually interfered by noise, especially high-frequency drift noise. Therefore, in this step, high-frequency drift preprocessing is performed on each signal (PPG signal, ECG signal, GSR signal) to remove the high-frequency drift in the PPG signal by using a band-pass filter to extract a clear pulse wave signal. For the ECG signal, an adaptive filtering technique is used to remove the high-frequency components introduced by poor electrode contact or environmental noise to ensure a clear electrocardiogram signal. The GSR signal is amplified for the low-frequency part by a differential amplifier and the power supply interference is removed through filtering to obtain a clean GSR signal, and finally, a clear PPG signal, a clear ECG signal, and a clear GSR signal are generated.

[0071] S3: A biological tissue feature analysis module, which is used to perform biological tissue feature analysis on the clear PPG signal, the clear ECG signal, and the clear GSR signal to obtain a wrist biological tissue signal feature set;

[0072] In an embodiment of the present invention, after high-frequency drift preprocessing of the PPG, ECG, and GSR signals, biological tissue feature analysis is performed. In the analysis of the PPG signal, features reflecting blood circulation and cardiovascular health, such as heart rate, pulse wave amplitude, etc., are extracted. A combination of time-domain analysis and frequency-domain analysis is used to calculate the average value, standard deviation, and spectral features of the signal. During the extraction of ECG signal features, attention is paid to the time and frequency features of key electrocardiogram waveforms such as P waves, QRS complexes, and T waves, and physiological parameters such as heart rate variability and QT interval are extracted to further analyze the heart health status. For the GSR signal, the change law of skin conductance is extracted, and the emotional response or stress state of an individual is inferred by analyzing the change trend of sweat gland activities. Through the extraction of these features, a feature set of wrist biological tissue signals is constructed, covering multi-dimensional physiological information such as the cardiovascular system and emotional state, and finally, a wrist biological tissue signal feature set is obtained.

[0073] S4: A wrist biometric status evaluation module, which is used to perform feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set to generate a wrist biological tissue signal feature vector; and use a convolutional neural network to perform wrist physiological state evaluation analysis on the wrist biological tissue signal feature vector to obtain a wrist biological tissue state evaluation result.

[0074] In an embodiment of the present invention, through feature redundancy and dimensionality reduction processing on the feature sets of the extracted clear PPG signals, clear ECG signals, and clear GSR signals, in the process of feature redundancy processing, the principal component analysis (PCA) method is adopted to remove redundant information in the signals to retain effective features that can most reflect the physiological state to the greatest extent. For feature dimensionality reduction, independent component analysis (ICA) or linear discriminant analysis (LDA) is used to further reduce the dimensions, eliminate noise and interference, ensure that the extracted features are more concise and effective, and the generated wrist biological tissue signal feature vectors are then input into a convolutional neural network (CNN) for deep learning processing. Through the multi-layer convolution and pooling operations of the CNN, combined with the trained model, the feature vectors are evaluated and analyzed for the physiological state of the wrist. The convolutional neural network can automatically learn key physiological state patterns from the input signal features and output comprehensive wrist biological tissue state evaluation results such as heart health status, stress level, and emotional state. This evaluation result provides a basis for further health monitoring and physiological state early warning, can reflect the physiological condition of the wearer in real time, and finally obtains the wrist biological tissue state evaluation result.

[0075] Furthermore, the wrist device sensing optimization design module includes the following functions:

[0076] Design an integrated wrist-worn device corresponding to a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor;

[0077] Obtain the distribution characteristics of wrist biological tissues, including the distribution positions of the radial artery in the wrist, the distribution positions of the wrist muscles, and the distribution positions of the wrist nerves;

[0078] Based on the distribution position of the radial artery in the wrist, conduct PPG sensing optimization design on the photoplethysmogram sensor in the wrist-worn device to accurately align the photoplethysmogram sensor with the distribution position of the radial artery in the wrist, and ensure tight contact between the sensor and the wrist skin through a flexible fitting technology to generate an optimized design scheme for the pulse wave sensing position;

[0079] Based on the distribution position of the wrist muscles, conduct ECG electrode layout optimization on the electrocardiogram sensor in the wrist-worn device to generate an optimized design scheme for the ECG sensing electrode layout; based on the distribution position of the wrist nerves, conduct skin interference optimization design on the galvanic skin response sensor in the wrist-worn device to generate an optimized design scheme for the galvanic skin response sensing interference;

[0080] Based on the optimized design scheme of the pulse wave sensing position, the optimized design scheme of the electrocardiogram sensing electrode layout, and the optimized design scheme of the galvanic skin response sensing interference, the position and layout structure of the photoplethysmogram sensor, electrocardiogram sensor, and galvanic skin response sensor corresponding to the wrist wearable device are optimized to generate an optimized device for multi-signal acquisition on the wrist.

[0081] As an embodiment of the present invention, referring to Figure 2 shown, it is Figure 1 the functional flow diagram of the wrist device sensing optimization design module in

[0082] S11: Design an integrated wrist wearable device corresponding to the photoplethysmogram sensor, electrocardiogram sensor, and galvanic skin response sensor;

[0083] In the embodiment of the present invention, when designing an integrated wrist wearable device with a photoplethysmogram sensor (PPG sensor), an electrocardiogram sensor (ECG sensor), and a galvanic skin response sensor (GSR sensor), first determine the performance requirements of each sensor. The photoplethysmogram sensor needs to have high sensitivity to accurately capture the pulse waveform signal. The electrocardiogram sensor requires high electrode conductivity to obtain a clear electrocardiogram signal. The galvanic skin response sensor needs to have a strong response ability to skin electrical signals. For the performance requirements of each sensor, high-performance optical components, conductive electrode materials, and high-precision current detection modules are used to ensure the stable operation of the sensors. When designing, the fit between the wrist skin and the sensors also needs to be considered. Flexible materials are combined with the sensor components, and it is ensured that each sensor has low power consumption, small volume, and is suitable for wearing on the wrist, improving the comfort and usage effect of the wearable device. Finally, a wrist wearable device is designed and generated.

[0084] S12: Obtain the distribution characteristics of wrist biological tissues, including the distribution positions of the radial artery, muscles, and nerves in the wrist;

[0085] In the embodiments of the present invention, by using medical imaging technologies such as ultrasonic imaging, MRI imaging, or CT scanning, the organizational structure information of the wrist is obtained in detail. Through image processing software, the imaging data is analyzed to accurately calibrate the specific positions and distribution characteristics of the radial artery, muscles, and nerves in the wrist. The radial artery in the wrist is usually located on the palmar side of the wrist near the wrist joint and is relatively concentrated in distribution; the wrist muscles are mainly distributed in the forearm area, especially near the radius; the wrist nerves are relatively thin and are usually located in the area between the radial artery and the muscles. According to this biological tissue distribution information, the next-step sensor optimization design plan is formulated, and the layout and positioning of the sensors can be carried out more accurately, and finally the distribution characteristics of the wrist biological tissues are obtained, including the distribution position of the radial artery in the wrist, the distribution position of the wrist muscles, and the distribution position of the wrist nerves.

[0086] S13: Based on the distribution position of the radial artery in the wrist, perform PPG sensing optimization design on the photoplethysmogram sensor in the wrist wearable device to accurately align the photoplethysmogram sensor with the distribution position of the radial artery in the wrist, and ensure the close contact between the sensor and the wrist skin through a flexible fitting technology to generate an optimized design plan for the pulse wave sensing position.

[0087] In the embodiments of the present invention, when optimizing the design of the photoplethysmogram sensor based on the distribution position of the radial artery in the wrist, first use simulation software to analyze the relationship between the position of the radial artery in the wrist and blood flow, and accurately design the incident angle and detection angle of the photoplethysmogram sensor using optical principles to ensure that the sensor can accurately align with the surface area of the radial artery. Then, design an adaptable and comfortable fitting structure through a flexible fitting technology to ensure the close contact between the photoplethysmogram sensor and the wrist skin, so as to minimize the gap between the sensor and the skin to the greatest extent, thereby ensuring the accurate acquisition of the pulse wave signal. This design plan also needs to consider individual differences such as the wrist circumference and skin type of different users to ensure that the device has a certain adjustment space when worn to meet the needs of different users, and finally generate an optimized design plan for the pulse wave sensing position.

[0088] S14: Based on the distribution position of the wrist muscles, perform ECG electrode layout optimization on the electrocardiogram sensor in the wrist wearable device to generate an optimized design plan for the ECG sensing electrode layout; based on the distribution position of the wrist nerves, perform skin interference optimization design on the galvanic skin response sensor in the wrist wearable device to generate an optimized design plan for the galvanic skin response sensing interference.

[0089] In an embodiment of the present invention, when optimizing the design of the electrocardiogram (ECG) sensor based on the muscle distribution position on the wrist, first, through the muscle distribution image data, the relationship between the main positions of the muscles in the wrist area and the ECG signals is analyzed. The electrode layout of the ECG sensor needs to ensure that it can capture the electrophysiological activity signals in the wrist area. An electrode array design is adopted, and by accurately measuring the distance and angle between the electrodes, the accurate transmission of the electrocardiogram signals is ensured, thereby designing and generating an optimized design scheme for the ECG sensing electrode layout. For the interference optimization design of the galvanic skin response (GSR) sensor, considering the relationship between the position of the nerves in the wrist and the GSR signals, appropriate interference shielding technologies are used, and the sensor is designed as a multi-layer structure to reduce the influence of external interference on the signals. In addition, the sensitivity of the GSR sensor is optimized to enable it to capture the GSR signals of the skin more accurately, ensuring the effectiveness and accuracy of the data, and finally generating an optimized design scheme for the GSR sensing interference.

[0090] S15: Based on the optimized design scheme for the pulse wave sensing position, the optimized design scheme for the ECG sensing electrode layout, and the optimized design scheme for the GSR sensing interference, the position and layout structure of the corresponding positions of the photoplethysmogram (PPG) sensor, the ECG sensor, and the GSR sensor in the wrist wearable device are optimized in terms of position and structure to generate an optimized device for multi-signal acquisition in the wrist.

[0091] In an embodiment of the present invention, based on the above optimization scheme, when specifically designing the layout of the PPG sensor, the ECG sensor, and the GSR sensor in the wrist wearable device, first, according to the distribution positions of the radial artery, muscles, and nerves, the optimal position of each sensor is accurately determined. The layout of the sensors adopts a symmetric design to ensure the accuracy of signal acquisition during wearing. Specifically, the PPG sensor is designed to fit the surface of the wrist and is accurately aligned with the position of the radial artery. The electrodes of the ECG sensor are arranged in an array according to the muscle distribution on the wrist to ensure the optimal contact area. The layout of the GSR sensor is designed to minimize interference according to the nerve distribution. The overall structure design needs to consider the portability, comfort, and stability of the device, and flexible materials and lightweight components are used to ensure the comfort of the wearing experience. Through structural optimization, an integrated, multi-signal acquisition, accurate, and comfortable wrist wearable device is formed, and finally an optimized device for multi-signal acquisition in the wrist is generated.

[0092] Furthermore, the optimization of the ECG electrode layout for the ECG sensor in the wrist wearable device based on the muscle distribution position on the wrist includes:

[0093] Based on the muscle distribution position on the wrist, the electro interference influence of the corresponding electrode layout of the ECG sensor in the wrist wearable device is calculated using the muscle electro interference influence calculation formula to obtain the electro interference influence factor of the ECG electrode layout;

[0094] In the embodiment of the present invention, a suitable calculation formula for the influence of muscle electrical interference is constructed by combining the spatial area where the wrist-worn device is located, the layout positions of the electrocardiogram (ECG) sensing electrodes, the number of muscle electrical activity distribution position points, the muscle electrical activity intensity, the maximum value of the muscle electrical activity intensity, the influence weight of the muscle activity points on electrical interference, the layout positions corresponding to the muscle activity points, the standard deviation of the layout positions corresponding to the muscle activity points, the maximum frequency of the muscle electrical activity, the muscle electrical interference adjustment coefficient, and related parameters for calculation, so as to further accurately adjust the electrode spacing and electrode positions, and then obtain the electrical interference influence factors under each electrode layout, and finally obtain the electrical interference influence factors of the ECG electrode layout.

[0095] Preferably, based on the electrical interference influence factors of the ECG electrode layout, the ECG electrodes in the wrist-worn device are optimized for the ECG electrode layout to generate an optimized design scheme for the layout of the ECG sensing electrodes.

[0096] In the embodiment of the present invention, by combining the previously obtained electrical interference influence factors of the ECG electrode layout, the ECG electrodes in the wrist-worn device are further optimized for the ECG electrode layout to generate an optimized electrode layout design scheme. During the implementation process, first, the previously calculated electrical interference influence factors need to be used as input parameters and considered as key indicators in the optimization algorithm. Using computer-aided design (CAD) tools and finite element analysis (FEA) models, the electrode positions are optimized. This optimization process relies on the dynamic simulation of electrical interference and combines the specific anatomical features of the human wrist area (such as bone structure, muscle distribution, skin resistance, etc.) to adjust the electrode positions. During the optimization process, physical factors such as the geometric shape of the electrode arrangement and the contact resistance between the electrode and the skin need to be considered. Through the iterative adjustment algorithm, the electrical interference influence factors are gradually reduced, and the electrode layout is optimized to improve the acquisition accuracy of the ECG signal. Specifically, optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.) are used to automatically adjust the positions, spacings, and angles of the electrodes to ensure the optimal acquisition effect of the ECG signal and to be unaffected by the wrist muscle electrical interference. At the same time, the optimization process should also ensure the comfort and operability of the wearable device, and finally generate an optimized design scheme for the layout of the ECG sensing electrodes.

[0097] Furthermore, the specific calculation formula for the influence of muscle electrical interference is:

[0098] ;

[0099] In the formula, is the electrical interference influence factor of the ECG electrode layout, is the spatial area where the wrist-worn device is located, is the layout position of the ECG sensing electrodes, is the number of muscle electrical activity distribution position points, is the The muscle activity point is at the position where the muscle electrical activity intensity is the maximum value of the muscle electrical activity intensity is the influence weight of the th muscle activity point on the electrical interference is an exponential function is the layout position corresponding to the th muscle activity point is the standard deviation of the layout position corresponding to the th muscle activity point is the maximum frequency of the muscle electrical activity is the interference adjustment coefficient of the muscle electrical activity is the correction coefficient of the influence factor of the electrical interference of the ECG electrode layout

[0100] The present invention obtains a calculation formula for the influence of electromyographic interference through the use of a specific mathematical model and verification, which is used to calculate the influence of electromyographic interference on the electrode layout corresponding to the electrocardiogram sensor in a wrist-worn device. By considering factors such as muscle activity distribution, interference intensity, and electrode layout position, this calculation formula for the influence of electromyographic interference can accurately calculate the influence of electromyographic interference on the electrocardiogram signal. Through the integral formula, combined with the influence weight and muscle activity intensity of each muscle activity point, it is possible to quantify the interference degree of different position points on the ECG signal, which can help identify which areas have a greater interference impact, thereby guiding the optimization of the electrode layout. This formula involves integral calculations in a spatial region, and it considers the relative relationship between the electrode layout position and the muscle activity points, which provides a basis for designers on how to select electrode positions. By reducing the influence of electromyographic activity interference, a better electrode layout can be designed to ensure the quality and accuracy of the ECG signal. Multiple parameters in this formula can be adjusted to reflect different environmental or individual differences. For example, the frequency and adjustment coefficient of electromyographic activity can be adjusted for different application scenarios to provide flexibility. By reasonably setting these parameters, the influence degree of electrical interference can be precisely controlled, making the optimization scheme more personalized and efficient. Since wrist-worn devices are often affected by electromyographic interference, especially the noise caused by wrist movement or muscle activity, optimizing the electrode layout through this formula can effectively reduce these interferences, thereby improving the quality of the ECG signal. By optimizing the electrode layout of the electrocardiogram sensor based on the electrical interference influence factor, the wrist-worn device can be made more adaptable to the physiological differences and exercise states of different users. For example, through an optimization algorithm, the device can adapt to differences in different body types, muscle activity intensities, frequencies, etc., providing a personalized signal acquisition scheme, further enhancing the adaptability and user experience of the device. In addition, the correction coefficient in this formula can be used for further adjustment and optimization, reflecting the deviations and errors in actual applications, ensuring an ideal evaluation of the influence of electrical interference in different application scenarios. This correction makes the calculation more accurate and avoids the error influence caused by environmental factors or individual differences. In summary, this formula fully considers the electrical interference influence factor of the ECG electrode layout , the spatial region where the wrist-worn device is located , the layout position of the electrocardiogram sensing electrodes , the number of position points of the muscle electrical activity distribution , the th muscle activity point is at the position with the muscle electrical activity intensity , the maximum value of the muscle electrical activity intensity , the th muscle activity point's influence weight on electrical interference , the exponential function , the The layout positions corresponding to muscle activity points , the standard deviation of the layout positions corresponding to the th muscle activity point, the maximum frequency of muscle electrical activity , the interference regulation coefficient of muscle electrical activity , the correction coefficient of the electrical interference influence factor of the ECG electrode layout , based on the electrical interference influence factor of the ECG electrode layout and the mutual correlation relationships among the above parameters constitute a functional relationship , this formula can realize the calculation process of the electrical interference influence of the electrode layout corresponding to the ECG sensor in the wrist-worn device. At the same time, through the correction coefficient of the electrical interference influence factor of the ECG electrode layout, it can be adjusted according to the error situation in the calculation process, so as to improve the accuracy and applicability of the calculation formula of the muscle electrical interference influence.

[0101] Furthermore, the wrist multi-signal preprocessing module includes the following functions:

[0102] Utilize the photoplethysmogram (PPG) sensor in the wrist multi-signal acquisition optimization device to collect the PPG signal corresponding to the wrist biological tissue in real time, so as to generate the wrist biological tissue PPG signal;

[0103] In the embodiment of the present invention, in the wrist multi-signal acquisition optimization device, the integrated photoplethysmogram (PPG) sensor is used to collect the real-time signal of the wrist biological tissue. The PPG sensor is composed of a light source and a photodetector. The light source emits light of a certain wavelength. After penetrating the wrist epidermal tissue, it is reflected back to the photodetector. The detector obtains the signal data related to blood flow according to the change in the intensity of the reflected light. By analyzing the reflected light signal on the surface of the wrist skin, the PPG signal of the wrist biological tissue is generated in real time. To ensure the signal quality, the light source wavelength of the PPG sensor should be selected as the light wavelength suitable for skin permeability and blood absorption characteristics, such as red light or near-infrared light. During the data acquisition process, the processing module in the device preliminarily filters the collected PPG signal to remove environmental light interference, and ensures the reliability of the signal through signal enhancement technology, and finally generates the wrist biological tissue PPG signal.

[0104] Preferably, utilize the ECG sensor in the wrist multi-signal acquisition optimization device to collect the ECG signal corresponding to the wrist biological tissue in real time, so as to generate the wrist biological tissue ECG signal;

[0105] In an embodiment of the present invention, within the wrist multi-signal acquisition optimization device, real-time ECG signal acquisition of wrist biological tissue is performed through an integrated electrocardiogram sensor. The electrocardiogram sensor is attached to the skin surface through a set of electrodes to detect the electrical activity signals generated by the heart. The electrodes sense the electrical wave conduction signals of the heart through the skin potential difference and convert them into digital signals. By analyzing the electrocardiogram signals, each electrical activity of the heart can be accurately captured, thereby generating the ECG signal of the wrist biological tissue, and finally generating the ECG signal of the wrist biological tissue.

[0106] Preferably, the galvanic skin response sensor within the wrist multi-signal acquisition optimization device is used to real-time collect the GSR signal corresponding to the wrist biological tissue to generate the GSR signal of the wrist biological tissue;

[0107] In an embodiment of the present invention, within the wrist multi-signal acquisition optimization device, the integrated galvanic skin response (GSR) sensor is used to real-time collect the GSR signal of the wrist biological tissue. The GSR signal reflects the physiological and psychological states by measuring the conductance change of the skin. This sensor usually consists of two electrodes. The electrodes contact the wrist skin and obtain the GSR signal by measuring the change of the skin resistance. The skin conductance change is related to the sweat gland activity, and the sweat secretion will cause the change of the skin conductance. During the signal acquisition process, the GSR sensor within the device real-time monitors the fluctuation of the skin conductance value, converts it into an electrical signal and outputs it as the GSR circuit signal, and finally generates the GSR signal of the wrist biological tissue.

[0108] Preferably, high-frequency drift preprocessing is performed on the PPG signal, ECG signal, and GSR signal of the wrist biological tissue to generate a clear PPG signal, clear ECG signal, and clear GSR signal.

[0109] In an embodiment of the present invention, when performing high-frequency drift preprocessing on the PPG signal, ECG signal, and GSR signal of the wrist biological tissue, frequency domain filtering technology is used to process these signals. First, the original signals are converted from the time domain to the frequency domain using the fast Fourier transform (FFT) to analyze the spectral distribution of each signal. In the PPG signal and ECG signal, there are high-frequency drift components generated due to power supply interference or other external factors, and these components will affect the accuracy of the signals. By setting the filter parameters, a low-pass filter is applied to remove the signal components with frequencies higher than a specific threshold, such as the interference signals above 50 Hz. For the GSR signal, a similar filtering method is also used to remove the noise caused by environmental temperature changes or skin friction. After the frequency domain filtering process, the high-frequency drift components of the PPG signal, ECG signal, and GSR signal are effectively removed, and a clearer and smoother clear PPG signal, clear ECG signal, and clear GSR signal are obtained, and finally the clear PPG signal, clear ECG signal, and clear GSR signal are generated.

[0110] Further, the high-frequency drift preprocessing of the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal includes:

[0111] Perform timestamp alignment and synchronization processing on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to compensate for the time deviation corresponding to the signal based on the interpolation algorithm, and ensure that the wrist biological tissue PPG signal, ECG signal, and GSR signal are completely aligned on the time axis to obtain a multi-source synchronous alignment signal set of wrist biological tissues;

[0112] In the embodiment of the present invention, by ensuring that the timestamps of the wrist biological tissue PPG (photoplethysmogram) signal, ECG (electrocardiogram) signal, and GSR (galvanic skin response) signal are completely aligned, this process is achieved through an interpolation algorithm. For the three signal sources, first extract their sampling time points and corresponding numerical data, and through interpolation methods such as linear interpolation or cubic interpolation, generate new data points between their timestamps to make the sampling times of the three signals exactly the same. The interpolation process should be calculated based on the smallest time unit to ensure precise synchronization of the signal time intervals, thereby eliminating the time deviation caused by inconsistent sampling frequencies of different signal sources. For the interpolated data, perform error analysis to ensure that the time deviation of all signals is less than a certain threshold to ensure the accuracy of synchronization. If the time deviation exceeds the predetermined threshold, interpolation correction needs to be performed again. Through these processes, ensure that the PPG signal, ECG signal, and GSR signal are precisely aligned on the same time axis, and finally obtain a multi-source synchronous alignment signal set of wrist biological tissues.

[0113] Preferably, perform wavelet transform decomposition on the multi-source synchronous alignment signal set of wrist biological tissues to separate the high-frequency and low-frequency components in the wrist biological tissue signal, and obtain a multi-source high-frequency signal and a multi-source low-frequency signal of wrist biological tissues;

[0114] In the embodiments of the present invention, the synchronized and aligned signal is decomposed by wavelet transform to separate the high-frequency and low-frequency components in the signal for subsequent high-frequency drift compensation and low-frequency noise removal. Wavelet transform uses a time-frequency analysis method based on wavelet functions and can obtain high resolution both in time and frequency. Daubechies wavelet or Coiflet wavelet is selected for signal decomposition. The discrete wavelet transform (DWT) is applied to decompose the multi-source synchronized and aligned signal set into sub-signals in different frequency ranges. By selecting appropriate wavelet functions and their scale factors, the signal is decomposed at multiple levels to obtain high-frequency (such as heartbeat signals, sudden changes in skin conductance responses, etc.) and low-frequency (such as heart rate fluctuations, slow waves, etc.) components. The decomposed signal can clearly distinguish high-frequency noise and low-frequency background components. During the wavelet transform process, attention is paid to dealing with the boundary effect of the signal. Symmetric extension or periodic extension can be used to avoid the influence of the edge effect on signal decomposition and ensure the accuracy of the decomposition result at each level. Finally, multi-source high-frequency signals of wrist biological tissues and multi-source low-frequency signals of wrist biological tissues are obtained.

[0115] Preferably, dynamic compensation for high-frequency drift is performed on the multi-source high-frequency signals of wrist biological tissues to obtain high-frequency drift compensation signals of wrist biological tissues;

[0116] In the embodiments of the present invention, by performing drift compensation on multi-source high-frequency signals, the changes in high-frequency components caused by device drift, noise, or other external factors in the signal are corrected. First, a time-domain or frequency-domain method is used to detect drift in high-frequency signals. A drift detection algorithm based on the difference method can be used to analyze the change trend of the signal through the differences between consecutive sampling points, identify whether there is drift, and by using a high-order filter or adaptive filtering technology, dynamically adjust the amplitude and frequency of high-frequency signals to eliminate the drift caused by device deviation or environmental factors. Typical methods include Kalman filtering or mean shift correction. These methods ensure the stability of the signal by dynamically estimating and correcting the trend changes in the signal. Finally, high-frequency drift compensation signals of wrist biological tissues are obtained.

[0117] Preferably, low-frequency noise elimination processing is performed on the multi-source low-frequency signals of wrist biological tissues to further eliminate the residual noise in the low-frequency components corresponding to the wrist biological tissue signals to obtain low-frequency noise removal signals of wrist biological tissues;

[0118] In the embodiment of the present invention, by using a low-pass filter to preliminarily suppress low-frequency noise, a low-pass filter of types such as Butterworth filter or Chebyshev filter can be used, and an appropriate cut-off frequency is set to suppress the noise components below this frequency. To further eliminate noise, a wavelet denoising method is adopted. By selecting an appropriate wavelet basis function and using threshold denoising technology in the decomposition process of low-frequency signals, the noise coefficients are filtered out, and only the main components in the signals are retained. During the noise removal process, different denoising strategies should be set according to the characteristics of low-frequency signals. For example, for signals with stronger slow-wave components, an adaptive threshold denoising method can be used to dynamically adjust the denoising intensity to avoid excessive elimination of useful signals. After the low-frequency noise elimination process, the noise components of the signal are significantly reduced, and more meaningful low-frequency information is retained, which can provide a clearer signal for subsequent analysis, and finally obtain the low-frequency noise removal signal of wrist biological tissue.

[0119] Preferably, multi-modal signal fusion is performed on the high-frequency drift compensation signal of wrist biological tissue and the low-frequency noise removal signal of wrist biological tissue to generate a multi-modal clear signal map of wrist biological tissue; based on the multi-modal clear signal map of wrist biological tissue and combined with the signal reconstruction method, clear signal reconstruction output is performed on the PPG signal of wrist biological tissue, the ECG signal of wrist biological tissue, and the GSR signal of wrist biological tissue to generate clear PPG signal, clear ECG signal, and clear GSR signal.

[0120] In the embodiment of the present invention, by fusing the high-frequency signal and the low-frequency signal after drift compensation and noise elimination processing to generate a clear multi-modal signal map, a weighted average method or a signal fusion technology based on the least squares method is used to combine the high-frequency drift compensation signal and the low-frequency noise removal signal. During the fusion process, the weights of each signal are optimized to ensure that the fused signal can not only retain the characteristics of high-frequency and low-frequency components but also reduce interference noise. When performing signal fusion, considering the timing relationship between the high-frequency signal and the low-frequency signal, a phase synchronization algorithm is adopted to make the two coordinate and cooperate both in the time domain and the frequency domain, avoiding introducing new inconsistencies during the fusion process, thereby generating a multi-modal clear signal map of wrist biological tissue. At the same time, by combining the previously generated multi-modal clear signal map of wrist biological tissue, the PPG signal, ECG signal, and GSR signal of wrist biological tissue are reconstructed, and signal reconstruction algorithms (such as inverse Fourier transform or inverse wavelet transform) are applied to reconstruct fine clear PPG, ECG, and GSR signals, ensuring that as little information loss as possible occurs during the signal recovery process, and finally generating clear PPG signal, clear ECG signal, and clear GSR signal.

[0121] Further, the high-frequency drift dynamic compensation for the multi-source high-frequency signals of wrist biological tissue includes:

[0122] Perform high-frequency drift component analysis on the multi-source high-frequency signals of wrist biological tissue to generate a high-frequency drift component map of wrist biological tissue;

[0123] In an embodiment of the present invention, a multi-source high-frequency signal of wrist biological tissue is collected by a high-frequency signal acquisition device. This device needs to have a multi-channel signal receiving function to be able to collect biological signals from different parts (such as the skin surface, deep tissues, etc.) simultaneously. The high-frequency signal collector can use a sensor array. By contacting the wrist skin surface, it captures multi-level bioelectric current signals within the wrist tissue. Then, through a high-frequency signal processing algorithm, spectral analysis is performed on the collected original signal to extract the high-frequency drift components. The high-frequency drift components refer to the change trend of the signal in the high-frequency range, which is related to the bioelectric characteristics of the tissue and external environment changes. Through frequency domain analysis tools such as Fourier transform, the frequency change of the analyzed signal is obtained, and a high-frequency drift component map of wrist biological tissue is generated. This map can present the distribution of different biological signal sources in terms of time and frequency, and finally generate a high-frequency drift component map of wrist biological tissue.

[0124] Preferably, perform high-frequency baseline drift fitting processing on the high-frequency drift component map of wrist biological tissue to obtain a high-frequency baseline drift signal tensor of wrist biological tissue;

[0125] In an embodiment of the present invention, after obtaining the high-frequency drift component map of wrist biological tissue, baseline drift fitting processing of the signal is entered. First, polynomial fitting, spline curve fitting, or other suitable mathematical models are used to model the baseline drift components in the map. Baseline drift refers to the long-term trend change existing in the signal, which is caused by factors such as environmental interference and equipment noise. By fitting the drift components, the baseline drift characteristics of the signal are obtained. The least squares method and other optimization algorithms are used in the fitting process to ensure the accuracy of the fitting result, and the baseline drift components in the signal are removed according to the fitting result, which can accurately represent the high-frequency drift signal in the wrist biological tissue, and finally obtain a high-frequency baseline drift signal tensor of wrist biological tissue.

[0126] Preferably, perform high-frequency drift dynamic compensation on the multi-source high-frequency signals of wrist biological tissue based on the high-frequency baseline drift signal tensor of wrist biological tissue to obtain a high-frequency drift compensation signal of wrist biological tissue.

[0127] In the embodiments of the present invention, by utilizing the previously obtained high-frequency baseline drift signal tensor of wrist biological tissue, dynamic compensation is performed on multi-source high-frequency signals. Specifically, a time-domain compensation algorithm or a frequency-domain compensation method is adopted, and the baseline drift signal tensor is applied to the original multi-source signal. The time-domain compensation method dynamically adjusts parameters such as the amplitude and frequency of the signal by performing real-time correction on the signal to reduce or eliminate the influence of high-frequency drift on the signal quality. The frequency-domain compensation method suppresses or corrects the high-frequency drift components in the frequency spectrum domain to restore the accuracy of the original signal. During the compensation process, the sliding window technique is used to locally process the signal at each moment to ensure the accuracy and real-time performance of the compensation. After the high-frequency drift dynamic compensation, the high-frequency drift compensation signal of the wrist biological tissue is finally obtained.

[0128] Further, the biological tissue feature analysis module includes the following functions:

[0129] Obtain the corresponding pulse wave rise time and pulse wave fall time through the clear PPG signal, and analyze the peak amplitude and dicrotic wave characteristics of the clear PPG signal based on the pulse wave rise time and pulse wave fall time to obtain the pulse wave PPG signal fluctuation characteristics, including the fluctuation amplitude and fluctuation phase corresponding to the pulse wave peak and dicrotic wave;

[0130] In the embodiments of the present invention, a high-quality PPG signal is obtained through a sensor, and noise is removed from the original PPG signal through an appropriate filter to ensure the clarity of the signal. Then, the pulse wave rise time (i.e., the time from the starting point of the signal to the maximum amplitude) and the pulse wave fall time (i.e., the time from the maximum amplitude to the recovery of the baseline) are extracted through a signal analysis algorithm. These time characteristics can accurately identify the peaks and valleys of the signal through digital signal processing techniques such as waveform detection and edge detection methods. During the pulse wave feature extraction process, the amplitude of the peak also needs to be analyzed to judge its amplitude change, especially paying attention to whether there is a dicrotic phenomenon. The characteristics of the dicrotic wave usually show an additional small peak and have the same rise and fall times as the main pulse wave. By combining time-domain and frequency-domain analysis, the fluctuation amplitude and fluctuation phase of the pulse wave are identified to comprehensively describe the dynamic characteristics of the pulse wave signal, and finally the pulse wave PPG signal fluctuation characteristics are obtained.

[0131] Preferably, perform electrocardiogram signal feature analysis on the corresponding P wave group, QRS wave group, and T wave group in the clear ECG signal to obtain the electrocardiogram ECG signal fluctuation characteristics, including the amplitude, width, and morphology corresponding to the P wave, QRS wave, and T wave;

[0132] In an embodiment of the present invention, an electrocardiogram (ECG) sensor is used to collect an ECG signal. After the signal is filtered to remove power frequency interference and electromyogram noise, a QRS detection algorithm (such as the Pan-Tompkins algorithm) is used to identify the starting point and ending point of the QRS complex, thereby determining the peak position of the R wave in the ECG signal. Based on the position of the R wave, the characteristics of the P wave group, QRS wave group, and T wave group are further extracted. The P wave reflects the activation of the atrium, the QRS wave group represents the depolarization process of the ventricle, and the T wave represents the repolarization process of the ventricle. During the analysis process, the amplitude, width, and morphological characteristics of these wave groups are calculated through feature extraction methods. For example, the amplitude of the P wave is usually small and the duration is short, the amplitude of the QRS wave group is large and the width is wide, while the T wave shows a relatively gentle rising and falling trend. Through these characteristics, the health status of the ECG signal can be judged more accurately, and finally the fluctuation characteristics of the ECG signal are obtained, including the amplitude, width, and morphology corresponding to the P wave, QRS wave, and T wave.

[0133] Preferably, the skin conductance characteristic analysis is performed on the clear GSR signal to obtain the fluctuation characteristics of the skin conductance GSR signal, including the skin conductance change rate and the average value of skin conductance fluctuations.

[0134] In an embodiment of the present invention, a galvanic skin response (GSR) sensor is used to obtain a skin conductance signal, which can reflect the change of moisture on the surface of the human skin and thus reflect emotional or physiological changes. First, the collected GSR signal is filtered to remove environmental noise and high-frequency interference to ensure the accuracy of the data. Then, the change rate of the skin conductance is calculated through time-domain analysis. The change rate is the change amplitude of the skin conductance per unit time, which can reveal the physiological response of the body in different situations. Next, the average value of the skin conductance fluctuations is calculated, which reflects the change trend of the skin conductance over a long period of time. By analyzing the fluctuation characteristics of the skin conductance, the stress response or other physiological characteristics of the human body can be evaluated, and the relationship between the physiological state and environmental stimuli can be further understood. Finally, the fluctuation characteristics of the skin conductance GSR signal are obtained, including the skin conductance change rate and the average value of skin conductance fluctuations.

[0135] Preferably, the fluctuation characteristics of the photoplethysmogram (PPG) signal, the fluctuation characteristics of the ECG signal, and the fluctuation characteristics of the GSR signal are combined to obtain a set of wrist biological tissue signal characteristics.

[0136] In the embodiment of the present invention, through data fusion technology, signal features from different sensors are combined. First, the pulse wave PPG signal fluctuation features extracted in step S31 (including pulse wave rise time, fall time, peak amplitude, and dicrotic wave fluctuation features), the electrocardiogram ECG signal fluctuation features extracted in step S32 (including the amplitude, width, and morphological features of P wave, QRS wave, and T wave), and the skin conductance GSR signal fluctuation features extracted in step S33 (including skin conductance change rate and fluctuation mean) are integrated. This integration process can be carried out by weighted average or machine learning methods to find effective relationships and patterns among different signal features, thereby forming a comprehensive wrist biological tissue signal feature set. This signal feature set can comprehensively reflect the physiological state and health status of an individual, and finally obtain the wrist biological tissue signal feature set.

[0137] Furthermore, the wrist biometric state evaluation module includes the following functions:

[0138] Remove feature redundancy from the wrist biological tissue signal feature set to obtain a wrist biological tissue signal redundancy-removed feature set;

[0139] In the embodiment of the present invention, by analyzing the correlation in the wrist biological tissue signal feature set, feature redundancy is removed. Redundant features refer to those features that affect each other in the feature space or repeatedly provide the same information. By discarding redundant or low-variance features, a redundancy-removed feature set is obtained, where each feature has high independence and information content, avoiding the interference brought by redundant features to the subsequent analysis process, and finally obtaining the wrist biological tissue signal redundancy-removed feature set.

[0140] Preferably, perform principal component feature dimensionality reduction processing on the wrist biological tissue signal redundancy-removed feature set and combine them to form a comprehensive feature vector to generate a wrist biological tissue signal feature vector;

[0141] In the embodiment of the present invention, based on the wrist biological tissue signal redundancy-removed feature set, principal component analysis (PCA) is applied for feature dimensionality reduction processing. The PCA algorithm projects the original multi-dimensional data into a lower-dimensional space through linear transformation of the feature set, retaining the principal components with the largest variance. Specifically, in implementation, by calculating the covariance matrix of the redundancy-removed feature set, the first few most representative principal components are extracted to reduce the feature dimension and improve the calculation efficiency. Then, these principal components are combined according to their importance weights in the dataset to form a comprehensive feature vector. This comprehensive feature vector can better capture the essential features of the wrist biological tissue signal, while reducing the computational complexity and the risk of overfitting, and finally generating a wrist biological tissue signal feature vector by combination.

[0142] Preferably, perform a topological distribution analysis on the wrist biological tissue signal feature vector to obtain the wrist biological tissue signal feature space distribution density;

[0143] In the embodiment of the present invention, through performing a topological distribution analysis based on the wrist biological tissue signal feature vector, the purpose of this analysis is to reveal the distribution density of signal features in a high-dimensional space and the topological relationship between features. Use clustering analysis or density estimation methods, such as Gaussian Mixture Model (GMM) or k-means clustering, to group the points in the feature space into multiple dense regions and sparse regions. Specifically, when implementing, dimensionality reduction methods such as t-SNE (t-Distributed Stochastic Neighbor Embedding) or UMAP (Uniform Manifold Approximation and Projection) can be used to map the high-dimensional features to a two-dimensional or three-dimensional space for visualizing and analyzing the distribution of features in the space. This analysis can help identify key regions or patterns in the wrist signal, and finally obtain the wrist biological tissue signal feature space distribution density.

[0144] Preferably, based on the wrist biological tissue signal feature space distribution density and combined with a convolutional neural network, perform an adaptive convolutional kernel design to use a 5x5 convolutional kernel to capture the global features of the wrist biological tissue signal in the signal feature region corresponding to dense features and gentle changes, and use a 1x1 convolutional kernel to capture the local detail features of the wrist biological tissue signal in the signal feature region corresponding to sparse features and drastic changes, so as to design and generate a wrist feature enhanced adaptive convolutional kernel model;

[0145] In the embodiment of the present invention, through designing an adaptive convolutional kernel according to the previously obtained feature space distribution density and combined with a convolutional neural network (CNN), first, based on the signal features in the dense region, use a 5x5 convolutional kernel to capture the global features of the signal, because the signal in these regions changes gently and is suitable for global perception with a larger convolutional kernel. For the sparse region, use a 1x1 convolutional kernel to capture the detail features, because the signal in these regions changes drastically, and a smaller convolutional kernel can effectively capture local rapid changes and detail information. The design of the adaptive convolutional kernel is optimized through the training process of the convolutional neural network. The size and shape of the convolutional kernel will be dynamically adjusted according to the characteristics of the signal to ensure the best feature extraction effect. Specifically, through the backpropagation algorithm, adjust the convolutional kernel weights so that the 5x5 convolutional kernel adapts to the dense region and the 1x1 convolutional kernel adapts to the sparse region, thereby enhancing the network's ability to capture different signal features, and finally design and generate a wrist feature enhanced adaptive convolutional kernel model.

[0146] Preferably, an adaptive convolution kernel model using wrist features is utilized to perform multi-scale convolution operations on the wrist biological tissue signal feature vector to generate a wrist multi-scale convolution enhanced feature vector; based on the wrist multi-scale convolution enhanced feature vector and combined with a fuzzy logic inference method, wrist state assessment is carried out to obtain a wrist physiological tissue state assessment index; based on the wrist physiological tissue state assessment index and combined with a multi-objective optimization algorithm, physiological assessment calibration is carried out to obtain a wrist biological tissue state assessment result.

[0147] In an embodiment of the present invention, by using the designed adaptive convolution kernel model to perform multi-scale convolution operations on the wrist biological tissue signal feature vector, by applying convolution kernels of different sizes (such as 5x5 and 1x1) to the signal feature vector, the global information and local detail information of the signal are extracted and enhanced. After convolution processing, the obtained multi-scale convolution enhanced feature vector can more comprehensively reflect the complexity and diversity of the wrist biological tissue signal. Then, combined with the fuzzy logic inference method, the multi-scale convolution enhanced feature vector is evaluated to generate a wrist state assessment index. The fuzzy logic inference method can perform state assessment by setting a rule base and a membership function under the condition of high uncertainty and ambiguity, so as to obtain a quantified assessment index. Based on this assessment index and combined with a multi-objective optimization algorithm, physiological assessment calibration is carried out to optimize the assessment result to obtain an accurate wrist biological tissue state assessment result. This process can provide an objective and accurate evaluation of the wrist health status, provide a basis for further health monitoring and intervention, and finally obtain a wrist biological tissue state assessment result.

[0148] Furthermore, the present invention also provides a wrist signal acquisition and processing method based on biological tissue analysis for implementing the wrist signal acquisition and processing system based on biological tissue analysis as described above. The wrist signal acquisition and processing method based on biological tissue analysis includes the following steps:

[0149] Step S1: Design a wrist wearable device integrating a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; obtain the distribution characteristics of wrist biological tissues, and based on the distribution characteristics of wrist biological tissues, optimize the position and structure of the wrist wearable device to generate an optimized wrist multi-signal acquisition device;

[0150] Step S2: Use the optimized wrist multi-signal acquisition device to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissue in real time to generate a wrist biological tissue PPG signal, a wrist biological tissue ECG signal, and a wrist biological tissue GSR signal; perform high-frequency drift preprocessing on the wrist biological tissue PPG signal, the wrist biological tissue ECG signal, and the wrist biological tissue GSR signal to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal;

[0151] Step S3: Perform biotissue feature analysis on the clear PPG signal, clear ECG signal, and clear GSR signal to obtain a wrist biotissue signal feature set;

[0152] Step S4: Perform feature redundancy and dimensionality reduction processing on the wrist biotissue signal feature set to generate a wrist biotissue signal feature vector; use a convolutional neural network to perform wrist physiological state assessment and analysis on the wrist biotissue signal feature vector to obtain a wrist biotissue state assessment result.

[0153] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application document are intended to be encompassed within the present invention.

[0154] The above description is only a specific implementation manner of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. A wrist signal acquisition and processing system based on biological tissue analysis, characterized in that, It includes the following modules: The wrist device sensing optimization design module is used to design a wrist wearable device integrated with a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; Obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist wearable device based on the distribution characteristics of wrist biological tissues to generate an optimized multi-signal acquisition device for the wrist; among them, it includes the following functions: Design a wrist wearable device integrated with a photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; Obtain the distribution characteristics of wrist biological tissues, including the distribution position of the radial artery in the wrist, the distribution position of the muscles in the wrist, and the distribution position of the nerves in the wrist; Based on the distribution position of the radial artery in the wrist, perform PPG sensing optimization design on the photoplethysmogram sensor in the wrist wearable device to accurately align the photoplethysmogram sensor with the distribution position of the radial artery in the wrist, and ensure close contact between the sensor and the wrist skin through a flexible fitting technology to generate an optimized design scheme for the pulse wave sensing position; Based on the distribution position of the muscles in the wrist, optimize the ECG electrode layout of the electrocardiogram sensor in the wrist wearable device to generate an optimized design scheme for the ECG sensing electrode layout; based on the distribution position of the nerves in the wrist, perform skin interference optimization design on the galvanic skin response sensor in the wrist wearable device to generate an optimized design scheme for the galvanic skin response sensing interference; among which the ECG electrode layout optimization includes: Based on the distribution position of the muscles in the wrist, use the muscle electrical interference impact calculation formula to calculate the electrical interference impact on the electrode layout corresponding to the electrocardiogram sensor in the wrist wearable device to obtain the ECG electrode layout electrical interference impact factor; among them, the muscle electrical interference impact calculation formula is specifically: ； Wherein, is the electro-interference influence factor of the ECG electrode layout, is the spatial area where the wrist-worn device is located, is the layout position of the electrocardiogram sensing electrode, is the number of position points of the electromyographic activity distribution, is the th muscle activity point, is the electromyographic activity intensity at the position of the th muscle activity point, is the maximum value of the electromyographic activity intensity, is the influence weight of the th muscle activity point on the electro-interference, is the exponential function, is the layout position corresponding to the th muscle activity point, is the standard deviation of the layout position corresponding to the th muscle activity point, is the maximum frequency of the electromyographic activity, is the correction factor of the electro-interference influence factor of the ECG electrode layout; Based on the ECG electrode layout electrical interference impact factor, optimize the ECG electrode layout of the electrocardiogram sensor in the wrist wearable device to generate an optimized design scheme for the ECG sensing electrode layout; Based on the optimized design scheme for the pulse wave sensing position, the optimized design scheme for the ECG sensing electrode layout, and the optimized design scheme for the galvanic skin response sensing interference, perform position and structure optimization design on the positions and layout structures corresponding to the photoplethysmogram sensor, the electrocardiogram sensor, and the galvanic skin response sensor in the wrist wearable device to generate an optimized multi-signal acquisition device for the wrist; The wrist multi-signal preprocessing module is used to use the optimized multi-signal acquisition device for the wrist to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissues in real time to generate the PPG signal of the wrist biological tissues, the ECG signal of the wrist biological tissues, and the GSR signal of the wrist biological tissues; perform high-frequency drift preprocessing on the PPG signal of the wrist biological tissues, the ECG signal of the wrist biological tissues, and the GSR signal of the wrist biological tissues to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal; The biological tissue feature analysis module is used to perform biological tissue feature analysis on the clear PPG signal, the clear ECG signal, and the clear GSR signal to obtain the wrist biological tissue signal feature set; A wrist biometric status evaluation module is used to perform feature redundancy and dimensionality reduction processing on the wrist biological tissue signal feature set to generate a wrist biological tissue signal feature vector; and use a convolutional neural network to perform wrist physiological status evaluation and analysis on the wrist biological tissue signal feature vector to obtain a wrist biological tissue status evaluation result.

2. The wrist signal acquisition and processing system based on biological tissue analysis according to claim 1, characterized in that The wrist multi-signal preprocessing module includes the following functions: Use the photoplethysmogram sensor in the wrist multi-signal acquisition optimization device to collect the PPG signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue PPG signal; Use the electrocardiogram sensor in the wrist multi-signal acquisition optimization device to collect the ECG signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue ECG signal; Use the galvanic skin response sensor in the wrist multi-signal acquisition optimization device to collect the GSR signal corresponding to the wrist biological tissue in real time to generate the wrist biological tissue GSR signal; Perform high-frequency drift preprocessing on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to generate a clear PPG signal, clear ECG signal, and clear GSR signal.

3. The wrist signal acquisition and processing system based on biological tissue analysis according to claim 2, wherein The high-frequency drift preprocessing of the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal includes: Perform timestamp alignment and synchronization processing on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to compensate for the time deviation corresponding to the signal based on the interpolation algorithm and ensure that the wrist biological tissue PPG signal, ECG signal, and GSR signal are completely aligned on the time axis to obtain a wrist biological tissue multi-source synchronous alignment signal set; Perform wavelet transform decomposition on the wrist biological tissue multi-source synchronous alignment signal set to separate the high-frequency and low-frequency components in the wrist biological tissue signal to obtain a wrist biological tissue multi-source high-frequency signal and a wrist biological tissue multi-source low-frequency signal; Perform high-frequency drift dynamic compensation on the wrist biological tissue multi-source high-frequency signal to obtain a wrist biological tissue high-frequency drift compensation signal; Perform low-frequency noise elimination processing on the wrist biological tissue multi-source low-frequency signal to further eliminate the residual noise in the low-frequency component corresponding to the wrist biological tissue signal to obtain a wrist biological tissue low-frequency noise removal signal; Perform multi-modal signal fusion on the wrist biological tissue high-frequency drift compensation signal and the wrist biological tissue low-frequency noise removal signal to generate a wrist biological tissue multi-modal clear signal map; based on the wrist biological tissue multi-modal clear signal map and combined with the signal reconstruction method, perform clear signal reconstruction output on the wrist biological tissue PPG signal, wrist biological tissue ECG signal, and wrist biological tissue GSR signal to generate a clear PPG signal, clear ECG signal, and clear GSR signal.

4. The wrist signal acquisition and processing system based on biological tissue analysis according to claim 3, characterized in that, The high-frequency drift dynamic compensation of the wrist biological tissue multi-source high-frequency signal includes: Perform high-frequency drift component analysis on the wrist biological tissue multi-source high-frequency signal to generate a wrist biological tissue high-frequency drift component map; Performing high-frequency baseline drift fitting processing on the high-frequency drift component map of the wrist biological tissue to obtain the high-frequency baseline drift signal tensor of the wrist biological tissue; Based on the high-frequency baseline drift signal tensor of the wrist biological tissue, high-frequency drift dynamic compensation is performed on the multi-source high-frequency signals of the wrist biological tissue to obtain a high-frequency drift compensation signal of the wrist biological tissue.

5. The wrist signal acquisition and processing system based on biological tissue analysis according to claim 1, characterized in that, The biological tissue feature analysis module includes the following functions: Obtain the corresponding pulse wave rise time and pulse wave fall time from the PPG clear signal, and perform peak amplitude and dicrotic fluctuation characteristic analysis on the PPG clear signal based on the pulse wave rise time and pulse wave fall time to obtain the pulse wave PPG signal fluctuation characteristics, including the fluctuation amplitude and fluctuation phase corresponding to the pulse wave peak and dicrotic; Perform ECG signal feature analysis on the corresponding P wave group, QRS wave group and T wave group in the clear ECG signal to obtain ECG signal fluctuation characteristics, including the amplitude, width and shape of the P wave, QRS wave and T wave; Perform skin conductance feature analysis on the GSR clear signal to obtain the skin conductance GSR signal fluctuation characteristics, including the skin conductance change rate and the skin conductance fluctuation mean; The pulse wave PPG signal fluctuation features, electrocardiogram (ECG) signal fluctuation features and skin conductance (GSR) signal fluctuation features are merged to obtain the wrist biological tissue signal feature set.

6. The wrist signal acquisition and processing system based on biological tissue analysis according to claim 1, wherein The wrist biometric status assessment module includes the following functions: performing feature redundancy removal on a wrist biological tissue signal feature set to obtain a wrist biological tissue signal redundancy removal feature set; Performing principal component feature dimensionality reduction processing on the redundant removal feature set of wrist biological tissue signals and combining them to form a comprehensive feature vector to generate a wrist biological tissue signal feature vector; Performing feature space topological distribution analysis on wrist biological tissue signal feature vectors to obtain the wrist biological tissue signal feature space distribution density; Based on the spatial distribution density of wrist tissue signal features and combined with convolutional neural networks, an adaptive convolution kernel design is performed. A 5x5 convolution kernel is used to capture the global features of wrist tissue signals in signal feature regions with dense features and gentle changes, and a 1x1 convolution kernel is used to capture the local details of wrist tissue signals in signal feature regions with sparse features and dramatic changes. This design generates a wrist feature enhancement adaptive convolution kernel model. A wrist feature enhancement adaptive convolution kernel model is used to perform multi-scale convolution operations on the wrist biological tissue signal feature vector to generate a wrist multi-scale convolution enhanced feature vector; based on the wrist multi-scale convolution enhanced feature vector and combined with the fuzzy logic reasoning method, the wrist state is assessed to obtain the wrist physiological tissue state assessment index; based on the wrist physiological tissue state assessment index and combined with the multi-objective optimization algorithm, physiological assessment calibration is performed to obtain the wrist biological tissue state assessment result.

7. A wrist signal acquisition and processing method based on biological tissue analysis, characterized in that, The wrist signal acquisition and processing system based on biological tissue analysis according to claim 1 is used to execute the wrist signal acquisition and processing method based on biological tissue analysis, comprising the following steps: Step S1: Design a wrist-worn device corresponding to an integrated photoplethysmogram sensor, an electrocardiogram sensor, and a galvanic skin response sensor; obtain the distribution characteristics of wrist biological tissues, and optimize the position and structure of the wrist-worn device based on the distribution characteristics of wrist biological tissues to generate an optimized multi-signal acquisition device for the wrist. Step S2: Use the optimized multi-signal acquisition device for the wrist to collect the PPG signal, ECG signal, and GSR signal corresponding to the wrist biological tissues in real time to generate the PPG signal of the wrist biological tissues, the ECG signal of the wrist biological tissues, and the GSR signal of the wrist biological tissues; perform high-frequency drift preprocessing on the PPG signal of the wrist biological tissues, the ECG signal of the wrist biological tissues, and the GSR signal of the wrist biological tissues to generate a clear PPG signal, a clear ECG signal, and a clear GSR signal. Step S3: Analyze the biological tissue characteristics of the clear PPG signal, the clear ECG signal, and the clear GSR signal to obtain a signal feature set of the wrist biological tissues. Step S4: Perform feature redundancy and dimensionality reduction processing on the signal feature set of the wrist biological tissues to generate a signal feature vector of the wrist biological tissues; use a convolutional neural network to perform wrist physiological state assessment and analysis on the signal feature vector of the wrist biological tissues to obtain a wrist biological tissue state assessment result.

Images