A method and system for bio-signal unscrambling based on skin deformation

Abstract

The application provides a skin deformation-based biosignal unwinding method and system, which is applied to the field of data processing; the original time sequence signals of skin micro displacement, strain and vibration change are collected by a deformation detection device based on a smart bracelet, combined with the motion information obtained by an acceleration and angular velocity sensor, motion interference can be recognized and inhibited in real time in a dynamic wearing environment, low-frequency motion components are separated from the skin deformation signal through a time-frequency masking strategy, and further matching, reconstruction and post-processing are performed on the sub-signal components, the continuity and stability of the target physiological signal can be effectively restored, and outliers caused by phase jumps and burst noises are eliminated.

Description

Technical Field

[0001] This invention relates to the field of data processing, and in particular to a method and system for untangling biological signals based on skin deformation. Background Technology

[0002] In the field of wearable smart devices, devices such as smart bracelets, flexible electronic patches, and electronic skin typically attach to the surface of human skin to sense minute deformations in real time and obtain the user's biometric signals, such as heart rate, respiratory rate, blood flow changes, and muscle activity. These technologies generally rely on piezoresistive, piezoelectric, or optical flexible sensors to convert the microscopic deformations of the skin driven by physiological activity into electrical or optical signals, thereby achieving indirect detection of physiological states.

[0003] However, in actual wearing scenarios, the sources of skin deformation signals are highly complex. On the one hand, various physiological activities of the human body (such as the periodic dilation of blood vessels caused by heartbeat, the slow displacement of tissues caused by respiration, and muscle contraction) will produce deformations of different frequencies and amplitudes on the skin surface. These signals superimpose with each other in time and space, forming strongly coupled mixed signals. On the other hand, the body motion interference generated by users during daily activities (such as walking, hand swinging, and posture changes) will introduce larger amplitude and non-rigid motion deformations, thereby further aggravating the degree of signal aliasing. Summary of the Invention

[0004] This invention aims to solve the problem that when users wear smart devices, the multi-source physiological information and motion interference in the skin deformation signal are severely mixed and difficult to separate effectively, and provides a biological signal untangling method and system based on skin deformation.

[0005] The present invention employs the following technical means to solve the technical problem: This invention provides a method for untangling biological signals based on skin deformation, comprising: Based on the deformation detection device preset in the smart bracelet, skin deformation information of the area where the user wears the bracelet is collected. Specifically, the skin deformation information is the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin. Determine whether the skin deformation information detects a preset motion interference; If so, the user's motion information is obtained through the accelerometer preset by the smart bracelet. According to the time-frequency masking strategy preset by the smart bracelet, the low-frequency motion component of the motion information is decomposed from the skin deformation information to remove interference components from non-physiological sources, and an intermediate signal after motion interference suppression is obtained. The intermediate signal is then processed to obtain each sub-signal component. Specifically, the motion information includes acceleration signal and angular velocity signal. Determine whether the sub-signal component matches the preset target physiological signal; If a match is found, the sub-signal component is reconstructed to restore the target biological feature of the sub-signal component. The target biological feature is then post-processed to eliminate phase jumps caused by measurement limitations and dynamically remove outliers caused by sudden noise. Specifically, the post-processing includes phase continuity correction and outlier detection and repair.

[0006] Furthermore, the step of decomposing the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet further includes: Based on the time-frequency masking strategy, a time-frequency masking matrix corresponding to the motion information is constructed. According to the distribution characteristics of the low-frequency motion components in the time-frequency masking matrix, the low-frequency region in the time-frequency spectrum is marked to generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency ranges. Determine whether the frequency band energy of the masked area exceeds a preset component threshold; If so, the frequency band energy is attenuated to zero. Based on the attenuation to zero, the user's current motion intensity is identified. Based on the current motion intensity, the strategy parameters of the time-frequency masking strategy are dynamically adjusted to generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include the masking threshold, the masking frequency band range, and the time window length.

[0007] Furthermore, before the step of removing non-physiological interference components to obtain the intermediate signal after motion interference suppression, and performing component processing on the intermediate signal to obtain each sub-signal component from the intermediate signal, the method further includes: Based on a preset time window, the intermediate signal is divided into several time segment signals. Feature sets of sub-signal components are extracted from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity, and stability parameters. Determine whether the energy proportion of the sub-signal component meets the preset range; If possible, the consistency of the changes of the sub-signal components in adjacent time windows is detected based on the degree of matching between the feature set and the physiological signal features. Based on the consistency of changes, unstable components of the sub-signal components are dynamically filtered out. Specifically, the unstable components are sub-signal components that exhibit abrupt changes or discontinuities in time.

[0008] Furthermore, the step of performing signal reconstruction operation on the sub-signal components to recover the target biological features of the sub-signal components further includes: Based on the candidate component set corresponding to the sub-signal component and the target organism feature, the main frequency deviation of each candidate component is detected from the candidate component set; Determine whether the main frequency deviation meets the preset threshold condition; If possible, a consistency check is performed on each sub-signal component to obtain the physical parameters of each sub-signal component. Based on the physical parameters, the physiological source of each sub-signal component is identified. Based on the physiological source, the phase difference of each sub-signal component is calculated. The phase compensation of each sub-signal component is dynamically adjusted through the phase difference. Specifically, the physical parameters include main frequency consistency, period stability, and phase change trend.

[0009] Furthermore, the step of determining whether the skin deformation information detects a preset motion interference also includes: Based on the characteristic parameters of the skin deformation information, the amplitude variation range of the skin deformation information within the current time window is calculated. Specifically, the characteristic parameters include the signal amplitude change rate, short-time energy, spectral distribution, main frequency variation, and signal stability. Determine whether the amplitude variation range is greater than a preset amplitude threshold; If so, the user's current motion state is identified, and the spectral expansion degree of the skin deformation information is obtained based on the current motion state. The bandwidth distribution of the spectrum is detected based on the spectral expansion degree, and the adaptive threshold of the smart bracelet is dynamically adjusted through the bandwidth distribution. The current motion state specifically includes stillness, light movement, and vigorous movement.

[0010] Furthermore, the step of determining whether the sub-signal component matches a preset target physiological signal further includes: Based on the periodic changes of the sub-signal components within a continuous time window, the periodic fluctuations of the sub-signal components are collected. Determine whether the periodic fluctuation is less than a preset fluctuation threshold; If so, then construct the physiological rhythm features of the sub-signal component, extract the corresponding fluctuation index from the physiological rhythm features, compare the fluctuation index with the preset template waveform for matching degree, and determine the sub-signal component with matching degree higher than the preset component as conforming to the target physiological signal according to the comparison result. The fluctuation index specifically includes peak position, rising edge feature, falling edge feature and waveform similarity.

[0011] Furthermore, the step of collecting skin deformation information in the user's wearing area based on the deformation detection device preset in the smart bracelet also includes: Based on the preset wearing and collection area of ​​the smart bracelet, the signal quality of the skin deformation information is identified; Determine whether the signal quality is lower than a preset quality threshold; If so, the gain parameter of the deformation detection device is dynamically adjusted according to the positional difference between the wearing acquisition area and the user wearing area, and the acquisition signal response of the smart bracelet is generated based on the gain parameter.

[0012] The present invention also provides a biosignal unwrapping system based on skin deformation, comprising: The acquisition module is used to acquire skin deformation information of the user's wearing area based on the deformation detection device preset in the smart bracelet. Specifically, the skin deformation information is the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin. The judgment module is used to determine whether the skin deformation information detects a preset motion interference; The execution module is configured to, if so, acquire the user's motion information through the preset accelerometer of the smart bracelet, decompose the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet, remove non-physiological interference components, obtain an intermediate signal after motion interference suppression, perform component processing on the intermediate signal, and obtain each sub-signal component from the intermediate signal, wherein the motion information specifically includes acceleration signal and angular velocity signal; The second judgment module is used to determine whether the sub-signal component matches a preset target physiological signal; The second execution module is used to perform signal reconstruction operation on the sub-signal component if a match is found, to restore the target biological feature of the sub-signal component, to perform post-processing operation on the target biological feature, and to eliminate phase jumps in the target biological feature caused by measurement limitations and dynamically remove outliers caused by sudden noise based on the post-processing operation. The post-processing operation specifically includes phase continuity correction and outlier detection and repair.

[0013] Furthermore, the execution module also includes: The generation unit is used to construct a time-frequency masking matrix corresponding to the motion information based on the time-frequency masking strategy, mark the low-frequency region in the time-frequency spectrum according to the distribution characteristics of the low-frequency motion component in the time-frequency masking matrix, and generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency intervals. The judgment unit is used to determine whether the frequency band energy of the masked area exceeds a preset component threshold. An execution unit is configured to, if so, attenuate the frequency band energy to zero, identify the user's current motion intensity based on the attenuation to zero, dynamically adjust the strategy parameters of the time-frequency masking strategy based on the current motion intensity, and generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include a masking threshold, a masking frequency band range, and a time window length.

[0014] Furthermore, it also includes: The extraction module is used to divide the intermediate signal into several time segment signals based on a preset time window, and extract the feature set of the sub-signal components from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity and stability parameters. The third judgment module is used to determine whether the energy ratio of the sub-signal component meets the preset range. The third execution module is used to, if possible, detect the consistency of changes of the sub-signal components in adjacent time windows based on the degree of matching between the feature set and the physiological signal features, and dynamically filter unstable components of the sub-signal components based on the consistency of changes, wherein the unstable components are specifically sub-signal components that exhibit abrupt changes or discontinuities in time.

[0015] This invention provides a method and system for untangling biological signals based on skin deformation, which has the following beneficial effects: This invention acquires raw time-series signals of minute displacement, strain, and vibration changes in the skin using a deformation detection device based on a smart bracelet. Combined with motion information obtained from acceleration and angular velocity sensors, it can identify and suppress motion interference in real time under dynamic wearing conditions. By using a time-frequency masking strategy to separate low-frequency motion components from the skin deformation signal, and further matching, reconstructing, and post-processing the sub-signal components, it can effectively restore the continuity and stability of the target physiological signal, eliminate outliers caused by phase jumps and sudden noise, and achieve accurate separation of multi-source physiological information from motion interference. This improves the extraction accuracy and robustness of physiological parameters such as heart rate and respiration of wearable devices under exercise conditions, providing reliable data support for user health monitoring and physiological state assessment, while enhancing the applicability of signal processing in complex dynamic environments. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating an embodiment of the biosignal unwrapping method based on skin deformation according to the present invention; Figure 2This is a structural block diagram of an embodiment of the biosignal unwrapping system based on skin deformation of the present invention. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The realization of the purpose, functional features, and advantages of the invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings.

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Reference Appendix Figure 1 The present invention provides a method for untangling biological signals based on skin deformation, comprising: S1: Based on the deformation detection device preset in the smart bracelet, collect skin deformation information of the area where the user wears the bracelet, wherein the skin deformation information is specifically the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin; S2: Determine whether the skin deformation information detects a preset motion interference; S3: If so, the user's motion information is obtained through the accelerometer preset by the smart bracelet. According to the time-frequency masking strategy preset by the smart bracelet, the low-frequency motion component of the motion information is decomposed from the skin deformation information to remove non-physiological interference components and obtain the intermediate signal after motion interference suppression. The intermediate signal is processed by component processing to obtain each sub-signal component from the intermediate signal. The motion information specifically includes acceleration signal and angular velocity signal. S4: Determine whether the sub-signal component matches the preset target physiological signal; S5: If a match is found, the sub-signal component is reconstructed to restore the target biological feature of the sub-signal component. The target biological feature is then post-processed to eliminate phase jumps caused by measurement limitations and dynamically remove outliers caused by sudden noise. Specifically, the post-processing operation includes phase continuity correction and outlier detection and repair.

[0020] In this embodiment, the system collects skin deformation information from the user's wearing area based on a pre-installed deformation detection device on the smart bracelet. Specifically, the skin deformation information is a raw time-series signal reflecting minute displacements, strains, and vibration changes in the skin. The system then determines whether this skin deformation information detects pre-defined motion interference and executes corresponding steps accordingly. For example, if the system determines that the skin deformation information in the user's wearing area does not detect pre-defined motion interference, the system assumes that the currently collected skin deformation signal mainly originates from the user's own physiological activities and is in a relatively stable state. The system then performs multi-component decomposition on the skin deformation information as an intermediate signal, extracting each sub-signal component. Subsequently, feature analysis and matching are performed on each sub-signal component to identify components that match the preset target physiological signal. Further signal reconstruction and post-processing operations are then performed, thereby improving processing efficiency and reducing unnecessary computational overhead. Conversely, if the system determines that the skin deformation information in the user's wearing area detects pre-defined motion interference, the system assumes that the currently collected skin deformation signal may not necessarily originate from the user's own physiological activities and is in an unstable state. The system then uses an accelerometer pre-installed on the smart bracelet to acquire the user's motion information, specifically including acceleration signals and... The angular velocity signal, based on a pre-set time-frequency masking strategy of the smart bracelet, decomposes the low-frequency motion component of motion information from skin deformation information, removes non-physiological interference components, and obtains intermediate signals after motion interference suppression. These intermediate signals undergo component processing to extract individual sub-signal components. The system, by introducing an accelerometer to acquire the user's motion information when motion interference is detected, and combining this with the time-frequency masking strategy to decompose and suppress low-frequency motion components, can effectively identify and remove non-physiological interference components caused by body movement. Compared to processing methods relying solely on a single skin deformation signal, this method significantly improves... The system's anti-interference capability in dynamic environments avoids the interference of motion artifacts on physiological signal extraction. Simultaneously, by performing component decomposition on the intermediate signal after interference removal, the originally coupled multi-source information is separated into multiple sub-signal components, which facilitates effective decoupling between different physiological signals. This process can separate signals with different frequency characteristics, such as heart rate and respiration, from complex mixed signals, improving the identifiability and extraction accuracy of the target physiological signal, thereby solving the problem of difficulty in separating multi-source physiological information from motion interference. The system then determines whether these sub-signal components match the pre-set target physiological signal to execute the corresponding steps.For example, when the system determines that these sub-signal components cannot match the pre-set target physiological signal, the system will consider that the currently decomposed sub-signal components do not contain effective physiological signals that meet the expected characteristics. The system will then re-execute the motion interference suppression process to enhance the ability to remove residual interference, expand or dynamically adjust the feature matching range of the target physiological signal to adapt to individual differences or state changes, and trigger the signal quality assessment and enhancement steps, i.e., perform signal-to-noise ratio analysis on the current signal and enhance the signal through filtering, smoothing, or gain adjustment. Conversely, when the system determines that these sub-signal components can match the pre-set target physiological signal, the system will consider that the currently decomposed sub-signal components effectively contain effective physiological signals that meet the expected characteristics. The system will then perform signal reconstruction operations on these sub-signal components to restore the target biological characteristics of these sub-signal components and perform post-processing operations on the target biological characteristics. The post-processing operations specifically include phase continuity correction and outlier detection and repair. Depending on the specific post-processing operation, phase jumps caused by measurement limitations in the target biological features are eliminated, and outliers caused by sudden noise are dynamically removed. The system reconstructs the physiological signal waveform by fusing and restoring the dispersed effective information through signal reconstruction of the matched sub-signal components. This process helps enhance the structural features of the target physiological signal, reduces information loss that may occur during component decomposition, improves the overall signal reconstruction accuracy, and makes the restored biological features closer to the actual physiological state. Simultaneously, by performing post-processing operations such as phase continuity correction and outlier detection and repair on the reconstructed target biological features, phase jump problems caused by measurement range limitations or signal processing can be effectively eliminated, and outliers caused by sudden noise are dynamically removed, thereby significantly improving the continuity, smoothness, and stability of the signal.

[0021] In another specific embodiment, the system continuously collects skin deformation information of the user's wearing area for 10 seconds at a sampling frequency of 100Hz, and simultaneously collects acceleration signals as motion information input. Within the time window from the 3rd to the 6th second, significant low-frequency fluctuations in the acceleration signal are detected. Spectral analysis determines that this low-frequency component is mainly concentrated in the 0.5Hz to 2Hz range. After mapping the skin deformation signal to the time-frequency domain, the system constructs a time-frequency masking matrix within the corresponding time window and frequency range, attenuating the signal amplitude within this frequency band, reducing the energy in this region by approximately 60%, thus obtaining the intermediate signal after motion interference suppression. Subsequently, the intermediate signal is divided into multiple sub-signal components according to frequency ranges, and the dominant frequency of each component is determined based on the spectral peak position. The energy percentage is calculated by the ratio of the corresponding frequency band energy to the total energy, and the period fluctuation is calculated by the period difference between adjacent time windows. For example, if a sub-signal component has a dominant frequency of 1.2Hz, an energy percentage of 35%, and a period fluctuation less than a set threshold of 5%, the system will assign this component to the appropriate sub-signal component. As a candidate physiological signal component, it is further matched with a preset heart rate template to calculate the peak position deviation and waveform similarity index. The matching degree is 0.92, which is higher than the preset threshold of 0.85, thus determining that the component is a valid target physiological signal component. At the same time, the phase difference of multiple sub-signal components from the same physiological source is calculated. For example, there is a time offset of about 0.15 seconds between adjacent components. The system performs phase compensation accordingly to align the reconstructed signal on the time axis. In terms of result verification, by comparing the signals before and after processing, the signal-to-noise ratio is calculated based on the ratio of the effective frequency band energy of the signal to the noise energy. The average results of multiple time windows are statistically analyzed. It can be observed that the main frequency of the target signal after processing is concentrated around 1.2Hz, the spectral bandwidth is significantly narrowed, the periodic fluctuation is reduced from the original 12% to 4%, the signal-to-noise ratio is improved by about 40%, and the abnormal fluctuation is significantly reduced. This shows that the method can effectively suppress motion interference, achieve the separation of multi-source signals, and improve the accuracy and stability of target physiological signal extraction.

[0022] It should be noted that the user's motion information is acquired through the accelerometer preset in the smart bracelet. Based on the time-frequency masking strategy preset in the smart bracelet, the low-frequency motion components of the motion information are decomposed from the skin deformation information to remove interference components from non-physiological sources, obtaining an intermediate signal after motion interference suppression. Component processing is then performed on the intermediate signal to extract each sub-signal component, specifically: First, the system acquires the user's motion information, including acceleration and angular velocity signals, using a preset accelerometer sensor to reflect changes in the user's movement during wear. When motion interference is detected in the skin deformation information, the system analyzes the motion information based on a preset time-frequency masking strategy, extracting low-frequency motion components and determining the temporal variation range and corresponding frequency range of these components. Subsequently, the skin deformation information is mapped to a time-frequency domain representation, and the time and frequency intervals corresponding to the low-frequency motion components are located within this domain, thus determining the specific distribution location of motion interference in the skin deformation information. Based on this, the system constructs a time-frequency masking range corresponding to the low-frequency motion components and suppresses the signal within this range, for example, by attenuating or limiting the amplitude of this portion of the signal, thereby weakening the low-frequency large-amplitude variation components introduced by motion, achieving the removal of non-physiological interference information, and obtaining an intermediate signal after motion interference suppression. Through this process, the main interference components caused by user movement in the original mixed signal are effectively separated, while retaining the effective signal components related to physiological activities. Specific examples are as follows: For example, when a user wears a smart bracelet and walks or swings their hand, the accelerometer detects periodically changing motion information and extracts the corresponding low-frequency motion component. This low-frequency motion component typically appears as a signal with a large amplitude and slow changes in skin deformation information, superimposed on the tiny periodic signals caused by heartbeat or respiration. Based on a preset time-frequency masking strategy, the system locates the time interval and frequency range consistent with this low-frequency motion component in the time-frequency distribution of skin deformation information and suppresses this part of the signal, thereby obtaining an intermediate signal with significantly reduced motion interference. Subsequently... The system performs component processing on the intermediate signal, splitting it into multiple sub-signal components based on the differences in frequency range and time variation characteristics of different signals. For example, it distinguishes between residual components with slower changes and lower frequencies and components with faster changes and stable periodicity. The sub-signal components with stable periodicity characteristics may correspond to the target physiological signal, while the remaining components may be residual interference or noise. Through the above processing, the system effectively splits the multi-source signals in the skin deformation information, clearly separating the target physiological signal from the complex mixed signal, providing a reliable foundation for subsequent matching and reconstruction.

[0023] It should be added that the sub-signal components are reconstructed to recover the target organism features of the sub-signal components. Post-processing is then performed on the target organism features to eliminate phase jumps caused by measurement limitations and dynamically remove outliers caused by sudden noise. Specifically: First, after the system obtains each sub-signal component from the intermediate signal, it performs signal reconstruction on the sub-signal components that match the preset target physiological signal characteristics to restore the target biological characteristics. Specifically, based on the sampling sequence and amplitude characteristics of each sub-signal component on the time axis, the system superimposes or fuses the originally dispersed sub-signal components in chronological order, so that the physiological signal information contained in different components complements each other, thereby generating a continuous and complete physiological signal waveform. During the reconstruction process, the system can normalize or adjust the amplitude of the signal to ensure that the contributions of different sub-signal components are coordinated and consistent, avoiding signal distortion caused by the amplitude of a certain component being too large or too small. In addition, the system can smooth the local time window of the reconstructed waveform to reduce the abrupt impact of short-term fluctuations and ensure the continuity and stability of the target biological characteristics. Subsequently, the system performs post-processing operations on the reconstructed target biosignal features, including phase continuity correction and outlier detection and repair. Phase continuity correction is used to handle signal periodic abrupt changes or phase jumps caused by measurement limitations, sensor sampling errors, or motion interference. By comparing and adjusting the phase relationship between adjacent periods, the reconstructed waveform maintains a smooth and continuous periodic change on the time axis. Outlier detection and repair is used to identify outlier amplitude points caused by sudden noise or external interference, such as short-term abnormal peaks or amplitude drops. The system automatically replaces, smooths, or interpolates outliers by comparing them with the amplitude average or trend of adjacent time windows, thereby eliminating the impact of outliers on the overall physiological signal waveform. Specific examples are as follows: For example, when a user wears a smart bracelet for daily activities, the wrist may swing slightly or be accidentally touched, causing some sub-signal components to momentarily exhibit abnormal amplitude or periodic jumps. If these components are directly used for target physiological signal extraction, it may lead to abnormal peaks in the heart rate signal or abrupt changes in the respiratory signal. Through signal reconstruction, the system fuses multiple sub-signal components in chronological order to form a complete and continuous heartbeat or respiratory signal waveform. Subsequently, phase continuity correction is performed to adjust the periodic jumps caused by sampling errors or short-term motion interference, restoring the waveform to a smooth and continuous period. At the same time, abnormal points with sudden increases or decreases in the waveform are detected and repaired, replacing outliers with nearby mean values ​​or performing smoothing. The final target biofeatures retain the periodicity and amplitude changes of the real physiological signals while effectively eliminating noise and abnormal interference, ensuring the accuracy and stability of physiological parameter extraction.

[0024] In this embodiment, step S3, which decomposes the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet, further includes: S31: Based on the time-frequency masking strategy, construct a time-frequency masking matrix corresponding to the motion information. According to the distribution characteristics of the low-frequency motion components in the time-frequency masking matrix, mark the low-frequency region in the time-frequency spectrum to generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency ranges. S32: Determine whether the frequency band energy of the masking area exceeds a preset component threshold; S33: If so, the frequency band energy is attenuated to zero. Based on the attenuation to zero, the user's current motion intensity is identified. Based on the current motion intensity, the strategy parameters of the time-frequency masking strategy are dynamically adjusted to generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include the masking threshold, the masking frequency band range, and the time window length.

[0025] In this embodiment, the system constructs a time-frequency masking matrix corresponding to motion information based on a time-frequency masking strategy. This matrix characterizes the intensity distribution of motion interference within different time windows and frequency ranges. Based on the distribution characteristics of low-frequency motion components in the time-frequency masking matrix, low-frequency regions in the time-spectrum graph are marked to generate masking regions corresponding to the motion interference. The system then determines whether the frequency band energy of the masking region exceeds a preset component threshold to execute the corresponding steps. For example, if the system determines that the frequency band energy of the masking region does not exceed the preset component threshold, the system considers the time window and frequency range to be within the specified range. The intensity of motion interference within the interval is low, and its impact on skin deformation signals is negligible. The system marks the signal within the masked area as a low-interference zone and directly participates in the generation of intermediate signals. Simultaneously, it records the energy information of this area as a reference for subsequent judgment and processing. This improves processing efficiency, avoids unnecessary signal suppression that could lead to the loss of effective physiological information, and enhances the fidelity of the target physiological signal in the original skin deformation signal, providing a reliable data foundation for subsequent component decomposition, matching, and reconstruction. For example, when the system determines that the frequency band energy of the masked area exceeds a pre-set component threshold, this... The system determines that the motion interference intensity within the specified time window and frequency range is strong, significantly impacting the skin deformation signal. Therefore, the system attenuates the frequency band energy to zero. Based on this attenuation, the system identifies the user's current motion intensity and dynamically adjusts the time-frequency masking strategy parameters according to different motion intensities. These parameters include the masking threshold, masking frequency band range, and time window length, generating an analysis of the energy changes in the skin deformation signal before and after processing. The system identifies the user's current motion intensity based on the attenuation zeroing operation and dynamically adjusts the time-frequency masking strategy parameters, including the masking threshold and masking frequency band range, according to different motion intensities. With its adaptive adjustment mechanism including the time window length, the system can flexibly adjust the interference suppression level according to the user's motion state. This ensures sufficient suppression of interference during high-motion states while avoiding excessive suppression of effective signals during low-motion states, thus achieving intelligent optimization of the signal processing strategy. Furthermore, by generating the energy changes of skin deformation signals before and after processing, the system can quantify the motion interference suppression effect and provide a reliable basis for subsequent component decomposition, signal reconstruction, and extraction of target physiological signals. This not only improves the controllability and stability of the signal processing flow but also enhances the system's adaptability to multi-source signal aliasing in actual wearing environments.

[0026] It should be noted that the frequency band energy is attenuated to zero. Based on this attenuation to zero, the user's current motion intensity is identified. Using this current motion intensity, the strategy parameters of the time-frequency masking strategy are dynamically adjusted to generate the energy change of the skin deformation signal before and after processing. Specifically: When the system determines that the frequency band energy of the masking area exceeds a preset component threshold, it indicates that the intensity of motion interference within that time window and frequency range is high, which may significantly affect the target physiological information in the skin deformation signal. To reduce the contribution of motion interference, the system first performs attenuation and zeroing of the frequency band energy, that is, reduces the signal amplitude within the corresponding time period and frequency range to near zero or completely eliminates it, thereby weakening the masking effect of motion interference on the target physiological signal. After the attenuation and zeroing operation is completed, the system analyzes the amplitude change, duration, and distribution characteristics of the attenuated signal to identify the user's current motion intensity, including slight, moderate, or vigorous motion, providing data for the dynamic adjustment of subsequent signal processing strategies. Based on this, the system dynamically adjusts the parameters of the time-frequency masking strategy according to the identified motion intensity. Specifically... In this system, the strategy parameters include the masking threshold, masking frequency band range, and time window length. Under high motion intensity, the system can increase the masking threshold and masking frequency band range, while extending the time window length to ensure sufficient suppression of high-intensity motion interference. Under low motion intensity, the system can reduce the masking amplitude, narrow the masking frequency band range, and shorten the time window length to avoid unnecessary suppression of effective physiological signals. Through this dynamic adjustment mechanism, the system can flexibly optimize the interference suppression effect under different motion states while ensuring the continuity and stability of the target physiological signal. At the same time, the system generates the energy change of the skin deformation signal before and after processing. This can not only be used to quantify the interference suppression effect but also provide a reliable basis for subsequent component decomposition, signal reconstruction, and target physiological signal extraction, providing controllability and transparency to the entire signal processing flow. Specific examples are as follows: For example, when a user wears a smart bracelet and walks briskly or swings their wrist, the low-frequency movement of the wrist will produce significant amplitude changes in the skin deformation signal, causing the energy of that frequency band to exceed a preset threshold. The system attenuates the amplitude of this frequency band signal to zero, effectively weakening the motion interference component. Subsequently, the system analyzes the magnitude and duration of the attenuation amplitude to determine that the user is in a high-intensity exercise state, and dynamically adjusts the time-frequency masking strategy parameters accordingly, including increasing the masking threshold, expanding the masking frequency band range, and extending the time window, to enhance the suppression capability of low-frequency interference during that time period. After the above processing, the generated skin deformation signal shows a significant difference in energy before and after processing. The changes in the data clearly show that motion interference has been effectively suppressed, while target physiological signals such as heart rate and respiration remain stable and continuous, avoiding period jumps or amplitude abnormalities caused by motion interference. Even when the user is engaged in moderate or slight exercise, the system will adjust the masking parameters according to the actual exercise intensity to avoid excessive suppression of effective signals. In this way, the system can adaptively optimize the interference suppression strategy in various wearing scenarios, improve the accuracy and stability of target physiological signal extraction, and ensure that the processing flow is highly robust to different exercise states, providing reliable support for physiological monitoring of smart bracelets in complex dynamic environments.

[0027] In this embodiment, before step S3, which involves removing non-physiological interference components to obtain an intermediate signal after motion interference suppression, and performing component processing on the intermediate signal to obtain each sub-signal component from the intermediate signal, the method further includes: S301: Based on a preset time window, the intermediate signal is divided into several time segment signals, and feature sets of sub-signal components are extracted from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity and stability parameters. S302: Determine whether the energy ratio of the sub-signal component meets the preset range; S303: If possible, then based on the degree of matching between the feature set and the physiological signal features, detect the consistency of the changes of the sub-signal components in adjacent time windows, and dynamically filter the unstable components of the sub-signal components based on the consistency of changes, wherein the unstable components are specifically sub-signal components that exhibit abrupt changes or discontinuities in time.

[0028] In this embodiment, the system divides the intermediate signal into several time-segment signals based on a pre-set time window, the length of which is dynamically set according to the periodic characteristics of the target physiological signal. From these time-segment signals, the system extracts feature sets of sub-signal components. These feature sets specifically include the dominant frequency, energy distribution, periodicity index, waveform similarity, and stability parameters. The system then determines whether the energy proportion of these sub-signal components meets a pre-set range to execute corresponding steps. For example, if the system determines that the energy proportion of these sub-signal components does not meet the pre-set range, it considers that the sub-signal components extracted from the intermediate signal may not fully reflect the effective characteristics of the target physiological signal. The system will then execute corresponding processing strategies, such as adjusting the time window length or re-dividing the intermediate signal to generate new time-segment signals, from which sub-signal components and their feature sets are re-extracted. Simultaneously, a time-frequency masking strategy is used to further suppress or correct frequency bands that may be affected by motion interference, thereby improving the energy proportion and reliability of the extracted sub-signal components. For example, if the system determines that the energy proportion of these sub-signal components meets the pre-set range, the system considers that the energy proportion of these sub-signal components extracted from the intermediate signal does not meet the pre-set range. The extracted sub-signal components can fully reflect the effective features of the target physiological signal. The system detects the consistency of changes of these sub-signal components in adjacent time windows based on the degree of matching between these feature sets and physiological signal features. Based on this consistency, the system dynamically filters unstable components, which are sub-signal components that exhibit abrupt changes or discontinuities in time. The system further detects the consistency of changes of sub-signal components in adjacent time windows based on the degree of matching between the extracted feature sets and the target physiological signal features. It dynamically filters unstable components, which are usually manifested as abrupt changes or discontinuities in the signal on the time axis. By removing or marking these components, the system can maintain the continuity and stability of the signal used for reconstruction, thereby reducing the interference of abnormal fluctuations or sudden noise on the final physiological signal. At the same time, through dynamic filtering and stability assurance, the system can adaptively handle the multi-source signal aliasing problem in complex wearing environments, and realize reliable monitoring and analysis of physiological parameters such as heart rate and respiration. This ensures that the sub-signal components used for the final reconstruction of the target physiological signal have both sufficient energy and temporal continuity and stability, thereby significantly improving the accuracy and precision of target physiological signal feature extraction.

[0029] It should be noted that, based on the degree of matching between the feature set and the physiological signal features, the consistency of changes in the sub-signal components within adjacent time windows is detected. Based on this consistency, unstable components of the sub-signal components are dynamically filtered out. Specifically: The system first evaluates the effectiveness of each sub-signal component based on the degree of matching between the extracted sub-signal component feature set and the preset target physiological signal features. The feature set includes dominant frequency, energy distribution, periodicity index, waveform similarity, and stability parameters. By comparing the similarity of these features with the target physiological signal features, the system can initially determine whether each sub-signal component may contain valid physiological information. Based on this, the system further detects the consistency of changes in the sub-signal components within adjacent time windows, i.e., observing whether the amplitude changes, periodic changes, and phase relationships of the same component remain stable in continuous time segments. If a sub-signal component exhibits abrupt changes, drastic amplitude fluctuations, or discontinuous periods within adjacent time windows, it is judged as an unstable component. Based on these detection results, the system dynamically filters unstable components, i.e., marking, removing, or reducing their weight, to ensure the continuity and stability of the signal used for subsequent reconstruction, thereby reducing the impact of external interference or noise on the extraction of the target physiological signal. Specific examples are as follows: For example, when a user wears a smart bracelet and performs light hand movements, wrist swings may cause some sub-signal components to experience a sudden increase in amplitude within a certain time window, followed by a rapid decrease in amplitude in the next time window, forming discontinuous or abrupt waveforms. By comparing the degree of matching between the dominant frequency and periodicity indicators of these sub-signal components and the characteristics of the heart rate signal, and combining waveform similarity and stability parameters, the system can identify these abnormally changing components that do not conform to the stable characteristics of the target physiological signal. Subsequently, the system dynamically filters and removes these unstable components so that they do not interfere with the final waveform in subsequent signal reconstruction and target physiological signal extraction. Through this mechanism, even in motion or interference environments, the system can still maintain the stability and continuity of the physiological signal components used for reconstruction, improving the accuracy and reliability of target physiological signal feature extraction such as heart rate or respiration.

[0030] In this embodiment, step S5, which involves performing signal reconstruction on the sub-signal components to recover the target biological characteristics of the sub-signal components, further includes: S51: Based on the candidate component set corresponding to the sub-signal component and the target organism feature, detect the main frequency deviation of each candidate component from the candidate component set; S52: Determine whether the main frequency deviation can meet the preset threshold condition; S53: If possible, perform consistency verification on each sub-signal component, obtain the physical parameters of each sub-signal component, identify the physiological source of each sub-signal component based on the physical parameters, calculate the phase difference of each sub-signal component based on the physiological source, and dynamically adjust the phase compensation of each sub-signal component through the phase difference. The physical parameters specifically include main frequency consistency, period stability, and phase change trend.

[0031] In this embodiment, the system detects the dominant frequency deviation of each candidate component from the candidate component set corresponding to the sub-signal components and the target biological characteristics. The system then determines whether these dominant frequency deviations meet a preset threshold condition to execute corresponding steps. For example, if the system determines that the dominant frequency deviation of each candidate component does not meet the preset threshold condition, the system considers that there is a significant difference between the current candidate component and the target biological characteristics, i.e., the dominant frequency of these components is inconsistent with the typical frequency of the target physiological signal. The system will remove components with excessively large dominant frequency deviations from the candidate component set, or, combined with the energy proportion, waveform similarity, and stability parameters of the aforementioned feature set, re-select components that better match the target physiological characteristics, improving the quality of the sub-signal components used for reconstruction and ensuring that the final extracted target physiological signal is highly consistent with the actual physiological state in terms of frequency characteristics, thereby improving the accuracy and stability of physiological parameter extraction. Conversely, if the system determines that the dominant frequency deviation of each candidate component meets the preset threshold condition, the system considers that there is no difference between the current candidate component and the target biological characteristics. The system will then perform consistency verification processing on each sub-signal component to obtain the... The system identifies the physical parameters of each sub-signal component, including frequency consistency, period stability, and phase change trend. Based on these parameters, the physiological origin of each sub-signal component is determined. The phase difference between each sub-signal component is calculated based on its physiological origin, and the phase compensation is dynamically adjusted accordingly. The system can distinguish different types of physiological signal components (such as heart rate, respiration, or muscle micro-motion signals) by identifying the physiological origin, ensuring correct classification of each component in subsequent signal reconstruction. By identifying the physiological origin, the system can adopt targeted processing strategies for different components, avoiding feature distortion caused by signal aliasing, thereby improving the accuracy of target physiological signal extraction. Simultaneously, the system calculates the phase difference of each sub-signal component based on the identified physiological origin and dynamically adjusts the corresponding phase compensation to correct the alignment of each component on the time axis. Through phase compensation, the system can eliminate period offsets caused by measurement limitations, sensor errors, or external interference, ensuring the continuity and stability of the final reconstructed target physiological signal in amplitude and period, significantly improving the extraction accuracy and reliability of physiological parameters (such as heart rate or respiratory rate).

[0032] It should be noted that consistency verification is performed on each sub-signal component to obtain the physical parameters of each sub-signal component. Based on the physical parameters, the physiological origin of each sub-signal component is identified. Based on the physiological origin, the phase difference of each sub-signal component is calculated. The phase compensation of each sub-signal component is dynamically adjusted based on the phase difference. Specifically: The system first performs consistency verification on each sub-signal component. By analyzing the amplitude changes, frequency stability, period continuity, and phase change trends of each component within adjacent time windows, components with abrupt amplitude changes, period discontinuities, or sudden anomalies are eliminated, thus ensuring the reliability and stability of the components used for subsequent signal reconstruction and physiological feature extraction. Based on this, the system further extracts the physical parameters of each component, including frequency consistency, period stability, and phase change trends. These parameters are used to identify the physiological origin of each sub-signal component, such as distinguishing between heart rate signals, respiratory signals, or micro-motion signals. Identifying the physiological origin enables targeted treatment of different signal types. Corresponding processing strategies are implemented to avoid feature distortion caused by signal aliasing or interference, and to provide a reference for dynamic phase compensation. The system calculates the phase difference between each sub-signal component based on the identified physiological source and performs dynamic phase compensation. By correcting the offset of each component on the time axis, the reconstructed target physiological signal maintains continuity and stability in amplitude, period, and phase, eliminating period drift and abrupt changes caused by wearing position, sensor error, or motion interference. This ensures the accuracy and reliability of the finally extracted physiological parameters such as heart rate and respiration. Even under conditions of user movement or complex wearing environment, the system can still maintain signal continuity, stability, and reconstructability. Specific examples are as follows: For example, when a user wears a smart bracelet and walks briskly or swings their wrist slightly, the heart rate signal component may experience slight amplitude jumps in certain time windows, while the respiratory signal component may have time delays or period drift. The system first performs consistency checks on the sub-signal components, eliminating components with abrupt amplitude changes or discontinuous periods, and extracts physical parameters such as the main frequency consistency, period stability, and phase change trend of each component. Subsequently, the system identifies the physiological sources of the heart rate and respiratory signals and calculates the phase difference between them, for example, the heart rate component leads the respiratory component by 0.2 seconds in time. Based on the calculation results, the system performs dynamic phase compensation on each component, aligning the reconstructed signals on the time axis, ensuring the continuity and periodic stability of the target physiological signal waveforms such as heart rate and respiration. Even under motion interference, it can accurately extract the target physiological characteristics, significantly improving the accuracy and reliability of physiological parameter measurement, while providing a stable signal basis for subsequent analysis and monitoring.

[0033] In this embodiment, step S2, which determines whether the skin deformation information detects a preset motion interference, further includes: S21: Based on the characteristic parameters of the skin deformation information, calculate the amplitude variation range of the skin deformation information within the current time window, wherein the characteristic parameters specifically include the signal amplitude change rate, short-time energy, spectral distribution, main frequency variation, and signal stability; S22: Determine whether the amplitude variation range is greater than a preset amplitude threshold; S23: If so, identify the user's current motion state, obtain the spectral expansion degree of the skin deformation information based on the current motion state, detect the bandwidth distribution of the spectrum based on the spectral expansion degree, and dynamically adjust the adaptive threshold of the smart bracelet through the bandwidth distribution. The current motion state specifically includes stillness, light movement, and vigorous movement.

[0034] In this embodiment, the system calculates the amplitude variation range of skin deformation information within the current time window based on characteristic parameters of skin deformation information, specifically including signal amplitude change rate, short-time energy, spectral distribution, dominant frequency variation, and signal stability. The system then determines whether this amplitude variation range exceeds a preset amplitude threshold to execute corresponding steps. For example, if the system determines that the amplitude variation range of skin deformation information within the current time window does not exceed the preset amplitude threshold, the system considers the skin deformation signal within that time window to be relatively stable in amplitude and with minimal signal fluctuation. The system utilizes skin deformation information within the current time window to extract sub-signal components and reconstruct target physiological signals. This stable signal serves as a reliable foundation for subsequent analysis and feature extraction, ensuring the accuracy and continuity of target physiological parameters such as heart rate and respiration. This improves the signal processing efficiency and reliability of the entire system under low-interference conditions. For example, if the system determines that the amplitude variation range of skin deformation information within the current time window exceeds a pre-set amplitude threshold, the system considers the skin deformation signal within that time window to have significant amplitude fluctuations. The system will then identify the user's current motion state. The specific motion states include stillness, mild motion, and vigorous motion. Based on different motion states, the system acquires the spectral spread of skin deformation information and detects the bandwidth distribution of the spectrum. Through different bandwidth distributions, the system dynamically adjusts the smart bracelet's adaptive threshold. The system acquires the spectral spread of skin deformation information and detects the bandwidth distribution of the spectrum based on the identified motion state. Spectral spread and bandwidth distribution reflect the energy distribution of the signal in different frequency ranges. By analyzing these parameters, the system can understand the intensity and characteristics of motion interference in the signal, thereby determining which frequency components may be affected by motion and require focused processing or suppression. Simultaneously, based on different bandwidth distributions, the system dynamically adjusts the smart bracelet's adaptive threshold, enabling subsequent sub-signal component extraction, signal decomposition, and target physiological signal reconstruction to be optimized for different motion states. Through adaptive threshold adjustment, the system can effectively reduce the impact of motion interference on skin deformation signals, improve signal continuity and reliability, and ensure the extraction accuracy and stability of target physiological parameters (such as heart rate and respiration). This allows the smart bracelet to maintain efficient and accurate physiological monitoring capabilities even when the user is moving or in complex wearing environments.

[0035] It should be noted that the process involves identifying the user's current motion state, obtaining the spectral spread of the skin deformation information based on the current motion state, detecting the bandwidth distribution of the spectrum based on the spectral spread, and dynamically adjusting the adaptive threshold of the smart bracelet using the bandwidth distribution. Specifically: The system first identifies the user's current motion state, such as stillness, mild movement, or vigorous movement, by analyzing the amplitude variation range of skin deformation information. The identification of motion state is based not only on the amplitude variation but also on the short-time energy and spectral distribution characteristics of the signal, thus determining whether signal fluctuations primarily originate from user movement or other interference. Subsequently, based on the identified motion state, the system calculates the spectral spread of skin deformation information within the current time window, analyzing the energy distribution range and diffusion degree of the signal in different frequency ranges, thereby quantifying the impact of motion interference on the signal. Based on the spectral spread, the system further detects the bandwidth distribution of the spectrum to determine the energy concentration in low-frequency or high-frequency regions, providing a reference for subsequent signal decomposition and motion interference suppression. By analyzing different bandwidth distributions, the system can dynamically adjust the smart bracelet's adaptive threshold, adapting the threshold to the signal fluctuation characteristics under stillness, mild movement, and vigorous movement states, ensuring the accuracy and stability of subsequent sub-signal component extraction and target physiological signal reconstruction. Specific examples are as follows: For example, when a user wears the bracelet and walks briskly, the amplitude fluctuation of the skin deformation signal increases significantly. Spectral analysis shows increased energy in the high-frequency region and a wider spectral spread. The bandwidth distribution indicates that the low-frequency signal is slightly interfered with, while the mid-to-high-frequency region is significantly affected by motion. The system identifies this state as mild to moderate exercise and dynamically increases the adaptive threshold based on the bandwidth distribution. This allows high-frequency motion interference components to be effectively suppressed during subsequent signal decomposition, while low-frequency physiological signals related to heart rate remain intact. Conversely, when the user is sitting still, the amplitude fluctuation is small and the spectral spread is narrow. The system automatically lowers the adaptive threshold, making the extraction of sub-signal components more sensitive and ensuring that weak physiological signals in the static state can be accurately captured. Through this mechanism, the smart bracelet can adaptively optimize the signal processing strategy under different exercise states, improving the extraction accuracy and stability of the target physiological signal.

[0036] In this embodiment, step S4, which determines whether the sub-signal component matches a preset target physiological signal, further includes: S41: Based on the periodic changes of the sub-signal component within a continuous time window, collect the periodic fluctuations of the sub-signal component; S42: Determine whether the periodic fluctuation is less than a preset fluctuation threshold; S43: If so, construct the physiological rhythm features of the sub-signal component, extract the corresponding fluctuation index from the physiological rhythm features, compare the fluctuation index with the preset template waveform for matching degree, and determine the sub-signal component with matching degree higher than the preset component as conforming to the target physiological signal according to the comparison result. The fluctuation index specifically includes peak position, rising edge feature, falling edge feature and waveform similarity.

[0037] In this embodiment, the system collects the periodic fluctuations of the sub-signal components based on their periodic changes within a continuous time window. The system then determines whether these periodic fluctuations are less than a preset fluctuation threshold to execute corresponding steps. For example, if the system determines that the periodic fluctuations of the sub-signal components are not less than the preset fluctuation threshold, the system considers the current sub-signal component to have a large periodic change within the continuous time window, indicating insufficient signal stability. The system will mark this sub-signal component as an unstable component and may take further processing measures, such as removing or reducing its weight in subsequent signal reconstruction. Simultaneously, it will comprehensively evaluate the sub-signal components based on the aforementioned physical parameters (such as frequency consistency, amplitude stability, and phase change trend) to ensure that the sub-signal components ultimately used for target physiological signal reconstruction have sufficient stability and continuity, thereby improving the accuracy and reliability of extracting physiological parameters such as heart rate and respiration. Conversely, if the system determines that the periodic fluctuations of the sub-signal components are less than the preset fluctuation threshold, the system considers the current sub-signal component to have no large periodic change within the continuous time window. The system will then construct the physiological rhythm characteristics of the sub-signal components and extract data from these physiological rhythm characteristics. The system extracts corresponding fluctuation indicators, including peak position, rising edge characteristics, falling edge characteristics, and waveform similarity. These fluctuation indicators are compared with pre-set template indicators. Based on the comparison results, sub-signal components with a matching degree higher than a preset component are identified as conforming to the target physiological signal. The system then constructs the physiological rhythm characteristics of these sub-signal components, extracting fluctuation indicators such as peak position, rising edge characteristics, falling edge characteristics, and waveform similarity. These indicators can describe the morphological characteristics and dynamic changes of the signal from multiple dimensions. By extracting these fine-grained features, the system can not only reflect the periodicity of the signal but also characterize its waveform structure, making the description of the target physiological signal more comprehensive and accurate. Simultaneously, the system compares the extracted fluctuation indicators with the preset template indicators and filters out sub-signal components with a matching degree higher than a preset component, thus determining that they conform to the target physiological signal characteristics. This mechanism achieves accurate signal identification through template matching, effectively reducing misjudgments and omissions, improving the accuracy and stability of target physiological signal extraction, and enabling the system to reliably identify key physiological parameters such as heart rate and respiration even in complex environments.

[0038] It should be noted that the process involves constructing the physiological rhythm features of the sub-signal components, extracting corresponding fluctuation indices from these features, comparing the fluctuation indices with a preset template waveform, and determining the sub-signal components with a matching degree higher than the preset component as conforming to the target physiological signal based on the comparison results. Specifically: After obtaining a sub-signal component that satisfies periodic stability, the system first constructs its physiological rhythm features. This involves modeling the periodic structure, amplitude fluctuations, and rhythmic patterns of the sub-signal component based on its waveform changes within a continuous time window, forming rhythmic features that characterize physiological activity. On this basis, the system extracts corresponding fluctuation indices from these physiological rhythm features, including parameters such as peak position, rising edge trend, falling edge trend, and overall waveform similarity. These fluctuation indices describe the morphological characteristics and variation patterns of the signal within one or more cycles. Subsequently, the system compares the extracted fluctuation indices with a pre-defined template waveform, evaluating the similarity between the current sub-signal component and the target physiological signal in terms of rhythmic structure and waveform morphology by calculating the consistency between various indices. When the matching degree is higher than a preset component threshold, the system determines that the sub-signal component conforms to the target physiological signal characteristics and uses it as a valid component in subsequent signal reconstruction and physiological parameter extraction, thereby ensuring the accuracy and stability of the final output signal in terms of rhythm and morphology. Specific examples are as follows: For example, in the process of heart rate signal extraction, the system constructs the physiological rhythm characteristics from a certain sub-signal component. It finds that this component exhibits regular periodic fluctuations within a continuous time window, with relatively consistent peak intervals, steep rising edges, and gentle falling edges. The overall waveform has a high similarity to the preset heart rate template waveform. The system further extracts the fluctuation index of this component and compares it with the corresponding index of the heart rate template waveform. The calculated consistency result is high, exceeding the preset threshold. Therefore, the sub-signal component is determined to be a valid heart rate signal component. Conversely, if some components have a certain periodicity but their peak positions are unstable or their waveforms are severely distorted, their matching degree is low and they will not be selected as the target physiological signal. Through this process, the system can accurately screen out the effective components that meet the target physiological characteristics from multi-source mixed signals, improving the accuracy and reliability of physiological signal extraction.

[0039] In this embodiment, step S1, which involves collecting skin deformation information of the user's wearing area based on a deformation detection device pre-set in the smart bracelet, further includes: S11: Based on the preset wearing and acquisition area of ​​the smart bracelet, identify the signal quality of the skin deformation information; S12: Determine whether the signal quality is lower than a preset quality threshold; S13: If so, then based on the positional difference between the wearing acquisition area and the user wearing area, the gain parameter of the deformation detection device is dynamically adjusted, and the acquisition signal response of the smart bracelet is generated according to the gain parameter.

[0040] In this embodiment, the system identifies the signal quality of the user's skin deformation information based on the pre-set wearing and acquisition area of ​​the smart bracelet. The system then determines whether the signal quality is lower than a pre-set quality threshold to execute corresponding steps. For example, if the system determines that the signal quality of the user's skin deformation information is not lower than the pre-set quality threshold, the system considers the currently acquired signal to have high integrity and stability, with minimal noise interference. The system will then use this skin deformation information for subsequent processing, such as motion interference detection, time-frequency masking, sub-signal component extraction, and target physiological signal matching and reconstruction. Simultaneously, the system will mark this signal as high-quality data for priority use in physiological parameter calculations or as a reference signal to improve the overall signal processing efficiency and the reliability of the results, thereby ensuring the accuracy and continuity of target physiological parameters such as heart rate and respiration extraction. Conversely, if the system determines that the signal quality of the user's skin deformation information is lower than the pre-set quality threshold... At this point, the system will consider the currently acquired signal to be incomplete or unstable. Based on the positional difference between the acquisition area and the user's actual wearing area, the system will dynamically adjust the gain parameter of the deformation detection device and generate the smart bracelet's acquisition signal response accordingly. The system adjusts the gain based on the positional difference between the acquisition area and the user's actual wearing area, allowing the signal acquisition process to adapt to different users' wearing methods and deviations. This adaptive adjustment capability effectively addresses practical usage scenarios such as inconsistent wearing tightness and positional shifts, improving the device's adaptability under various wearing conditions and enhancing the user experience. Simultaneously, by generating the smart bracelet's acquisition signal response based on the adjusted gain parameter, the system can maintain the stability and continuity of signal output in low-signal-quality environments, providing reliable input for subsequent signal processing, component decomposition, and physiological parameter extraction. This not only improves the robustness of the overall signal processing flow but also ensures that target physiological parameters such as heart rate and respiration maintain high measurement accuracy and reliability even in complex environments.

[0041] It should be noted that, based on the positional difference between the wearing area and the user's wearing area, the gain parameter of the deformation detection device is dynamically adjusted, and the acquisition signal response of the smart bracelet is generated according to the gain parameter, specifically as follows: When the system detects low signal quality, it first analyzes the positional relationship between the smart bracelet's preset wearing area and the user's actual wearing area. By comparing the degree of spatial offset between the two, it determines whether there is a deviation in the current sensor-skin contact state, such as loose fit, uneven local pressure, or misalignment. Based on this, the system dynamically adjusts the gain parameter of the deformation detection device according to the positional difference. When loose contact or significant signal attenuation is detected, the gain parameter is appropriately increased to enhance the signal amplitude; when local pressure or excessively strong signal is detected, the gain parameter is decreased to avoid signal saturation or distortion. In this way, the deformation detection device can output a signal with moderate amplitude and stable variation under different wearing conditions. Subsequently, the system generates a corresponding acquisition signal response based on the adjusted gain parameter, making the output signal more consistent with the actual skin deformation in terms of amplitude, dynamic range, and trend. This provides a stable and reliable data foundation for subsequent signal decomposition, feature extraction, and target physiological signal reconstruction. Specific examples are as follows: For example, when a user wears the smart bracelet outside the preset collection area or wears it loosely, the contact between the skin and the sensor weakens, resulting in a smaller amplitude and less noticeable fluctuation in the collected skin deformation signal. By detecting this positional difference, the system automatically increases the gain parameter of the deformation detection device, thereby amplifying the signal amplitude and making the originally weak deformation changes clear. Conversely, when the user wears it too tightly or there is significant local pressure, the signal may have an excessively high amplitude or local saturation. The system then reduces the gain parameter to restore the collected signal to a reasonable range. Through this dynamic adjustment process, the smart bracelet can output a stable and consistent collection signal response under different wearing conditions, effectively improving signal quality and ensuring the accuracy and reliability of subsequent physiological signal extraction.

[0042] Reference Appendix Figure 2 A biological signal unwrapping system based on skin deformation, as described in one embodiment of the present invention, includes: The acquisition module 10 is used to acquire skin deformation information of the user's wearing area based on the deformation detection device preset in the smart bracelet. Specifically, the skin deformation information is the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin. The judgment module 20 is used to determine whether the skin deformation information detects a preset motion interference; The execution module 30 is configured to, if so, acquire the user's motion information through the preset accelerometer of the smart bracelet, decompose the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet, remove non-physiological interference components, obtain an intermediate signal after motion interference suppression, perform component processing on the intermediate signal, and obtain each sub-signal component from the intermediate signal, wherein the motion information specifically includes acceleration signal and angular velocity signal; The second judgment module 40 is used to determine whether the sub-signal component matches the preset target physiological signal; The second execution module 50 is used to perform signal reconstruction operation on the sub-signal component if a match is found, to restore the target biological feature of the sub-signal component, to perform post-processing operation on the target biological feature, and to eliminate phase jumps in the target biological feature caused by measurement limitations and dynamically remove outliers caused by sudden noise according to the post-processing operation. The post-processing operation specifically includes phase continuity correction and outlier detection and repair.

[0043] In this embodiment, the acquisition module 10 collects skin deformation information of the user's wearing area based on the deformation detection device pre-installed in the smart bracelet. Specifically, the skin deformation information is a raw time-series signal reflecting minute displacements, strains, and vibration changes in the skin. Then, the judgment module 20 determines whether these skin deformation information signals detect pre-set motion interference, and executes corresponding steps accordingly. For example, when the system determines that the skin deformation information of the user's wearing area does not detect pre-set motion interference, the system considers the currently collected skin deformation signal to mainly originate from the user's own physiological activities, and the signal is in a relatively stable state. The system will then use the skin deformation information as an intermediate signal. The signal is decomposed into multiple components to extract individual sub-signal components. Then, feature analysis and matching are performed on each sub-signal component to identify components that match the preset target physiological signal. Further signal reconstruction and post-processing are then performed, thereby improving processing efficiency and reducing unnecessary computational overhead. For example, when the system detects pre-set motion interference in the skin deformation information of the user's wearing area, the execution module 30 considers the currently collected skin deformation signal to be potentially unstable and not necessarily originating from the user's own physiological activity. The system then uses the accelerometer sensor pre-installed on the smart bracelet to acquire the user's motion information, specifically including acceleration. Based on a pre-set time-frequency masking strategy on the smart bracelet, the low-frequency motion components of motion information are decomposed from skin deformation information to remove non-physiological interference components, resulting in intermediate signals after motion interference suppression. These intermediate signals undergo component processing to extract individual sub-signal components. By introducing an accelerometer to acquire user motion information when motion interference is detected, and combining this with the time-frequency masking strategy to decompose and suppress low-frequency motion components, the system can effectively identify and remove non-physiological interference components caused by body movement. Compared to processing methods relying solely on a single skin deformation signal, this method significantly improves system performance. The system's anti-interference capability in dynamic environments avoids motion artifacts from interfering with the extraction of physiological signals. Furthermore, by performing component decomposition on the intermediate signals after interference removal, the originally coupled multi-source information is separated into multiple sub-signal components, which facilitates effective decoupling between different physiological signals. This process can separate signals with different frequency characteristics, such as heart rate and respiration, from complex mixed signals, improving the identifiability and extraction accuracy of the target physiological signal, thereby solving the problem of difficulty in separating multi-source physiological information from motion interference. Then, the second judgment module 40 determines whether these sub-signal components match the pre-set target physiological signal to execute the corresponding steps.For example, when the system determines that these sub-signal components cannot match the pre-set target physiological signal, the system will consider that the currently decomposed sub-signal components do not contain effective physiological signals that meet the expected characteristics. The system will then re-execute the motion interference suppression process to enhance the ability to remove residual interference, expand or dynamically adjust the feature matching range of the target physiological signal to adapt to individual differences or state changes, and trigger the signal quality assessment and enhancement steps, that is, to perform signal-to-noise ratio analysis on the current signal and enhance the signal through filtering, smoothing or gain adjustment. For example, when the system determines that these sub-signal components can match the pre-set target physiological signal, the second execution module 50 will consider that the currently decomposed sub-signal components effectively contain effective physiological signals that meet the expected characteristics. The system will then perform signal reconstruction operations on these sub-signal components to restore the target biological characteristics of these sub-signal components and perform post-processing on the target biological characteristics. The post-processing operations specifically include phase continuity correction and outlier detection and repair. Depending on the specific post-processing operation, phase jumps caused by measurement limitations in the target biological features are eliminated, and outliers caused by sudden noise are dynamically removed. The system reconstructs the matched sub-signal components, fusing and restoring the dispersed effective information to reconstruct a more complete and clear physiological signal waveform. This process helps enhance the structural features of the target physiological signal, reduces information loss that may occur during component decomposition, improves the overall signal reconstruction accuracy, and makes the restored biological features closer to the actual physiological state. Simultaneously, by performing post-processing operations such as phase continuity correction and outlier detection and repair on the reconstructed target biological features, phase jump problems caused by measurement range limitations or signal processing can be effectively eliminated, and outliers caused by sudden noise are dynamically removed, thereby significantly improving the continuity, smoothness, and stability of the signal.

[0044] In this embodiment, the execution module further includes: The generation unit is used to construct a time-frequency masking matrix corresponding to the motion information based on the time-frequency masking strategy, mark the low-frequency region in the time-frequency spectrum according to the distribution characteristics of the low-frequency motion component in the time-frequency masking matrix, and generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency intervals. The judgment unit is used to determine whether the frequency band energy of the masked area exceeds a preset component threshold. An execution unit is configured to, if so, attenuate the frequency band energy to zero, identify the user's current motion intensity based on the attenuation to zero, dynamically adjust the strategy parameters of the time-frequency masking strategy based on the current motion intensity, and generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include a masking threshold, a masking frequency band range, and a time window length.

[0045] In this embodiment, the system constructs a time-frequency masking matrix corresponding to motion information based on a time-frequency masking strategy. This matrix characterizes the intensity distribution of motion interference within different time windows and frequency ranges. Based on the distribution characteristics of low-frequency motion components in the time-frequency masking matrix, low-frequency regions in the time-spectrum graph are marked to generate masking regions corresponding to the motion interference. The system then determines whether the frequency band energy of the masking region exceeds a preset component threshold to execute the corresponding steps. For example, if the system determines that the frequency band energy of the masking region does not exceed the preset component threshold, the system considers the time window and frequency range to be within the specified range. The intensity of motion interference within the interval is low, and its impact on skin deformation signals is negligible. The system marks the signal within the masked area as a low-interference zone and directly participates in the generation of intermediate signals. Simultaneously, it records the energy information of this area as a reference for subsequent judgment and processing. This improves processing efficiency, avoids unnecessary signal suppression that could lead to the loss of effective physiological information, and enhances the fidelity of the target physiological signal in the original skin deformation signal, providing a reliable data foundation for subsequent component decomposition, matching, and reconstruction. For example, when the system determines that the frequency band energy of the masked area exceeds a pre-set component threshold, this... The system determines that the motion interference intensity within the specified time window and frequency range is strong, significantly impacting the skin deformation signal. Therefore, the system attenuates the frequency band energy to zero. Based on this attenuation, the system identifies the user's current motion intensity and dynamically adjusts the time-frequency masking strategy parameters according to different motion intensities. These parameters include the masking threshold, masking frequency band range, and time window length, generating an analysis of the energy changes in the skin deformation signal before and after processing. The system identifies the user's current motion intensity based on the attenuation zeroing operation and dynamically adjusts the time-frequency masking strategy parameters, including the masking threshold and masking frequency band range, according to different motion intensities. With its adaptive adjustment mechanism including the time window length, the system can flexibly adjust the interference suppression level according to the user's motion state. This ensures sufficient suppression of interference during high-motion states while avoiding excessive suppression of effective signals during low-motion states, thus achieving intelligent optimization of the signal processing strategy. Furthermore, by generating the energy changes of skin deformation signals before and after processing, the system can quantify the motion interference suppression effect and provide a reliable basis for subsequent component decomposition, signal reconstruction, and extraction of target physiological signals. This not only improves the controllability and stability of the signal processing flow but also enhances the system's adaptability to multi-source signal aliasing in actual wearing environments.

[0046] In this embodiment, it also includes: The extraction module is used to divide the intermediate signal into several time segment signals based on a preset time window, and extract the feature set of the sub-signal components from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity and stability parameters. The third judgment module is used to determine whether the energy ratio of the sub-signal component meets the preset range. The third execution module is used to, if possible, detect the consistency of changes of the sub-signal components in adjacent time windows based on the degree of matching between the feature set and the physiological signal features, and dynamically filter unstable components of the sub-signal components based on the consistency of changes, wherein the unstable components are specifically sub-signal components that exhibit abrupt changes or discontinuities in time.

[0047] In this embodiment, the system divides the intermediate signal into several time-segment signals based on a pre-set time window, the length of which is dynamically set according to the periodic characteristics of the target physiological signal. From these time-segment signals, the system extracts feature sets of sub-signal components. These feature sets specifically include the dominant frequency, energy distribution, periodicity index, waveform similarity, and stability parameters. The system then determines whether the energy proportion of these sub-signal components meets a pre-set range to execute corresponding steps. For example, if the system determines that the energy proportion of these sub-signal components does not meet the pre-set range, it considers that the sub-signal components extracted from the intermediate signal may not fully reflect the effective characteristics of the target physiological signal. The system will then execute corresponding processing strategies, such as adjusting the time window length or re-dividing the intermediate signal to generate new time-segment signals, from which sub-signal components and their feature sets are re-extracted. Simultaneously, a time-frequency masking strategy is used to further suppress or correct frequency bands that may be affected by motion interference, thereby improving the energy proportion and reliability of the extracted sub-signal components. For example, if the system determines that the energy proportion of these sub-signal components meets the pre-set range, the system considers that the energy proportion of these sub-signal components extracted from the intermediate signal does not meet the pre-set range. The extracted sub-signal components can fully reflect the effective features of the target physiological signal. The system detects the consistency of changes of these sub-signal components in adjacent time windows based on the degree of matching between these feature sets and physiological signal features. Based on this consistency, the system dynamically filters unstable components, which are sub-signal components that exhibit abrupt changes or discontinuities in time. The system further detects the consistency of changes of sub-signal components in adjacent time windows based on the degree of matching between the extracted feature sets and the target physiological signal features. It dynamically filters unstable components, which are usually manifested as abrupt changes or discontinuities in the signal on the time axis. By removing or marking these components, the system can maintain the continuity and stability of the signal used for reconstruction, thereby reducing the interference of abnormal fluctuations or sudden noise on the final physiological signal. At the same time, through dynamic filtering and stability assurance, the system can adaptively handle the multi-source signal aliasing problem in complex wearing environments, and realize reliable monitoring and analysis of physiological parameters such as heart rate and respiration. This ensures that the sub-signal components used for the final reconstruction of the target physiological signal have both sufficient energy and temporal continuity and stability, thereby significantly improving the accuracy and precision of target physiological signal feature extraction.

[0048] In this embodiment, the second execution module further includes: The detection unit is used to detect the main frequency deviation of each candidate component from the candidate component set based on the candidate component set corresponding to the sub-signal component and the target organism feature; The second judgment unit is used to determine whether the main frequency deviation can meet the preset threshold condition; The second execution unit is configured to, if possible, perform consistency verification processing on each sub-signal component, obtain the physical parameters of each sub-signal component, identify the physiological source of each sub-signal component based on the physical parameters, calculate the phase difference of each sub-signal component based on the physiological source, and dynamically adjust the phase compensation of each sub-signal component through the phase difference. The physical parameters specifically include main frequency consistency, period stability, and phase change trend.

[0049] In this embodiment, the system detects the dominant frequency deviation of each candidate component from the candidate component set corresponding to the sub-signal components and the target biological characteristics. The system then determines whether these dominant frequency deviations meet a preset threshold condition to execute corresponding steps. For example, if the system determines that the dominant frequency deviation of each candidate component does not meet the preset threshold condition, the system considers that there is a significant difference between the current candidate component and the target biological characteristics, i.e., the dominant frequency of these components is inconsistent with the typical frequency of the target physiological signal. The system will remove components with excessively large dominant frequency deviations from the candidate component set, or, combined with the energy proportion, waveform similarity, and stability parameters of the aforementioned feature set, re-select components that better match the target physiological characteristics, improving the quality of the sub-signal components used for reconstruction and ensuring that the final extracted target physiological signal is highly consistent with the actual physiological state in terms of frequency characteristics, thereby improving the accuracy and stability of physiological parameter extraction. Conversely, if the system determines that the dominant frequency deviation of each candidate component meets the preset threshold condition, the system considers that there is no difference between the current candidate component and the target biological characteristics. The system will then perform consistency verification processing on each sub-signal component to obtain the... The system identifies the physical parameters of each sub-signal component, including frequency consistency, period stability, and phase change trend. Based on these parameters, the physiological origin of each sub-signal component is determined. The phase difference between each sub-signal component is calculated based on its physiological origin, and the phase compensation is dynamically adjusted accordingly. The system can distinguish different types of physiological signal components (such as heart rate, respiration, or muscle micro-motion signals) by identifying the physiological origin, ensuring correct classification of each component in subsequent signal reconstruction. By identifying the physiological origin, the system can adopt targeted processing strategies for different components, avoiding feature distortion caused by signal aliasing, thereby improving the accuracy of target physiological signal extraction. Simultaneously, the system calculates the phase difference of each sub-signal component based on the identified physiological origin and dynamically adjusts the corresponding phase compensation to correct the alignment of each component on the time axis. Through phase compensation, the system can eliminate period offsets caused by measurement limitations, sensor errors, or external interference, ensuring the continuity and stability of the final reconstructed target physiological signal in amplitude and period, significantly improving the extraction accuracy and reliability of physiological parameters (such as heart rate or respiratory rate).

[0050] In this embodiment, the determination module further includes: The calculation unit is used to calculate the amplitude variation range of the skin deformation information within the current time window based on the feature parameters of the skin deformation information, wherein the feature parameters specifically include the signal amplitude change rate, short-time energy, spectral distribution, main frequency variation and signal stability; The third judgment unit is used to determine whether the amplitude change range is greater than a preset amplitude threshold. The third execution unit is used to identify the user's current motion state if the current motion state is true, obtain the spectral spread of the skin deformation information based on the current motion state, detect the bandwidth distribution of the spectrum based on the spectral spread, and dynamically adjust the adaptive threshold of the smart bracelet based on the bandwidth distribution. The current motion state specifically includes stillness, light movement, and vigorous movement.

[0051] In this embodiment, the system calculates the amplitude variation range of skin deformation information within the current time window based on characteristic parameters of skin deformation information, specifically including signal amplitude change rate, short-time energy, spectral distribution, dominant frequency variation, and signal stability. The system then determines whether this amplitude variation range exceeds a preset amplitude threshold to execute corresponding steps. For example, if the system determines that the amplitude variation range of skin deformation information within the current time window does not exceed the preset amplitude threshold, the system considers the skin deformation signal within that time window to be relatively stable in amplitude and with minimal signal fluctuation. The system utilizes skin deformation information within the current time window to extract sub-signal components and reconstruct target physiological signals. This stable signal serves as a reliable foundation for subsequent analysis and feature extraction, ensuring the accuracy and continuity of target physiological parameters such as heart rate and respiration. This improves the signal processing efficiency and reliability of the entire system under low-interference conditions. For example, if the system determines that the amplitude variation range of skin deformation information within the current time window exceeds a pre-set amplitude threshold, the system considers the skin deformation signal within that time window to have significant amplitude fluctuations. The system will then identify the user's current motion state. The specific motion states include stillness, mild motion, and vigorous motion. Based on different motion states, the system acquires the spectral spread of skin deformation information and detects the bandwidth distribution of the spectrum. Through different bandwidth distributions, the system dynamically adjusts the smart bracelet's adaptive threshold. The system acquires the spectral spread of skin deformation information and detects the bandwidth distribution of the spectrum based on the identified motion state. Spectral spread and bandwidth distribution reflect the energy distribution of the signal in different frequency ranges. By analyzing these parameters, the system can understand the intensity and characteristics of motion interference in the signal, thereby determining which frequency components may be affected by motion and require focused processing or suppression. Simultaneously, based on different bandwidth distributions, the system dynamically adjusts the smart bracelet's adaptive threshold, enabling subsequent sub-signal component extraction, signal decomposition, and target physiological signal reconstruction to be optimized for different motion states. Through adaptive threshold adjustment, the system can effectively reduce the impact of motion interference on skin deformation signals, improve signal continuity and reliability, and ensure the extraction accuracy and stability of target physiological parameters (such as heart rate and respiration). This allows the smart bracelet to maintain efficient and accurate physiological monitoring capabilities even when the user is moving or in complex wearing environments.

[0052] In this embodiment, the second determination module further includes: The acquisition unit is used to acquire the periodic fluctuation of the sub-signal component based on the periodic change of the sub-signal component within a continuous time window; The fourth judgment unit is used to determine whether the periodic fluctuation is less than a preset fluctuation threshold. The fourth execution unit is used to construct the physiological rhythm features of the sub-signal component if the condition is met, extract the corresponding fluctuation index from the physiological rhythm features, compare the fluctuation index with the preset template waveform, and determine the sub-signal component with a matching degree higher than the preset component as conforming to the target physiological signal according to the comparison result. The fluctuation index specifically includes peak position, rising edge feature, falling edge feature and waveform similarity.

[0053] In this embodiment, the system collects the periodic fluctuations of the sub-signal components based on their periodic changes within a continuous time window. The system then determines whether these periodic fluctuations are less than a preset fluctuation threshold to execute corresponding steps. For example, if the system determines that the periodic fluctuations of the sub-signal components are not less than the preset fluctuation threshold, the system considers the current sub-signal component to have a large periodic change within the continuous time window, indicating insufficient signal stability. The system will mark this sub-signal component as an unstable component and may take further processing measures, such as removing or reducing its weight in subsequent signal reconstruction. Simultaneously, it will comprehensively evaluate the sub-signal components based on the aforementioned physical parameters (such as frequency consistency, amplitude stability, and phase change trend) to ensure that the sub-signal components ultimately used for target physiological signal reconstruction have sufficient stability and continuity, thereby improving the accuracy and reliability of extracting physiological parameters such as heart rate and respiration. Conversely, if the system determines that the periodic fluctuations of the sub-signal components are less than the preset fluctuation threshold, the system considers the current sub-signal component to have no large periodic change within the continuous time window. The system will then construct the physiological rhythm characteristics of the sub-signal components and extract data from these physiological rhythm characteristics. The system extracts corresponding fluctuation indicators, including peak position, rising edge characteristics, falling edge characteristics, and waveform similarity. These fluctuation indicators are compared with pre-set template indicators. Based on the comparison results, sub-signal components with a matching degree higher than a preset component are identified as conforming to the target physiological signal. The system then constructs the physiological rhythm characteristics of these sub-signal components, extracting fluctuation indicators such as peak position, rising edge characteristics, falling edge characteristics, and waveform similarity. These indicators can describe the morphological characteristics and dynamic changes of the signal from multiple dimensions. By extracting these fine-grained features, the system can not only reflect the periodicity of the signal but also characterize its waveform structure, making the description of the target physiological signal more comprehensive and accurate. Simultaneously, the system compares the extracted fluctuation indicators with the preset template indicators and filters out sub-signal components with a matching degree higher than a preset component, thus determining that they conform to the target physiological signal characteristics. This mechanism achieves accurate signal identification through template matching, effectively reducing misjudgments and omissions, improving the accuracy and stability of target physiological signal extraction, and enabling the system to reliably identify key physiological parameters such as heart rate and respiration even in complex environments.

[0054] In this embodiment, the acquisition module further includes: The identification unit is used to identify the signal quality of the skin deformation information based on the preset wearing and collection area of ​​the smart bracelet; The fifth judgment unit is used to determine whether the signal quality is lower than a preset quality threshold. The fifth execution unit is configured to, if so, dynamically adjust the gain parameter of the deformation detection device based on the positional difference between the wearing acquisition area and the user wearing area, and generate the acquisition signal response of the smart bracelet based on the gain parameter.

[0055] In this embodiment, the system identifies the signal quality of the user's skin deformation information based on the pre-set wearing and acquisition area of ​​the smart bracelet. The system then determines whether the signal quality is lower than a pre-set quality threshold to execute corresponding steps. For example, if the system determines that the signal quality of the user's skin deformation information is not lower than the pre-set quality threshold, the system considers the currently acquired signal to have high integrity and stability, with minimal noise interference. The system will then use this skin deformation information for subsequent processing, such as motion interference detection, time-frequency masking, sub-signal component extraction, and target physiological signal matching and reconstruction. Simultaneously, the system will mark this signal as high-quality data for priority use in physiological parameter calculations or as a reference signal to improve the overall signal processing efficiency and the reliability of the results, thereby ensuring the accuracy and continuity of target physiological parameters such as heart rate and respiration extraction. Conversely, if the system determines that the signal quality of the user's skin deformation information is lower than the pre-set quality threshold... At this point, the system will consider the currently acquired signal to be incomplete or unstable. Based on the positional difference between the acquisition area and the user's actual wearing area, the system will dynamically adjust the gain parameter of the deformation detection device and generate the smart bracelet's acquisition signal response accordingly. The system adjusts the gain based on the positional difference between the acquisition area and the user's actual wearing area, allowing the signal acquisition process to adapt to different users' wearing methods and deviations. This adaptive adjustment capability effectively addresses practical usage scenarios such as inconsistent wearing tightness and positional shifts, improving the device's adaptability under various wearing conditions and enhancing the user experience. Simultaneously, by generating the smart bracelet's acquisition signal response based on the adjusted gain parameter, the system can maintain the stability and continuity of signal output in low-signal-quality environments, providing reliable input for subsequent signal processing, component decomposition, and physiological parameter extraction. This not only improves the robustness of the overall signal processing flow but also ensures that target physiological parameters such as heart rate and respiration maintain high measurement accuracy and reliability even in complex environments.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for untangling biological signals based on skin deformation, characterized in that, Includes the following steps: Based on the deformation detection device preset in the smart bracelet, skin deformation information of the area where the user wears the bracelet is collected. Specifically, the skin deformation information is the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin. Determine whether the skin deformation information detects a preset motion interference; If so, the user's motion information is obtained through the accelerometer preset by the smart bracelet. According to the time-frequency masking strategy preset by the smart bracelet, the low-frequency motion component of the motion information is decomposed from the skin deformation information to remove interference components from non-physiological sources, and an intermediate signal after motion interference suppression is obtained. The intermediate signal is then processed to obtain each sub-signal component. Specifically, the motion information includes acceleration signal and angular velocity signal. Determine whether the sub-signal component matches the preset target physiological signal; If a match is found, the sub-signal component is reconstructed to restore the target biological feature of the sub-signal component. The target biological feature is then post-processed to eliminate phase jumps caused by measurement limitations and dynamically remove outliers caused by sudden noise. Specifically, the post-processing includes phase continuity correction and outlier detection and repair.

2. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, The step of decomposing the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet further includes: Based on the time-frequency masking strategy, a time-frequency masking matrix corresponding to the motion information is constructed. According to the distribution characteristics of the low-frequency motion components in the time-frequency masking matrix, the low-frequency region in the time-frequency spectrum is marked to generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency ranges. Determine whether the frequency band energy of the shielded area exceeds a preset component threshold; If so, the frequency band energy is attenuated to zero. Based on the attenuation to zero, the user's current motion intensity is identified. Based on the current motion intensity, the strategy parameters of the time-frequency masking strategy are dynamically adjusted to generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include the masking threshold, the masking frequency band range, and the time window length.

3. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, Before the step of removing non-physiological interference components to obtain the intermediate signal after motion interference suppression, and performing component processing on the intermediate signal to obtain each sub-signal component from the intermediate signal, the method further includes: Based on a preset time window, the intermediate signal is divided into several time segment signals. Feature sets of sub-signal components are extracted from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity, and stability parameters. Determine whether the energy proportion of the sub-signal component meets the preset range; If possible, the consistency of the changes of the sub-signal components in adjacent time windows is detected based on the degree of matching between the feature set and the physiological signal features. Based on the consistency of changes, unstable components of the sub-signal components are dynamically filtered out. Specifically, the unstable components are sub-signal components that exhibit abrupt changes or discontinuities in time.

4. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, The step of performing signal reconstruction on the sub-signal components to recover the target biological features of the sub-signal components further includes: Based on the candidate component set corresponding to the sub-signal component and the target organism feature, the main frequency deviation of each candidate component is detected from the candidate component set; Determine whether the main frequency deviation meets the preset threshold condition; If possible, a consistency check is performed on each sub-signal component to obtain the physical parameters of each sub-signal component. Based on the physical parameters, the physiological source of each sub-signal component is identified. Based on the physiological source, the phase difference of each sub-signal component is calculated. The phase compensation of each sub-signal component is dynamically adjusted through the phase difference. Specifically, the physical parameters include main frequency consistency, period stability, and phase change trend.

5. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, The step of determining whether the skin deformation information detects a preset motion interference further includes: Based on the characteristic parameters of the skin deformation information, the amplitude variation range of the skin deformation information within the current time window is calculated. Specifically, the characteristic parameters include the signal amplitude change rate, short-time energy, spectral distribution, main frequency variation, and signal stability. Determine whether the amplitude variation range is greater than a preset amplitude threshold; If so, the user's current motion state is identified, and the spectral expansion degree of the skin deformation information is obtained based on the current motion state. The bandwidth distribution of the spectrum is detected based on the spectral expansion degree, and the adaptive threshold of the smart bracelet is dynamically adjusted through the bandwidth distribution. The current motion state specifically includes stillness, light movement, and vigorous movement.

6. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, The step of determining whether the sub-signal component matches the preset target physiological signal further includes: Based on the periodic changes of the sub-signal components within a continuous time window, the periodic fluctuations of the sub-signal components are collected. Determine whether the periodic fluctuation is less than a preset fluctuation threshold; If so, then construct the physiological rhythm features of the sub-signal component, extract the corresponding fluctuation index from the physiological rhythm features, compare the fluctuation index with the preset template waveform for matching degree, and determine the sub-signal component with matching degree higher than the preset component as conforming to the target physiological signal according to the comparison result. The fluctuation index specifically includes peak position, rising edge feature, falling edge feature and waveform similarity.

7. The biosignal unwrapping method based on skin deformation according to claim 1, characterized in that, The step of collecting skin deformation information in the user's wearing area based on the deformation detection device preset in the smart bracelet also includes: Based on the preset wearing and acquisition area of ​​the smart bracelet, the signal quality of the skin deformation information is identified; Determine whether the signal quality is lower than a preset quality threshold; If so, the gain parameter of the deformation detection device is dynamically adjusted according to the positional difference between the wearing acquisition area and the user wearing area, and the acquisition signal response of the smart bracelet is generated based on the gain parameter.

8. A biosignal unwrapping system based on skin deformation, characterized in that, include: The acquisition module is used to acquire skin deformation information of the user's wearing area based on the deformation detection device preset in the smart bracelet. Specifically, the skin deformation information is the original time-series signal reflecting the minute displacement, strain and vibration changes of the skin. The judgment module is used to determine whether the skin deformation information detects a preset motion interference; The execution module is configured to, if so, acquire the user's motion information through the preset accelerometer of the smart bracelet, decompose the low-frequency motion component of the motion information from the skin deformation information according to the preset time-frequency masking strategy of the smart bracelet, remove non-physiological interference components, obtain an intermediate signal after motion interference suppression, perform component processing on the intermediate signal, and obtain each sub-signal component from the intermediate signal, wherein the motion information specifically includes acceleration signal and angular velocity signal; The second judgment module is used to determine whether the sub-signal component matches a preset target physiological signal; The second execution module is used to perform signal reconstruction operation on the sub-signal component if a match is found, to restore the target biological feature of the sub-signal component, to perform post-processing operation on the target biological feature, and to eliminate phase jumps in the target biological feature caused by measurement limitations and dynamically remove outliers caused by sudden noise based on the post-processing operation. The post-processing operation specifically includes phase continuity correction and outlier detection and repair.

9. The biosignal unwrapping system based on skin deformation according to claim 8, characterized in that, The execution module further includes: The generation unit is used to construct a time-frequency masking matrix corresponding to the motion information based on the time-frequency masking strategy, mark the low-frequency region in the time-frequency spectrum according to the distribution characteristics of the low-frequency motion component in the time-frequency masking matrix, and generate a masking region corresponding to the motion interference. The time-frequency masking matrix is ​​used to characterize the intensity distribution of motion interference in different time windows and frequency intervals. The judgment unit is used to determine whether the frequency band energy of the masked area exceeds a preset component threshold. An execution unit is configured to, if so, attenuate the frequency band energy to zero, identify the user's current motion intensity based on the attenuation to zero, dynamically adjust the strategy parameters of the time-frequency masking strategy based on the current motion intensity, and generate the energy change of the skin deformation signal before and after processing. The strategy parameters specifically include a masking threshold, a masking frequency band range, and a time window length.

10. The biosignal unwrapping system based on skin deformation according to claim 8, characterized in that, Also includes: The extraction module is used to divide the intermediate signal into several time segment signals based on a preset time window, and extract the feature set of the sub-signal components from the time segment signals. The length of the time window is dynamically set according to the periodic characteristics of the target physiological signal. The feature set specifically includes the main frequency, energy distribution, periodicity index, waveform similarity and stability parameters. The third judgment module is used to determine whether the energy ratio of the sub-signal component meets the preset range. The third execution module is used to, if possible, detect the consistency of changes of the sub-signal components in adjacent time windows based on the degree of matching between the feature set and the physiological signal features, and dynamically filter unstable components of the sub-signal components based on the consistency of changes, wherein the unstable components are specifically sub-signal components that exhibit abrupt changes or discontinuities in time.

Images