Psychological risk dynamic assessment system and method based on sleep HRV characteristics

Abstract

The invention relates to the technical field of psychological health monitoring, and particularly discloses a psychological risk dynamic assessment system and method based on sleep HRV characteristics, the system comprises a signal acquisition module, the signal acquisition module is integrated with a piezoelectric film sensor, a signal amplification circuit, an MCU and a data transmission module, the piezoelectric film sensor is arranged under a mattress, and the signal amplification circuit is connected with the signal acquisition module; the signal acquisition module is used for acquiring a BCG signal; the data processing module is used for processing the BCG signal and converting the BCG signal into an HRV (High Rate Voltage) characteristic value; the risk assessment module is used for generating a dynamic psychological risk assessment result through an HRV-emotion index mapping model on the basis of the HRV feature value; the user interaction module is used for displaying the dynamic psychological risk assessment result through a man-machine interaction device; bCG signals during sleep are collected in a non-contact mode, and HRV features are extracted based on the BCG signals, so that dynamic assessment of psychological risks is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of mental health monitoring, and in particular to a system and method for dynamic assessment of mental risk based on sleep HRV characteristics. Background Art

[0002] Against the backdrop of a deepening global mental health crisis, early detection and intervention of mental illness have become crucial issues. Traditional psychological assessment methods have significant limitations: subjective assessment scales rely on recall: obtaining data through completing scales is subject to high recall bias, leading to distorted assessment results; contact monitoring devices are uncomfortable: using wearable devices such as chest patches and EEG patches to collect physiological data can interfere with nighttime sleep and affect data accuracy; and assessment data is fragmented: testing based on outpatient time points cannot capture the dynamic evolution of the disease and is difficult to reflect long-term changes in psychological status.

[0003] To overcome these limitations, the industry has explored two technical approaches: wearable physiological monitoring systems, which capture EEG signals through optical fibers. However, the results are only instantaneous data and require correction based on doctor's interview materials, making continuous monitoring impossible. Video behavioral analysis solutions, which use AI to analyze facial micro-expressions to assess psychological states. However, due to privacy concerns, their acceptance in home settings is low.

[0004] Based on this, there is an urgent need for a non-contact, non-interference and continuously monitored psychological risk assessment technology to achieve objective and dynamic psychological state assessment. Summary of the Invention

[0005] In order to overcome the above-mentioned technical problems existing in the prior art, an embodiment of the present invention provides a dynamic psychological risk assessment system and method based on sleep HRV characteristics, aiming to solve the problems existing in the existing psychological assessment methods, such as large subjective bias, interference of contact devices with sleep, and data fragmentation. By non-contact collection of BCG signals during sleep and extraction of HRV features based on BCG signals, dynamic assessment of psychological risks can be achieved.

[0006] In order to achieve the above-mentioned purpose, an embodiment of the present invention provides a dynamic psychological risk assessment system based on sleep HRV characteristics, including: a signal acquisition module, the signal acquisition module integrating a piezoelectric film sensor, a signal amplification circuit, an MCU and a data transmission module, the piezoelectric film sensor is placed under the mattress, and the signal acquisition module is used to collect BCG signals; a data processing module, the data processing module is used to process the BCG signals and convert the BCG signals into HRV characteristic values; a risk assessment module, based on the HRV characteristic values, and generating dynamic psychological risk assessment results through an HRV-emotion index mapping model; a user interaction module, displaying the dynamic psychological risk assessment results through a human-computer interaction device.

[0007] Preferably, the data processing module is used to process the BCG signal, specifically including: the processing steps are sampling conversion, preprocessing, calculation and HRV characteristic value conversion in sequence; the preprocessing includes filtering processing, detrending processing and short-time energy envelope processing; the calculation includes heart rate calculation and heart beat interval estimation comparison.

[0008] Preferably, the HRV-emotion index mapping model satisfies the following requirements: inputting HRV feature values ​​and outputting evaluation indicators of associated parasympathetic nervous system states; acquiring historical case data, and the model performing training verification through the historical case data.

[0009] Correspondingly, the present invention also provides a method for dynamic psychological risk assessment based on sleep HRV characteristics, which is applied to a dynamic psychological risk assessment system based on sleep HRV characteristics provided by the present invention, including the following steps: S1: collecting BCG signals during sleep through a piezoelectric film sensor placed under the mattress; S2: performing processing operations on the BCG signals to obtain processed BCG signals; S3: calculating HRV characteristic values ​​based on the processed BCG signals; S4: inputting the HRV characteristic values ​​into a preset HRV-emotion index mapping model to generate dynamic psychological risk assessment results.

[0010] Preferably, the S2 specifically includes: performing data conversion on the BCG signal through an analog-to-digital converter, and obtaining a physical signal value of the BCG signal after voltage conversion according to formula (1), wherein the formula (1) is: (1), Wherein, val is the physical signal value, ranging from [-refValue / 2, +refValue / 2], adcCode is the analog-to-digital converter sampling value, and refValue is the reference voltage; the physical signal value is preprocessed, and the preprocessing specifically includes: using an IIR high-pass filter with a cutoff frequency of 1Hz to filter out low-frequency noise; using an IIR low-pass filter with a cutoff frequency of 20Hz to filter out high-frequency noise; eliminating baseline drift through detrending processing, with a trend term length of 25, and the processed data is used as the data source for accurate heart rate estimation; the short-time energy envelope of the downsampled data is calculated with a window length of 150 and a step of 10, and the sampling rate is reduced to 60Hz, which is used as the data source for FFT pre-estimation of the heart rate.

[0011] Preferably, the S3 specifically includes: calculating the heart rate, SDNN, RMSSD, LF / HF, and SD1 / SD2 eigenvalues ​​based on the processed BCG signal using calculation formulas, wherein the calculation formulas for each eigenvalue are as follows: the heart rate eigenvalue is calculated using formula (2), wherein formula (2) is specifically: (2), where RR interval is the interval between adjacent heart beats; The SDNN eigenvalue is calculated using formula (3), which is specifically: (3), Where, NNi is the ith normal sinus interval, is the average value of all NN intervals, and N is the total number of NN intervals; The RMSSD eigenvalue is calculated using formula (4), which is specifically: (4), Where, NNi is the i-th normal sinus beat interval, and N is the total number of NN intervals; Calculate the SD1 eigenvalue and SD2 eigenvalue, specifically: sd=diff(NNi); rr=NNi; SD1=sqrt( 0.5*std(sd)^2 ); SD2=sqrt( 2*(std(rr)^2)-(0.5*std(sd)^2) ); Where NNi is the i-th normal sinus beat interval, and std is the standard deviation function.

[0012] Preferably, the calculation step of the LF / HF eigenvalue is as follows: using the Lomb-Scargle algorithm to calculate the power spectrum density by fitting a sine wave model, the calculation formula is: , Where xi is the i-th signal sampling value, ti is the non-uniform sampling time point, σ2 is the signal variance, and τ is the time shift factor; For frequency integration, the low-frequency (LF) power integration frequency band is 0.04–0.15 Hz, and the high-frequency (HF) power integration frequency band is 0.15–0.4 Hz; The LF eigenvalue and the HF eigenvalue are calculated using formula (6), where formula (5) is specifically: (5).

[0013] Preferably, the heart rate interval calculation for calculating the heart rate characteristic value in S3 includes: pre-estimating the heart rate range through FFT based on the downsampled data, and using the pre-estimated peak point position ±10 as the similarity calculation range; fusing the signal amplitude, absolute amplitude difference and autocorrelation within a preset limited range, and calculating the maximum similarity position as the real-time heart rate; verifying the accuracy based on the correlation coefficient between the heart rate interval sequence and the ECG data, and the correlation coefficient calculation formula is: , Wherein, HBecg is the HRV sequence calculated based on ECG, HBbcg is the HRV sequence calculated based on BCG, and they are the means of data x and y, respectively. N is the data length.

[0014] The technical solution provided by the present invention has at least the following technical effects: BCG signals are collected without any contact through sensors under the mattress, avoiding interference of contact equipment on the sleep of the person being assessed, ensuring that the data reflects the natural physiological state, and adopting non-contact collection to improve comfort and data accuracy; it can continuously monitor dynamically and capture long-term changes, and can continuously collect data throughout the night and for a long time, breaking through the limitations of fragmented outpatient assessments and dynamically tracking the evolution of psychological states; data accuracy is high: the correlation coefficient between the heart rate interval calculated based on BCG and the ECG is very high, and the average error is very small; the HRV characteristic values ​​(HR, SDNN, RMSSD, etc.) are highly consistent with the ECG results and meet clinical standards; it is highly targeted and adapts to the needs of psychological assessment. The HRV characteristic values ​​can effectively reflect the emotional state and provide an objective basis for the initial screening of psychological risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings: Figure 1 is a structural diagram of a psychological risk dynamic assessment system based on sleep HRV characteristics provided by an embodiment of the present invention; Figure 2 Schematic diagram of the structure of the data acquisition module in an embodiment of the present invention; Figure 3 is a data processing flow chart in an embodiment of the present invention; Figure 4 is a flow chart of a method for dynamic assessment of psychological risk based on sleep HRV characteristics provided by an embodiment of the present invention; Figure 5 1. This is a graph showing ECG and BCG heart beat interval estimation and deviation experimental data in an embodiment of the present invention; Figure 6 Schematic diagram of calculating HR feature consistency using the method provided by the present invention; Figure 7 Schematic diagram of calculating SDNN feature consistency using the method provided by the present invention; Figure 8 Schematic diagram of calculating RMSSD feature consistency using the method provided by the present invention; Figure 9Schematic diagram of calculating LF / HF feature consistency using the method provided by the present invention; Figure 10 This is a schematic diagram of calculating SD1 / SD2 feature consistency using the method provided by the present invention. DETAILED DESCRIPTION

[0016] The following describes the specific implementation of the embodiment of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present invention and is not used to limit the embodiment of the present invention.

[0017] The terms "system" and "network" in the embodiments of the present invention can be used interchangeably. "Multiple" refers to two or more. In view of this, "multiple" can also be understood as "at least two" in the embodiments of the present invention. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / ", unless otherwise specified, generally indicates that the previous and next associated objects are in an "or" relationship. In addition, it should be understood that in the description of the embodiments of the present invention, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0018] See Figure 1 The embodiment of the present invention provides a psychological risk dynamic assessment system based on sleep HRV characteristics, including a signal acquisition module, a data processing module, a risk assessment module and a user interaction module, each module collaboratively realizes the signal acquisition, processing and assessment functions; wherein, Figure 2 As shown, the signal acquisition module is used to collect BCG signals. The signal acquisition module integrates a piezoelectric film sensor, a signal amplification circuit, an MCU, and a data transmission module. The piezoelectric film sensor is placed under the mattress to collect tiny body vibration signals caused by cardiac ejection. The signal amplification circuit performs multi-stage amplification on the original vibration signal. The MCU obtains ADC sampling data through SPI. The data transmission module transmits the processed signal to the background server. The data transmission module can use a WiFi module for wireless transmission.

[0019] In the embodiment of the present invention, the data processing module is used to process the BCG signal and convert the BCG signal into the HRV characteristic value. Specifically, the data processing module receives the BCG signal transmitted by the signal acquisition module and executes the data processing process, such as Figure 3As shown, the steps of the processing flow are sampling conversion, preprocessing, calculation and HRV characteristic value conversion. The preprocessing includes filtering processing, detrending processing and short-time energy envelope processing. The calculation includes heart rate calculation and heart beat interval estimation comparison; the risk assessment module is based on the HRV characteristic value obtained by processing and generates a dynamic psychological risk assessment result through the HRV-emotion index mapping model. The HRV-emotion index mapping model meets the following requirements: input HRV characteristic value, output associated parasympathetic nerve state assessment index; obtain historical case data, and the model performs training verification through historical case data; the user interaction module communicates with the background server, and is used to display the dynamic psychological risk assessment result through the human-computer interaction device, receive and display HRV characteristic value and psychological risk assessment result for guardians or medical staff to view.

[0020] Based on the same inventive concept, Figure 4 As shown, an embodiment of the present invention provides a method for dynamic assessment of psychological risk based on sleep HRV characteristics. The method is implemented based on a system for dynamic assessment of psychological risk based on sleep HRV characteristics provided by an embodiment of the present invention, and includes the following steps: S1: BCG signals are collected during sleep using a piezoelectric film sensor placed under the mattress; S2: performing a processing operation on the BCG signal to obtain a processed BCG signal; S3: Calculating HRV characteristic values ​​based on the processed BCG signal; S4: Inputting the HRV characteristic value into a preset HRV-emotion index mapping model to generate a dynamic psychological risk assessment result.

[0021] In the embodiment of the present invention, step S1 specifically involves collecting the BCG signal of the person undergoing psychological assessment while sleeping through a piezoelectric film sensor placed under the mattress, transmitting the signal to the backend server after circuit processing, and then executing S2; S2 specifically involves the backend server performing processing operations on the BCG signal and obtaining a processed BCG signal; the processing operations specifically include sampling conversion, preprocessing, calculation, and HRV characteristic value conversion, and the sampling conversion specifically involves performing data conversion on the BCG signal through an analog-to-digital converter. Specifically, a 24-bit AD7190 can be used as the ADC sampling input, and the sampling frequency is configured to be 600 Hz. After the ADC successfully converts the sampled data, an interrupt is generated, and the MCU obtains the sampled data through the SPI method, and obtains the physical signal value of the BCG signal after voltage conversion according to formula (1). Formula (1) is: (1), Among them, val is the physical signal value, ranging from [-refValue / 2, +refValue / 2], adcCode is the analog-to-digital converter sampling value, and refValue is the reference voltage; the physical signal value is preprocessed, and the preprocessing specifically includes: using an IIR high-pass filter with a cutoff frequency of 1Hz to filter out low-frequency noise; using an IIR low-pass filter with a cutoff frequency of 20Hz to filter out high-frequency noise; performing detrending processing to eliminate baseline drift, with a trend term length of 25, and the processed data is used as the data source for accurate heart rate estimation; performing short-time energy envelope processing, calculating the short-time energy envelope of the downsampled data with a window length of 150 and a step of 10, and reducing the sampling rate to 60Hz, as the data source for FFT pre-estimation of the heart rate; then executing S3.

[0022] In the embodiment of the present invention, S3 specifically includes: calculating the heart rate (hereinafter referred to as HR), SDNN, RMSSD, LF / HF, and SD1 / SD2 eigenvalues ​​based on the processed BCG signal using calculation formulas. The calculation formulas for each eigenvalue are as follows: The heart rate characteristic value is calculated using formula (2), which is: (2), where RR interval is the interval between adjacent heart beats; The SDNN eigenvalue is calculated using formula (3), which is: (3), Where, NNi is the ith normal sinus interval, is the average value of all NN intervals, and N is the total number of NN intervals; The RMSSD eigenvalue is calculated using formula (4), which is: (4), Where, NNi is the i-th normal sinus beat interval, and N is the total number of NN intervals; Use the following formula to calculate the SD1 eigenvalue and SD2 eigenvalue. The calculation formula is as follows: sd=diff(NNi); rr=NNi; SD1=sqrt( 0.5*std(sd)^2 ); SD2=sqrt( 2*(std(rr)^2)-(0.5*std(sd)^2) ); Where NNi is the i-th normal sinus beat interval, and std is the standard deviation function.

[0023] In the embodiment of the present invention, the steps for calculating the LF / HF eigenvalue in S3 are as follows: First, the power spectral density is calculated by fitting a sine wave model using the Lomb-Scargle algorithm, and the calculation formula is: Where xi is the i-th signal sampling value, ti is the non-uniform sampling time point, σ2 is the signal variance, and τ is the time shift factor; The second step is to integrate the frequency band. The low-frequency (LF) power integration band is 0.04-0.15Hz, and the high-frequency (HF) power integration band is 0.15-0.4Hz. Formula (5) is used to calculate the LF eigenvalue and HF eigenvalue. Formula (5) is specifically: (5).

[0024] In the embodiment of the present invention, the minimum window length for heart rate calculation is specifically as follows: the heart rate calculation range is 40 to 180 bmp, the sampling rate is 60 Hz during pre-estimation of the heart rate, and the lower limit of the heart rate is 40 bmp. Therefore, it can be seen that the minimum window length for pre-estimation is 180. If the sampling rate is 600 Hz during detailed estimation, it can be seen that the minimum window length is 1800. The heart rate interval calculation for calculating the heart rate characteristic value in S3 specifically includes: synchronously collecting ECG data, calculating HRV sequences based on BCG signals and ECG data, performing FFT operation on the down-sampled data, and obtaining the pre-estimated heart rate. The range is based on the range corresponding to the maximum FFT peak point. The corresponding range can be used as the similarity calculation range of the pre-estimated peak point position ± 10. The peak point position is converted to the corresponding position of the maximum similarity in the detailed estimation as the approximate range for searching similarity. The signal amplitude, absolute amplitude difference, and autocorrelation are integrated within the preset limited range to calculate the maximum similarity position as the real-time heart rate. For example, the maximum peak point is used as the anchor point within the current calculation window. When the anchor point leaves the calculation window, it is a heartbeat process. The average of all calculation results of the current anchor point is calculated as the real-time heart rate output. Furthermore, heartbeat interval estimation and comparison are performed. Based on the collected BCG signal, heartbeat detection and heart rate and HRV are calculated. At the same time, the algorithm is analyzed and verified based on the ECG HRV data. For example, the subject lies quietly on the mattress, and the piezoelectric film sensor under the mattress is turned on to collect BCG signals and ECG data simultaneously. The two systems simultaneously record and store data, measure the subject's heartbeat-related HRV data, and the test lasts for 5 minutes. Finally, the corresponding data are compared and analyzed. Specifically, the accuracy is verified based on the correlation coefficient between the heartbeat interval sequence and ECG data. The correlation coefficient calculation formula is: , Among them, HBecg is the HRV sequence calculated based on ECG, HBbcg is the HRV sequence calculated based on BCG, and they are the mean of data x and y respectively. N is the data length. The interval of Corr is [-1,1]. Figure 5 As shown, the correlation of the average HRV sequence based on the collected data is above 0.98, and the average error obtained by BCG and ECG is within 2ms. Therefore, the heart beat interval obtained based on BCG and the heart beat interval obtained based on ECG calculation are consistent, with no significant difference.

[0025] In one embodiment, Figures 6 to 10 As shown in the figure, based on the relationship between BCG signal and ECG HRV calculation, four characteristic parameters, HR, SDNN, RMSSD, SD1 / SD2 and LF / HF, were calculated respectively. The Bland-Altman method was used to perform consistency test on the accuracy of the calculated HRV characteristics. Figures 6-10 In the figure, the horizontal axis represents the average value of HRV calculated by ECG and the HRV value calculated by BCG, and the vertical axis represents the difference in characteristic values. The consistency of HRV characteristic values ​​is 95%, and the measurement results are accurate.

[0026] In an embodiment of the present invention, the HRV characteristic value obtained through S3 is input into a preset HRV-emotion index mapping model to generate a dynamic psychological risk assessment result to achieve dynamic assessment of psychological risk. The user interaction module pushes a dynamic assessment report to the user, and the dynamic assessment report includes an HRV characteristic value trend graph and a psychological risk level.

[0027] By means of the dynamic psychological risk assessment system and method based on sleep HRV characteristics provided by the present invention, HRV characteristic values ​​are calculated based on BCG signals, and compared with HRV results obtained based on electrocardiogram (ECG). The heartbeat intervals calculated based on BCG are highly correlated with those calculated based on electrocardiogram (ECG). The heartbeat intervals based on BCG can effectively reflect changes in heartbeats. The relevant features HR, SDNN, RMSSD, LF / HF, and SD1 / SD2 calculated based on BCG signals and ECG signals are significantly consistent and fall within 95% of the effective interval. Therefore, the relevant features calculated based on BCG are consistent with those calculated based on ECG. The method provided based on BCG is convenient and simple to operate, and can be used for continuous monitoring of HRV. It has significant application prospects for daily, long-term continuous HRV monitoring.

[0028] The beneficial effects of the present invention are as follows: BCG signals are collected without feeling through sensors under the mattress, avoiding interference of contact equipment on the sleep of the psychological assessee, ensuring that the data reflects the natural physiological state, and adopting non-contact collection to improve comfort and data accuracy; continuous dynamic monitoring can be carried out to capture long-term change patterns, and data can be collected continuously throughout the night and for a long time, breaking through the limitations of fragmented outpatient assessments, and can dynamically track the evolution of the assessee's psychological state; data accuracy is high: the correlation coefficient between the heart rate interval calculated based on BCG and ECG is very high, and the average error is very small; the HRV characteristic values ​​(HR, SDNN, RMSSD, etc.) are highly consistent with the ECG results, which meet clinical standards; it is highly targeted and adapts to the needs of psychological assessment. The HRV characteristic values ​​can effectively reflect the emotional state and provide an objective basis for the initial screening of psychological risks.

[0029] The above describes in detail the optional implementation methods of the embodiments of the present invention in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above implementation methods. Within the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the scope of protection of the embodiments of the present invention.

[0030] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0031] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program. The program is stored in a storage medium and includes a number of instructions for causing a single-chip microcomputer, chip, or processor to execute all or part of the steps in the methods described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0032] In addition, various implementations of the embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A psychological risk dynamic assessment system based on sleep HRV characteristics, characterized by: include: A signal acquisition module, which integrates a piezoelectric film sensor, a signal amplification circuit, an MCU, and a data transmission module. The piezoelectric film sensor is placed under the mattress. The signal acquisition module is used to collect BCG signals; a data processing module, configured to process the BCG signal and convert the BCG signal into an HRV characteristic value; A risk assessment module generates a dynamic psychological risk assessment result based on the HRV characteristic value and through an HRV-emotion index mapping model; The user interaction module displays the dynamic psychological risk assessment results through a human-computer interaction device.

2. The system according to claim 1, wherein: The data processing module is used to process the BCG signal, specifically including: The processing steps are sampling conversion, preprocessing, calculation and HRV characteristic value conversion; The preprocessing includes filtering, detrending and short-time energy envelope processing; The calculation includes heart rate calculation and heart beat interval estimation comparison.

3. The system according to claim 1, wherein: The HRV-emotion index mapping model satisfies: Input HRV characteristic value and output the evaluation index of the associated parasympathetic nervous system state; Historical case data is obtained, and the model is trained and verified using the historical case data.

4. A method for dynamic assessment of psychological risk based on sleep HRV characteristics, applied to a system for dynamic assessment of psychological risk based on sleep HRV characteristics according to any one of claims 1 to 3, characterized in that: The steps include: S1: BCG signals are collected during sleep using a piezoelectric film sensor placed under the mattress; S2: performing a processing operation on the BCG signal to obtain a processed BCG signal; S3: Calculating HRV characteristic values ​​based on the processed BCG signal; S4: Inputting the HRV characteristic value into a preset HRV-emotion index mapping model to generate a dynamic psychological risk assessment result.

5. The method according to claim 4, characterized in that The S2 specifically includes: The BCG signal is converted into data by an analog-to-digital converter, and the physical signal value of the BCG signal after voltage conversion is obtained according to formula (1), and the formula (1) is: （1）， Where val is the physical signal value, ranging from [-refValue / 2, +refValue / 2], adcCode is the analog-to-digital converter sampling value, and refValue is the reference voltage; Preprocessing the physical signal value, wherein the preprocessing specifically includes: An IIR high-pass filter with a cutoff frequency of 1 Hz is used to filter out low-frequency noise; An IIR low-pass filter with a cutoff frequency of 20 Hz is used to filter out high-frequency noise; Baseline drift was eliminated by detrending, with a trend term length of 25. The processed data served as the data source for accurate heart rate estimation; The short-time energy envelope of the downsampled data was calculated with a window length of 150 and a step of 10, and the sampling rate was reduced to 60 Hz, which was used as the data source for the heart rate FFT pre-estimation.

6. The method according to claim 5, characterized in that The S3 specifically includes: Based on the processed BCG signal, the heart rate, SDNN, RMSSD, LF / HF, and SD1 / SD2 eigenvalues ​​are calculated using the following calculation formulas: The heart rate characteristic value is calculated using formula (2), which is specifically: （2）， Among them, RR interval is the interval between adjacent heart beats; The SDNN eigenvalue is calculated using formula (3), which is specifically: （3）， Where, NNi is the ith normal sinus interval, is the average value of all NN intervals, and N is the total number of NN intervals; The RMSSD eigenvalue is calculated using formula (4), which is specifically: （4）， Where, NNi is the i-th normal sinus beat interval, and N is the total number of NN intervals; Calculate the SD1 eigenvalue and SD2 eigenvalue, specifically: sd=diff(NNi); rr=NNi; SD1=sqrt( 0.5*std(sd)^2 ); SD2=sqrt( 2*(std(rr)^2)-(0.5*std(sd)^2) ); Where NNi is the i-th normal sinus beat interval, and std is the standard deviation function.

7. The method according to claim 6, characterized in that The calculation steps of the LF / HF eigenvalue are as follows: The power spectral density is calculated using the Lomb-Scargle algorithm by fitting a sine wave model. The calculation formula is: ， Where xi is the i-th signal sampling value, ti is the non-uniform sampling time point, σ2 is the signal variance, and τ is the time shift factor; For frequency integration, the low-frequency (LF) power integration frequency band is 0.04–0.15 Hz, and the high-frequency (HF) power integration frequency band is 0.15–0.4 Hz; The LF eigenvalue and the HF eigenvalue are calculated using formula (5), which is specifically: （5）。 8. The method according to claim 6, characterized in that The heartbeat interval calculation of the heart rate characteristic value in S3 includes: Based on the downsampled data, the heart rate range is estimated by FFT, and the estimated peak point position ±10 is used as the similarity calculation range; The signal amplitude, absolute amplitude difference and autocorrelation are integrated within a preset range, and the maximum similarity position is calculated as the real-time heart rate; The accuracy is verified based on the correlation coefficient between the heartbeat interval sequence and ECG data. The correlation coefficient calculation formula is: ， Wherein, HBecg is the HRV sequence calculated based on ECG, HBbcg is the HRV sequence calculated based on BCG, and they are the means of data x and y, respectively. N is the data length.

Images