An anti-interference filtering method and system of a multi-modal physiological signal collection device, a terminal and a storage medium

Abstract

The present application relates to physiological signal processing technical field, disclose a kind of multi-modal physiological signal acquisition equipment's anti-interference filter method, system, terminal and storage medium, comprising: the multi-modal physiological signal of acquisition is preprocessed;Modal identification and feature extraction are carried out to the multi-modal physiological signal after pre-processing;According to modal identification result, corresponding filter is selected, and filter parameter is dynamically adjusted;Filtering path and order are adjusted based on cascaded filter structure and dynamic switching mechanism;Based on crosstalk processing and reference channel compensation mechanism, adaptive interference estimation and offset are carried out;According to the extracted feature, parallel filtering is carried out, and the evaluation parameter of each filtering result is compared based on output result selection mechanism, and according to the comparison result, the final signal is selected and output.The present application realizes the effective suppression of power interference, motion artifact, crosstalk and multi-source interference such as high-frequency noise, improves the real-time and stability of filtering effect.

Description

Technical Field

[0001] This invention relates to the field of physiological signal processing technology, and in particular to an anti-interference filtering method, system, terminal, and storage medium for a multimodal physiological signal acquisition device. Background Technology

[0002] With the rapid development of wearable devices and telemedicine, physiological signal acquisition equipment has gradually evolved from single-modality to multi-modality. Common physiological signals include electrocardiogram (ECG), electroencephalogram (EEG), electrical activity of the skin (EDA), electromyography (EMG), and blood oxygen saturation (SpO2). However, these physiological signals are easily affected by various interferences during acquisition (e.g., power frequency noise, motion artifacts, electromyographic noise, environmental electromagnetic interference, and crosstalk between signals). Therefore, specific filtering methods are required to process these physiological signals after acquisition.

[0003] Existing filtering methods, such as FIR filters (finite-length impulse response filters) / IIR filters (infinite-impulse response digital filters), wavelet transforms, adaptive filters (e.g., LMS filters, least mean square filters), and blind source separation (ICA methods, independent component analysis denoising methods), while effective in certain specific cases, have the following shortcomings in multimodal signal processing:

[0004] 1) Insufficient applicability: It is difficult to apply to multiple signal modes at the same time, resulting in limited filtering effect in multimodal scenarios.

[0005] 2) Poor real-time performance: Some filtering algorithms are computationally complex and difficult to meet real-time processing requirements, especially in resource-constrained wearable devices.

[0006] 3) Limited filtering effect: In dynamic or high-interference environments, the filtering quality is difficult to guarantee.

[0007] Therefore, existing technologies still need improvement. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide an anti-interference filtering method, system, terminal and storage medium for a multimodal physiological signal acquisition device, in order to solve the problems of poor filtering effect and low efficiency of existing filtering methods in multimodal signal scenarios.

[0009] The technical solution adopted by this invention to solve the technical problem is as follows:

[0010] In a first aspect, the present invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising:

[0011] Preprocessing of the acquired multimodal physiological signals;

[0012] Modality recognition and feature extraction are performed on the preprocessed multimodal physiological signals;

[0013] Select the corresponding filter based on the modality recognition results, and dynamically adjust the filter parameters;

[0014] Adjusting the filtering path and sequence based on a cascaded filtering structure and dynamic switching mechanism;

[0015] Adaptive interference estimation and cancellation are based on crosstalk processing and reference channel compensation mechanisms.

[0016] Parallel filtering is performed based on the extracted features. The evaluation parameters of each filtering result are compared based on the output result selection mechanism, and the final signal is selected and output based on the comparison results.

[0017] In one implementation, the preprocessing of the acquired multimodal physiological signals includes:

[0018] The amplitude of the multimodal physiological signal is normalized, DC bias is removed by mean removal and high-pass filter method, and outliers are removed by standard fraction method, interquartile range method and sliding window method to obtain the preprocessed multimodal physiological signal.

[0019] In one implementation, the modality recognition and feature extraction of the preprocessed multimodal physiological signal includes:

[0020] Identify the signal category corresponding to each modality signal in the preprocessed multimodal physiological signal;

[0021] The time-domain signals of each modal signal are converted into frequency-domain signals by using spectrum analysis, and the frequency component with the largest amplitude is identified to obtain the corresponding dominant frequency characteristics.

[0022] The corresponding waveform energy characteristics are obtained by summing the squared amplitudes of each sampling point of each modal signal;

[0023] Calculate the difference between adjacent sampling points of each modal signal, count the number of mutation points exceeding a set threshold, and calculate the ratio of the number of mutation points to the total number of sampling points to obtain the corresponding mutation rate feature.

[0024] Calculate the ratio of signal power to noise power for each modal signal to obtain the corresponding signal-to-noise ratio characteristics.

[0025] In one implementation, the step of selecting the corresponding filter based on the modality recognition result and dynamically adjusting the filter parameters includes:

[0026] Select the corresponding filter combination based on the identified signal mode, and dynamically adjust the filter parameters in the combination;

[0027] The filter combination includes any combination of band-stop filters, wavelet denoising, high-pass filters, adaptive filters, independent component analysis denoising, low-pass filters, notch filters, and Kalman filters.

[0028] In one implementation, the adjustment of the filtering path and order based on the cascaded filtering structure and dynamic switching mechanism includes:

[0029] The signal-to-noise ratio (SNR) of each modal signal is obtained in real time, and a path that meets the current noise environment is selected from the preset filtering paths based on the SNR detection results.

[0030] The filter order and filter parameters in the filter combination are dynamically adjusted based on the current noise characteristics of each modal signal.

[0031] In one implementation, the adaptive interference estimation and cancellation based on crosstalk processing and reference channel compensation mechanism includes:

[0032] A reference channel is constructed, and the interference signal in the main channel is estimated using the signal corresponding to the reference channel. The estimated value corresponding to the interference signal is then subtracted from the signal corresponding to the main channel to cancel out the interference signal.

[0033] In one implementation, the step of performing parallel filtering based on the extracted features, comparing the evaluation parameters of each filtering result based on an output result selection mechanism, and selecting and outputting the final signal based on the comparison results includes:

[0034] Parallel filtering is performed based on the characteristics of each modal signal to generate the corresponding filtering results;

[0035] Based on a preset strategy or algorithm, the evaluation parameters of each filtering result are compared, and the result that meets the current signal characteristics is selected as the final signal for output.

[0036] Secondly, the present invention provides an anti-interference filtering system for a multimodal physiological signal acquisition device, comprising:

[0037] The signal preprocessing module is used to preprocess the acquired multimodal physiological signals;

[0038] The modality recognition and feature extraction module is used to perform modality recognition and feature extraction on preprocessed multimodal physiological signals;

[0039] The adaptive filter module is used to select the corresponding filter based on the modality recognition results and dynamically adjust the filter parameters;

[0040] The interference detection and dynamic adjustment module is used to adjust the filtering path and sequence based on the cascaded filtering structure and dynamic switching mechanism.

[0041] The crosstalk suppression and signal fusion module is used for adaptive interference estimation and cancellation based on crosstalk processing and reference channel compensation mechanisms.

[0042] The output selection and data transmission module is used to perform parallel filtering based on the extracted features, compare the evaluation parameters of each filtering result based on the output result selection mechanism, and select and output the final signal based on the comparison results.

[0043] Thirdly, the present invention provides a terminal, comprising: a processor and a memory, wherein the memory stores an anti-interference filtering program for a multimodal physiological signal acquisition device, and the anti-interference filtering program for the multimodal physiological signal acquisition device, when executed by the processor, is used to implement the operation of the anti-interference filtering method for the multimodal physiological signal acquisition device as described in the first aspect.

[0044] Fourthly, the present invention also provides a computer-readable storage medium storing an anti-interference filtering program for a multimodal physiological signal acquisition device, wherein the anti-interference filtering program for the multimodal physiological signal acquisition device, when executed by a processor, is used to implement the operation of the anti-interference filtering method for the multimodal physiological signal acquisition device as described in the first aspect.

[0045] The present invention, by employing the above technical solution, has the following effects:

[0046] This invention proposes an anti-interference filtering method for multimodal physiological signal acquisition equipment. Based on the spectral characteristics, noise types, and processing requirements of different physiological signal modes, a modality-adaptive filtering mechanism is constructed. Combined with a multi-stage filter cascade structure, it effectively suppresses multi-source interference such as power supply interference, motion artifacts, crosstalk, and high-frequency noise. Through interference identification and dynamic parameter adjustment mechanisms, the real-time performance and stability of the filtering effect are improved. This invention features good versatility, stability, and low computational resource consumption. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0048] Figure 1 This is a flowchart of the anti-interference filtering method for the multimodal physiological signal acquisition device in this invention.

[0049] Figure 2 This is a structural block diagram of the anti-interference filtering system of the multimodal physiological signal acquisition device in this invention.

[0050] Figure 3 This is a functional schematic diagram of the terminal in one implementation of the present invention.

[0051] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0053] Exemplary methods

[0054] Existing anti-interference filtering methods for single-modal physiological signal acquisition devices, such as FIR filters (finite-length impulse response filters) / IIR filters (infinite-impulse response digital filters), wavelet transforms, adaptive filters (e.g., LMS filters, least mean square filters), and blind source separation (ICA analysis methods, independent component analysis denoising methods), while effective in certain specific situations, have the following shortcomings in multimodal signal processing:

[0055] 1) Insufficient applicability: It is difficult to apply to multiple signal modes at the same time, resulting in limited filtering effect in multimodal scenarios.

[0056] 2) Poor real-time performance: Some filtering algorithms are computationally complex and difficult to meet real-time processing requirements, especially in resource-constrained wearable devices.

[0057] 3) Limited filtering effect: In dynamic or high-interference environments, the filtering quality is difficult to guarantee.

[0058] To address the above technical problems, this invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device. The method includes: preprocessing the acquired multimodal physiological signals; performing modality recognition and feature extraction on the preprocessed multimodal physiological signals; selecting a corresponding filter based on the modality recognition results and dynamically adjusting the filter parameters; adjusting the filtering path and order based on a cascaded filtering structure and a dynamic switching mechanism; performing adaptive interference estimation and cancellation based on crosstalk processing and a reference channel compensation mechanism; performing parallel filtering based on the extracted features; comparing the evaluation parameters of each filtering result based on an output result selection mechanism; and selecting and outputting the final signal based on the comparison results.

[0059] This invention can construct a modality-adaptive filtering mechanism tailored to the spectral characteristics, noise types, and processing requirements of different physiological signal modalities. Combined with a multi-stage filter cascade structure, it effectively suppresses multi-source interference such as power supply interference, motion artifacts, crosstalk, and high-frequency noise. Through interference identification and dynamic parameter adjustment mechanisms, the real-time performance and stability of the filtering effect are improved. It features good versatility, stability, and low computational resource consumption.

[0060] like Figure 1 As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0061] Step S100: Preprocess the acquired multimodal physiological signals.

[0062] Specifically, in one implementation of this embodiment, step S100 includes the following steps:

[0063] Step S101: The amplitude of the multimodal physiological signal is normalized, DC bias is removed by mean removal and high-pass filter method, and outliers are removed by standard fraction method, interquartile range method and sliding window method to obtain the preprocessed multimodal physiological signal.

[0064] In this embodiment, the anti-interference filtering method of the multimodal physiological signal acquisition device is implemented through the anti-interference filtering system of the multimodal physiological signal acquisition device.

[0065] like Figure 2 As shown, the anti-interference filtering system of the multimodal physiological signal acquisition device includes: a signal preprocessing module, a modality recognition and feature extraction module, an adaptive filter module, an interference detection and dynamic adjustment module, a crosstalk suppression and signal fusion module, and an output selection and data transmission module.

[0066] In this embodiment, physiological signals of the corresponding modalities are first collected by sensors such as ECG (electrocardiogram sensor), EMG (electromyogram sensor), EEG (electroencephalogram sensor), EDA (electrodermal response) sensor, and blood oxygen saturation sensor, thereby obtaining the multimodal physiological signals.

[0067] After obtaining the multimodal physiological signals, the raw signals of each modality are input into the signal preprocessing module described above for data preprocessing. The specific preprocessing process is as follows:

[0068] The input raw signal is normalized in amplitude, DC bias is removed by mean removal and high-pass filter method, and outliers in the raw signal of each mode are removed by Z-Score method (data standardization and outlier detection method), IQR method (statistical method for identifying outliers based on interquartile range) and sliding window method.

[0069] In this embodiment, the preprocessed multimodal physiological signals can be obtained by performing data preprocessing in the above manner. Of course, in other application scenarios, data preprocessing can also be performed by polynomial fitting methods, signal segmentation methods, time-domain superposition methods, and noise injection methods.

[0070] like Figure 1 As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0071] Step S200: Modality recognition and feature extraction are performed on the preprocessed multimodal physiological signals.

[0072] In this embodiment, after obtaining the preprocessed multimodal physiological signals, they are input into the modality recognition and feature extraction module described above. In this module, the signal type of each modality can be automatically identified, and the features of each signal can be extracted through the corresponding feature extraction method.

[0073] Specifically, in one implementation of this embodiment, step S200 includes the following steps:

[0074] Step S201: Identify the signal category corresponding to each modality signal in the preprocessed multimodal physiological signal;

[0075] Step S202: Convert the time-domain signal of each modal signal into a frequency-domain signal using a spectrum analysis method, identify the frequency component with the largest amplitude, and obtain the corresponding main frequency characteristics;

[0076] Step S203: Sum the squared amplitudes of each sampling point of each modal signal to obtain the corresponding waveform energy characteristics;

[0077] Step S204: Calculate the difference between adjacent sampling points of each modal signal, count the number of mutation points exceeding the set threshold, and calculate the ratio of the number of mutation points to the total number of sampling points to obtain the corresponding mutation rate feature.

[0078] Step S205: Calculate the ratio of signal power to noise power for each modal signal to obtain the corresponding signal-to-noise ratio characteristics.

[0079] In this embodiment, after automatically identifying the signal type (such as ECG signal, EEG signal, etc.), the corresponding common method is selected to extract the features that best characterize the signal based on the identified signal type. The extracted signal features include, but are not limited to: main frequency features, waveform energy features, mutation rate features, and signal-to-noise ratio features.

[0080] As an example, features of each modal signal can be extracted in this embodiment in the following way:

[0081] 1) By using spectral analysis methods, such as the Fast Fourier Transform (FFT) method, the time-domain signal is converted into a frequency-domain signal, and the frequency component with the largest amplitude is identified, thus obtaining the main frequency characteristics.

[0082] 2) The total energy of the signal is obtained by summing the squared amplitudes of each sampling point of the signal, and this is used as the waveform energy feature.

[0083] 3) By calculating the difference between adjacent sampling points, the number of mutation points exceeding the set threshold is counted, and the ratio of this number to the total number of sampling points is used as the mutation rate feature.

[0084] 4) The signal-to-noise ratio characteristics are obtained by calculating the ratio of signal power to noise power.

[0085] In this embodiment, in addition to performing modality recognition and feature extraction on the preprocessed multimodal physiological signals in the above manner, machine learning models can also be used to replace the modality recognition and feature extraction modules mentioned above. For example, CNN (Convolutional Neural Network) models, LSTM (Long Short-Term Memory) models, and GRU (Gated Recurrent Unit) models can be used for modality recognition and feature extraction.

[0086] like Figure 1 As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0087] Step S300: Select the corresponding filter based on the modality recognition result and dynamically adjust the filter parameters.

[0088] In this embodiment, after modality recognition and feature extraction, the above-mentioned adaptive filter module is used for filter selection and adaptive parameter tuning. That is, according to the recognized signal modality, a suitable filter combination is selected and the parameters are dynamically adjusted.

[0089] Specifically, in one implementation of this embodiment, step S300 includes the following steps:

[0090] Step S301: Select the corresponding filter combination according to the identified signal mode, and dynamically adjust the filter parameters in the combination.

[0091] In this embodiment, anti-interference filtering is performed by dynamically adjusting and combining different filters. Specifically, during the process of adjusting the filter combination, the optimal filter parameters can be adaptively selected based on the combined effect of the filters, thereby realizing the function of dynamic adjustment of filter parameters.

[0092] As an example, the filter combinations described in this embodiment include, but are not limited to, any combination of: band-stop filters, wavelet denoising, high-pass filters, adaptive filters, independent component analysis denoising, low-pass filters, notch filters, and Kalman filters.

[0093] In some application scenarios, the filter combinations selected by the adaptive filter module in this embodiment and their corresponding parameters are as follows:

[0094] 1) ECG signal: band-stop filter + wavelet noise reduction; wherein, the parameters of the band-stop filter are adjusted to: 50Hz±2Hz.

[0095] 2) EEG signal: High-pass filter + LMS adaptive filter (an adaptive filter that minimizes the error between the input signal and the desired signal by continuously adjusting the filter coefficients) + ICA (Independent Component Analysis for noise reduction); wherein, the parameters of the high-pass filter are adjusted to 1Hz.

[0096] 3) EMG signal: high-pass filter + low-pass filter; the parameters of the high-pass filter are adjusted to >20Hz; the parameters of the low-pass filter are adjusted to <500Hz.

[0097] 4) EDA signal: Low-pass filter removes fast disturbances; the parameters of the low-pass filter are adjusted to <5Hz.

[0098] 5) SpO2 signal: notch filter + Kalman filter to dynamically compensate for illumination artifacts.

[0099] In addition to the above filter combinations and parameter adjustment methods, in some application scenarios, wavelet denoising can be replaced with Empirical Mode Decomposition (EMD) method, or LMS adaptive filter can be replaced with RLS algorithm (adaptive filtering algorithm based on least squares method), or a separate mode-specific optimization algorithm can be selected for different mode usage scenarios to achieve the corresponding filtering method and parameter adjustment.

[0100] In this embodiment, by employing a multimodal adaptive filtering strategy, the filtering structure can be intelligently matched according to the signal type; and by employing a cascaded adjustable filtering mechanism, an adjustable processing chain can be formed by combining different filters.

[0101] like Figure 1As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0102] Step S400: Adjust the filtering path and sequence based on the cascaded filtering structure and dynamic switching mechanism.

[0103] In this embodiment, after filter selection and adaptive parameter tuning, based on the aforementioned interference detection and dynamic adjustment module, the cascaded filter structure and dynamic switching mechanism are used to adjust the filter path and filter order to adapt to different noise environments.

[0104] Specifically, in one implementation of this embodiment, step S400 includes the following steps:

[0105] Step S401: Obtain the real-time signal-to-noise ratio (SNR) detection results of each modal signal, and select a path that meets the current noise environment from the preset filtering paths based on the SNR detection results;

[0106] Step S402: Dynamically adjust the filter order and filter parameters in the filter combination according to the current noise characteristics of each modal signal.

[0107] In this embodiment, based on the real-time SNR (signal-to-noise ratio) detection results of each modal signal, the system selects the most suitable path from the preset filtering paths for the current noise environment. For example, in a high-noise environment, the system may select a path that includes an adaptive filter to more effectively suppress noise; and the system dynamically adjusts the order of filters according to the current noise characteristics. For example, when there is strong periodic interference, a notch filter is used first to remove noise at a specific frequency.

[0108] This embodiment employs a dynamic interference identification and switching mechanism, which utilizes SNR analysis to adjust the filter path, thereby achieving the functions of dynamic adjustment of the filter structure and adaptive adjustment of parameters.

[0109] like Figure 1 As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0110] Step S500: Adaptive interference estimation and cancellation are performed based on crosstalk processing and reference channel compensation mechanisms.

[0111] In this embodiment, in addition to adjusting the filtering path and sequence based on the cascaded filtering structure and dynamic switching mechanism, adaptive interference estimation and cancellation are also performed through crosstalk processing and reference channel compensation mechanisms; wherein, the crosstalk processing and reference channel compensation mechanisms are implemented through the aforementioned crosstalk suppression and signal fusion modules.

[0112] Specifically, in one implementation of this embodiment, step S500 includes the following steps:

[0113] Step S501: Construct a reference channel, estimate the interference signal in the main channel using the signal corresponding to the reference channel, and subtract the estimated value corresponding to the interference signal from the signal corresponding to the main channel to cancel the interference signal.

[0114] In this embodiment, reference channels for each modality first need to be constructed. In the signal acquisition system, the reference channels are used to capture signals related to interference sources in the main channel. For example, in electrocardiogram (ECG) signal processing, to eliminate electromyographic interference, additional electrodes can be placed near the muscle activity area to specifically acquire electromyographic signals as reference channel inputs. This reference signal should be highly correlated with the interference components in the main channel but should not contain the target signal components.

[0115] After constructing the reference channel, the interference signal in the main channel can be estimated using the signal corresponding to the reference channel, and the signal can be canceled based on the estimated value of the interference signal.

[0116] As an example, this embodiment employs the LMS algorithm for interference cancellation. The LMS algorithm iteratively adjusts the filter coefficients to minimize the error between the output signal and the desired signal. In interference cancellation, the LMS algorithm estimates the interference components in the main channel using the reference channel signal and subtracts this estimate from the main channel signal, thereby extracting a cleaner target signal. This process does not require prior knowledge of the specific characteristics of the interference and can adapt to changes in interference in real time, improving the robustness of signal processing.

[0117] In other application scenarios of this embodiment, an AI-assisted judgment mechanism can be added to achieve dynamic interference prediction, and a deep neural network can be used for unified modeling and processing of the entire solution.

[0118] like Figure 1 As shown in the figure, this embodiment of the invention provides an anti-interference filtering method for a multimodal physiological signal acquisition device, comprising the following steps:

[0119] Step S600: Perform parallel filtering based on the extracted features, compare the evaluation parameters of each filtering result based on the output result selection mechanism, and select and output the final signal based on the comparison results.

[0120] In this embodiment, after adaptive interference estimation and cancellation, multiple modal signals can be filtered in parallel during the actual filtering process. For each modal signal, multiple filters are also used in parallel processing. Then, by evaluating each filter scheme, the result that best matches the current noise environment and current signal characteristics is selected as the final output.

[0121] Specifically, in one implementation of this embodiment, step S600 includes the following steps:

[0122] Step S601: Perform parallel filtering processing based on the characteristics of each modal signal to generate the corresponding filtering results;

[0123] Step S602: Based on a preset strategy or algorithm, compare the evaluation parameters of each filtering result, and select the result that meets the current signal characteristics as the final signal for output.

[0124] In this embodiment, the parallel filtering process is as follows: the original signal is processed simultaneously through multiple filters, such as bandpass filters, notch filters, adaptive filters, etc., and then multiple filtering results are generated.

[0125] Furthermore, each filtering result is evaluated through feature evaluation; the feature evaluation method is as follows: for each filtering result, its key parameters such as SNR (signal-to-noise ratio), energy distribution, and spectral characteristics are calculated.

[0126] Optimal selection: Based on a preset strategy or algorithm, compare the evaluation parameters of each filtering result and select the result that best matches the current signal characteristics as the final output.

[0127] In this embodiment, the signal-to-noise ratio of the multimodal signal is significantly improved (by an average of 8-15 dB) through the above technical solutions; and the filtering delay is controlled within 10 ms to make it suitable for real-time systems.

[0128] It is worth mentioning that the above technical solutions in this embodiment support the deployment of edge computing platforms (such as ESP32 platform and STM32 platform) and can maintain high stability in dynamic motion or high interference environments.

[0129] As an example, the technical solution of this embodiment can be deployed in the following ways:

[0130] 1) Deployed on the STM32H7 platform, the filtering algorithm is implemented using CMSIS-DSP.

[0131] 2) Run FreeRTOS on the ESP32-S3 platform and dynamically switch between the LMS algorithm and the FIR algorithm.

[0132] 3) Combine AI models to identify interference types and call specific filtering schemes (such as edge neural networks to determine whether there is power frequency interference).

[0133] As an application prospect of this embodiment, the anti-interference filtering method of the multimodal physiological signal acquisition device can be deployed in combination with edge AI to improve the accuracy of interference identification; and it can be extended to multi-user, multi-channel remote physiological monitoring system; and it has good adaptability in Internet of Things (IoT) and can be used as the core algorithm of health management platform.

[0134] This embodiment achieves the following technical effects through the above technical solution:

[0135] This embodiment proposes an anti-interference filtering method for multimodal physiological signal acquisition equipment. Based on the spectral characteristics, noise types, and processing requirements of different physiological signal modes, a modality-adaptive filtering mechanism is constructed. Combined with a multi-stage filter cascade structure, it effectively suppresses multi-source interference such as power supply interference, motion artifacts, crosstalk, and high-frequency noise. Through interference identification and dynamic parameter adjustment mechanisms, the real-time performance and stability of the filtering effect are improved. This embodiment features good versatility, stability, and low computational resource consumption.

[0136] Exemplary device

[0137] Based on the above embodiments, the present invention also provides an anti-interference filtering system for a multimodal physiological signal acquisition device, comprising:

[0138] The signal preprocessing module is used to preprocess the acquired multimodal physiological signals;

[0139] The modality recognition and feature extraction module is used to perform modality recognition and feature extraction on preprocessed multimodal physiological signals;

[0140] The adaptive filter module is used to select the corresponding filter based on the modality recognition results and dynamically adjust the filter parameters;

[0141] The interference detection and dynamic adjustment module is used to adjust the filtering path and sequence based on the cascaded filtering structure and dynamic switching mechanism.

[0142] The crosstalk suppression and signal fusion module is used for adaptive interference estimation and cancellation based on crosstalk processing and reference channel compensation mechanisms.

[0143] The output selection and data transmission module is used to perform parallel filtering based on the extracted features, compare the evaluation parameters of each filtering result based on the output result selection mechanism, and select and output the final signal based on the comparison results.

[0144] This embodiment achieves the following technical effects through the above technical solution:

[0145] This embodiment proposes an anti-interference filtering system for multimodal physiological signal acquisition equipment. Based on the spectral characteristics, noise types, and processing requirements of different physiological signal modes, a modality-adaptive filtering mechanism is constructed. Combined with a multi-stage filter cascade structure, it effectively suppresses multi-source interference such as power supply interference, motion artifacts, crosstalk, and high-frequency noise. Through interference identification and dynamic parameter adjustment mechanisms, the real-time performance and stability of the filtering effect are improved. This embodiment features good versatility, stability, and low computational resource consumption.

[0146] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 3 As shown.

[0147] The terminal includes: a processor, a memory, an interface, a display screen, and a communication module connected via a system bus; wherein, the processor of the terminal provides computing and control capabilities; the memory of the terminal includes a computer-readable storage medium and internal memory; the computer-readable storage medium stores an operating system and computer programs; the internal memory provides an environment for the operation of the operating system and computer programs in the computer-readable storage medium; the interface is used to connect to external devices; the display screen is used to display relevant information; and the communication module is used to communicate with a cloud server or other devices.

[0148] When executed by the processor, this computer program is used to implement the anti-interference filtering method of the multimodal physiological signal acquisition device.

[0149] It will be understood by those skilled in the art that Figure 3 The schematic diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0150] In one embodiment, a terminal is provided, comprising: a processor and a memory, the memory storing an anti-interference filtering program for a multimodal physiological signal acquisition device, the anti-interference filtering program for the multimodal physiological signal acquisition device being executed by the processor to implement the operation of the anti-interference filtering method of the multimodal physiological signal acquisition device as described above.

[0151] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores an anti-interference filtering program for a multimodal physiological signal acquisition device, the anti-interference filtering program for the multimodal physiological signal acquisition device being executed by a processor to implement the operation of the anti-interference filtering method of the multimodal physiological signal acquisition device as described above.

[0152] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, database, or other media used in the embodiments provided by this invention can include both non-volatile and volatile memory.

[0153] In summary, this invention provides an anti-interference filtering method, system, terminal, and storage medium for a multimodal physiological signal acquisition device, comprising: preprocessing the acquired multimodal physiological signals; performing modality recognition and feature extraction on the preprocessed multimodal physiological signals; selecting the corresponding filter based on the modality recognition results and dynamically adjusting the filter parameters; adjusting the filtering path and order based on a cascaded filtering structure and a dynamic switching mechanism; performing adaptive interference estimation and cancellation based on crosstalk processing and a reference channel compensation mechanism; performing parallel filtering based on the extracted features; comparing the evaluation parameters of each filtering result based on an output result selection mechanism; and selecting and outputting the final signal based on the comparison results. This invention effectively suppresses multi-source interference such as power supply interference, motion artifacts, crosstalk, and high-frequency noise, improving the real-time performance and stability of the filtering effect.

[0154] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. An anti-interference filtering method for a multimodal physiological signal acquisition device, characterized in that, include: Preprocessing of the acquired multimodal physiological signals; Modality recognition and feature extraction are performed on the preprocessed multimodal physiological signals; Select the corresponding filter based on the modality recognition results, and dynamically adjust the filter parameters; Specific filter combinations and parameters are used to process different physiological signal modes: ECG signal: processed using a band-stop filter and wavelet noise reduction; wherein, the parameters of the band-stop filter are adjusted to: 50Hz±2Hz; EEG signal: Processed using a high-pass filter, an LMS adaptive filter, and ICA; wherein, the parameters of the high-pass filter are adjusted to 1Hz; EMG signal: Processed using a high-pass filter and a low-pass filter; wherein, the parameters of the high-pass filter are adjusted to >20Hz; and the parameters of the low-pass filter are adjusted to <500Hz. EDA signal: Processed by removing fast disturbances using a low-pass filter; wherein the parameters of the low-pass filter are adjusted to: <5Hz; SpO2 signal: Processed by using notch filter and Kalman filter to dynamically compensate for illumination artifacts; Adjusting the filtering path and sequence based on a cascaded filtering structure and dynamic switching mechanism; Adaptive interference estimation and cancellation are performed based on crosstalk processing and reference channel compensation mechanisms; an AI-assisted judgment mechanism is added to achieve dynamic interference prediction; and deep neural networks are used for unified modeling and processing of the entire scheme. Parallel filtering is performed based on the extracted features. The evaluation parameters of each filtering result are compared based on the output result selection mechanism, and the final signal is selected and output based on the comparison results. The adjustment of the filtering path and order based on the cascaded filtering structure and dynamic switching mechanism includes: The signal-to-noise ratio (SNR) of each modal signal is obtained in real time, and a path that meets the current noise environment is selected from the preset filtering paths based on the SNR detection results. The filter order and filter parameters in the filter combination are dynamically adjusted according to the current noise characteristics of each modal signal. The adaptive interference estimation and cancellation based on crosstalk processing and reference channel compensation mechanism includes: A reference channel is constructed, and the interference signal in the main channel is estimated using the signal corresponding to the reference channel. The estimated value corresponding to the interference signal is then subtracted from the signal corresponding to the main channel to cancel out the interference signal. The process of performing parallel filtering based on extracted features, comparing evaluation parameters of each filtering result based on an output result selection mechanism, and selecting and outputting the final signal based on the comparison results includes: Parallel filtering is performed based on the characteristics of each modal signal to generate the corresponding filtering results; Based on a preset strategy or algorithm, the evaluation parameters of each filtering result are compared, and the result that meets the current signal characteristics is selected as the final signal for output.

2. The anti-interference filtering method for the multimodal physiological signal acquisition device according to claim 1, characterized in that, The preprocessing of the acquired multimodal physiological signals includes: The amplitude of the multimodal physiological signal is normalized, DC bias is removed by mean removal and high-pass filter method, and outliers are removed by standard fraction method, interquartile range method and sliding window method to obtain the preprocessed multimodal physiological signal.

3. The anti-interference filtering method for the multimodal physiological signal acquisition device according to claim 1, characterized in that, The modality recognition and feature extraction of the preprocessed multimodal physiological signals includes: Identify the signal category corresponding to each modality signal in the preprocessed multimodal physiological signal; The time-domain signals of each modal signal are converted into frequency-domain signals by using spectrum analysis, and the frequency component with the largest amplitude is identified to obtain the corresponding dominant frequency characteristics. The corresponding waveform energy characteristics are obtained by summing the squared amplitudes of each sampling point of each modal signal; Calculate the difference between adjacent sampling points of each modal signal, count the number of mutation points exceeding a set threshold, and calculate the ratio of the number of mutation points to the total number of sampling points to obtain the corresponding mutation rate feature. Calculate the ratio of signal power to noise power for each modal signal to obtain the corresponding signal-to-noise ratio characteristics.

4. The anti-interference filtering method for the multimodal physiological signal acquisition device according to claim 1, characterized in that, The step of selecting the corresponding filter based on the modality recognition result and dynamically adjusting the filter parameters includes: Select the corresponding filter combination based on the identified signal mode, and dynamically adjust the filter parameters in the combination; The filter combination includes any combination of band-stop filters, wavelet denoising, high-pass filters, adaptive filters, independent component analysis denoising, low-pass filters, notch filters, and Kalman filters.

5. An anti-interference filtering system for a multimodal physiological signal acquisition device, used to implement the anti-interference filtering method for the multimodal physiological signal acquisition device as described in any one of claims 1-4, characterized in that, include: The signal preprocessing module is used to preprocess the acquired multimodal physiological signals; The modality recognition and feature extraction module is used to perform modality recognition and feature extraction on preprocessed multimodal physiological signals; The adaptive filter module is used to select the corresponding filter based on the modality recognition results and dynamically adjust the filter parameters; The interference detection and dynamic adjustment module is used to adjust the filtering path and sequence based on the cascaded filtering structure and dynamic switching mechanism. The crosstalk suppression and signal fusion module is used for adaptive interference estimation and cancellation based on crosstalk processing and reference channel compensation mechanisms. The output selection and data transmission module is used to perform parallel filtering based on the extracted features, compare the evaluation parameters of each filtering result based on the output result selection mechanism, and select and output the final signal based on the comparison results.

6. A terminal, characterized in that, include: The processor and memory, wherein the memory stores an anti-interference filtering program for a multimodal physiological signal acquisition device, and the anti-interference filtering program for the multimodal physiological signal acquisition device, when executed by the processor, is used to implement the operation of the anti-interference filtering method for the multimodal physiological signal acquisition device as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an anti-interference filtering program for a multimodal physiological signal acquisition device. When executed by a processor, the anti-interference filtering program for the multimodal physiological signal acquisition device is used to implement the operation of the anti-interference filtering method for the multimodal physiological signal acquisition device as described in any one of claims 1-4.

Images