Cross-scene reusable multi-modal physiological intelligent monitoring equipment and system

Abstract

The invention discloses cross-scene reusable multi-modal physiological intelligent monitoring equipment and system. The equipment comprises a multi-modal physiological acquisition sensing array, an acquisition control module, a placement detection module, a self-adaptive parameter adjustment module and a signal processing module. The multi-modal physiological acquisition sensing array comprises an ambient light sensor, a photoelectric volume pulse wave sensor, a temperature sensor, a sensing electrode, a pressurizing electrode, a sound collector, an inertial measurement unit and the like; the placement detection module is used for calculating the comprehensive confidence coefficient of the current physiological electric signal belonging to each physiological signal type for each signal acquired by the multi-modal physiological acquisition sensing array, so as to judge the physiological signal type to which the physiological electric signal belongs; the adaptive parameter adjusting module is used for generating a sampling rate, an amplification factor, a filtering bandwidth and an ADC bit depth corresponding to the current physiological signal type; and the signal processing module is used for carrying out signal processing according to the parameters of the self-adaptive parameter adjusting module. Multi-modal physiological signal monitoring can be realized by adopting one set of equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to signal acquisition technology, and in particular to a cross-scene reusable multi-modal physiological intelligent monitoring device and system. BACKGROUND

[0002] The mainstream wearable devices on the market at present (such as smart bracelets, smart watches, smart headbands, portable electrocardiographs, breathing bands, etc.) mainly collect physiological signals of specific and single points, and use built-in fixed algorithms for analysis, and the application scenarios are relatively limited. Taking products in the form of bracelets as an example, the bracelet can only be placed on the wrist and is usually used to collect local physiological parameters of specific parts such as heart rate, blood oxygen, gait, etc., and it is difficult to realize the synchronous acquisition of multi-modal and multi-site signals. Such devices are limited by the hardware structure and the closed nature of the algorithm, and usually only support a limited number of physiological index detection, which cannot meet the comprehensive perception and analysis needs of more physiological information such as electroencephalogram, electromyogram, respiration, skin electricity, etc. in complex scenes.

[0003] In the process of multi-modal physiological signal acquisition, the existing technology usually needs to use multiple independent physiological monitoring devices to collect different types of signals such as electroencephalogram, electrocardiogram, electromyogram, respiration, skin electricity, etc. Since these devices come from different manufacturers, use different sampling clock sources and communication protocols, there are often problems such as inconsistent time reference, different sampling frequency, and non-uniform data acquisition starting point. Therefore, most of the existing devices lack a unified data synchronization mechanism, and users usually rely on manual marking or post-alignment methods to process data in the use process, which is not only complicated to operate, but also difficult to meet the strict requirements of high-precision scientific research and clinical application for multi-channel physiological signal synchronous acquisition. SUMMARY

[0004] In view of the problems existing in the prior art, the purpose of the present application is to provide a cross-scene reusable multi-modal physiological intelligent monitoring device and system.

[0005] In order to achieve the above-mentioned purpose of the application, the present application provides the following technical scheme:

[0006] A cross-scene reusable multi-modal physiological intelligent monitoring device, comprising:

[0007] A multi-modal physiological acquisition sensor array, comprising a plurality of array-arranged sensor units, the sensor units comprising a sensing electrode, a pressure electrode, a sound collector, an inertial measurement unit, etc.

[0008] An acquisition control module for controlling the sensing electrode to collect physiological electrical signals spontaneously generated by the human body, and controlling the pressure electrode to generate an excitation current input to the human skin, and collecting physiological electrical signals generated by the human body under the excitation current through the sensing electrode;

[0009] The placement detection module is used to calculate the comprehensive confidence level of the current physiological electrical signal belonging to each physiological signal type from the signals collected by the multimodal physiological acquisition sensor array, thereby determining the physiological signal type to which the physiological electrical signal collected by the sensing electrode belongs. The physiological electrical signal types include electrocardiogram (ECG), electroencephalogram (EEG), electromyogram (EMG), electrodermal signal, and respiratory signal. Respiratory signal and EMG are physiological electrical signals acquired by the human body under excitation current, while ECG, EEG, and EMG are physiological electrical signals spontaneously generated by the human body.

[0010] An adaptive parameter adjustment module is used to generate the sampling rate, amplification factor, filter bandwidth, and ADC bit depth corresponding to the current physiological signal type;

[0011] The signal processing module is used to amplify the signal according to the amplification factor, perform analog-to-digital conversion according to the sampling rate and ADC bit depth, filter the signal according to the filtering bandwidth, and encode and compress the signal according to the encoding and compression strategy to obtain the processed physiological electrical signal.

[0012] Furthermore, the multimodal physiological acquisition sensor array is fixed to the surface of human skin by a stretchable flexible strap or patch.

[0013] Furthermore, the placement detection module includes:

[0014] The preprocessing unit is used to preprocess the physiological electrical signals collected by the sensing electrodes in the multimodal physiological acquisition sensor array, the sound signals collected by the sound acquisition unit, and the inertial signals collected by the inertial measurement unit.

[0015] The feature calculation unit is used to perform feature calculations on each signal within a preset window time after preprocessing. Specifically, it includes: calculating the waveform characteristics and contact impedance of the physiological electrical signal based on the physiological electrical signal; calculating the energy spectrum and envelope of the sound signal based on the sound signal; and calculating the attitude angle of the current position based on the inertial signal.

[0016] The position determination unit is used to determine the current acquisition position based on the attitude angle, and to determine the position confidence of the current physiological electrical signal belonging to each physiological signal type based on the current acquisition position;

[0017] The contact determination unit is used to determine the contact confidence level of the current physiological electrical signal belonging to each physiological signal type based on the contact impedance.

[0018] The waveform feature determination unit is used to determine the signal attribute confidence level of the physiological electrical signal belonging to each physiological signal type based on the waveform features of the physiological electrical signal.

[0019] The sound determination unit is used to determine whether a sound signal belongs to the heart sound component or the pulse component based on the energy spectrum and envelope of the sound signal, and to determine the sound attribute confidence of the current physiological electrical signal belonging to each physiological signal type based on the attributes of the sound signal.

[0020] The type determination unit is used to perform weighted fusion of location confidence, contact confidence, signal attribute confidence and sound attribute confidence to obtain a comprehensive confidence score, and select the type with the highest comprehensive confidence score as the physiological signal type.

[0021] Furthermore, the waveform feature determination unit specifically includes:

[0022] The ECG signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal to be the ECG signal when a periodic narrow QRS peak is detected in the waveform of the physiological electrical signal, and an SCG peak appears after the R wave in the QRS peak.

[0023] The EEG signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal to be an EEG signal when the waveform of the physiological electrical signal does not have QRS peaks, and the low-frequency energy is dominant and the amplitude is in the μV range.

[0024] The electromyography signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as an electromyography signal to be the highest when the high-frequency energy of the physiological electrical signal is dominant and appears in the form of a short burst.

[0025] The skin conductance signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal belonging to the skin conductance signal to the maximum when the spectral energy of the contact impedance signal is mainly concentrated in the low frequency part and the signal consists of a slowly changing baseline (skin conductance level, SCL) and a rapid transient response (skin conductance response, SCR).

[0026] The respiratory signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as a respiratory signal to be the highest when the low-frequency envelope of the contact impedance is synchronized with the envelope of the sound signal and has periodicity in the respiratory frequency band.

[0027] Furthermore, the sound determination unit specifically includes:

[0028] The first determination subunit is used to determine that the current sound signal belongs to the heart sound component when the energy spectrum of the sound signal detects the S1 / S2 energy peak and the sound signal envelope detects the short-term heart sound envelope. Then, the confidence level of the physiological electrical signal belonging to the signal attribute of the electrocardiogram signal and the respiratory signal is set to be the highest.

[0029] The second determination subunit is used to determine that the current sound signal belongs to the pulse component when the sound signal envelope detects the pulse envelope, and to set the signal attribute of the physiological electrical signal to have the highest confidence level as belonging to electromyography and electrodermal signals.

[0030] Furthermore, the adaptive sampling adjustment module specifically includes:

[0031] The modality determination unit is used to determine the current acquisition modality M* based on the comprehensive confidence level of the current physiological signal. The acquisition modality includes ECG acquisition modality, EEG acquisition modality, EMG acquisition modality, SCTA acquisition modality, and respiratory impedance acquisition modality.

[0032] The short-time observation unit is used to measure and estimate the following observations based on physiological signals: short-time signal amplitude estimation. rms Motion index M and available bandwidth BW avail ;

[0033] The target bandwidth and sampling rate calculation unit is used to determine the filtering bandwidth based on the nominal bandwidth of the current acquisition mode M*, the overall confidence level, and the motion index M. :

[0034]

[0035]

[0036] in, As a bandwidth security factor, it is determined by querying the preset parameters. The mapping table between M and bandwidth security factor is obtained. This represents the overall confidence level of the current acquisition mode M*. Indicates the sampling rate. The nominal bandwidth of the dominant mode M* is determined by the inherent spectral width occupied by the physiological signal corresponding to the dominant mode M*. Indicates the oversampling factor;

[0037] Amplification factor and ADC bit depth configuration unit, used to estimate signal based on short-time amplitude. rms Estimating short-time peak-to-peak value and based on and ADC full amplitude Calculate the amplification factor G, and calculate the ADC bit depth b based on the target signal-to-noise ratio:

[0038]

[0039]

[0040] in, This refers to the clip function. Indicates the target occupancy percentage. , Indicates the minimum and maximum magnification; Indicates the target signal-to-noise ratio. It represents rounding up;

[0041] The encoding and compression strategy setting unit is used to calculate the original code rate based on the ADC bit depth and sampling rate. And based on available bandwidth With energy budget EB, the original bitrate R raw Mapped to target bitrate This allows us to determine the required compression ratio. And select an encoding compression strategy with a compression ratio of CR;

[0042] The parameter sending unit is used to send the magnification factor G and sampling rate. The ADC bit depth (b) and encoding compression strategy are respectively distributed to the amplifier, analog-to-digital converter, and encoder.

[0043] Furthermore, the sensing unit also includes a photoplethysmography (PPG) sensor, which is used to acquire pulse wave signals.

[0044] The acquisition modes also include pulse wave acquisition mode, heart sound acquisition mode, and breath sound acquisition mode. When the photoplethysmography pulse wave sensor acquires a pulse wave signal and the physiological electrical signal acquired by the sensing electrode is lower than a preset threshold, the current acquisition mode is determined to be pulse wave acquisition mode. When the sound signal acquired by the sound acquisition device is a heart sound, the current acquisition mode is determined to be heart sound acquisition mode. When the sound signal acquired by the sound acquisition device is a breath sound, the current acquisition mode is determined to be breath sound acquisition mode.

[0045] Furthermore, the sensing unit also includes a temperature sensor and an ambient light sensor. The temperature sensor is used to collect skin temperature or ambient temperature, and the ambient light sensor is used to collect external light intensity.

[0046] Furthermore, the adaptive sampling adjustment module also includes:

[0047] The short-time acknowledgment and quality inspection unit is used to calculate the real-time signal-to-noise ratio (SNR) within the short-time acknowledgment window. est Key feature detection rate and reconstruction error ε, if SNR est If the values ​​are greater than the threshold and the key feature detection rate is greater than the threshold, then accept the current parameter configuration and continue running; otherwise, execute the closed-loop adjustment unit.

[0048] The closed-loop adjustment unit is used to adjust according to the following priority order:

[0049] (1) Increase the magnification;

[0050] (2) Increase the sampling rate;

[0051] (3) Reduce compression intensity or switch to near-lossless / lossless encoding;

[0052] (4) If the above measures fail to improve the situation, the user will be prompted to reposition the device or proceed with manual calibration.

[0053] Furthermore, the adaptive sampling adjustment module also includes:

[0054] The event-driven switching unit is used to immediately switch the device to event-driven mode when a predefined abnormal event is identified during normal data processing. This temporarily increases the sampling rate and reduces the compression intensity to record lossless or near-lossless data for the event period. After the event ends, the system returns to normal parameters.

[0055] A reusable, multimodal physiological intelligent monitoring system for various scenarios includes:

[0056] The aforementioned multimodal physiological intelligent monitoring equipment; and

[0057] The intelligent physiological signal analysis platform provides an external program interface to perform personalized and custom analysis and application of processed physiological signals.

[0058] Furthermore, the intelligent physiological signal analysis platform includes:

[0059] The host connection and control platform is used for bidirectional communication with multimodal physiological intelligent monitoring equipment to acquire data and send commands;

[0060] A signal decoding platform used to reconstruct physiological electrical signals through specific decoding methods;

[0061] The middle layer is used for alignment, standardization, and resource scheduling, enabling the unified management and use of physiological electrical signals of different frequencies and characteristics.

[0062] The signal analysis plugin platform is used to upload and download user-defined external programs. These external programs call physiological electrical signals through an intermediate layer to achieve targeted analysis and personalized applications of various physiological signals.

[0063] Compared with existing technologies, the advantages of this invention are as follows: The reusable multimodal physiological intelligent monitoring device of this invention can achieve simultaneous measurement of one or more of the following signals using a single device: electroencephalogram (EEG), electrocardiogram (ECG), respiratory signal, electrodermal signal, and electromyographic signal, without the need for separate measurements using multiple devices. Furthermore, based on the concepts of hardware modularization and algorithm plug-in, this invention proposes a reusable multimodal physiological intelligent monitoring system. The modular design fulfills the reusability requirements of the hardware devices, allowing users to combine the device with specific accessories in specific scenarios and needs, effectively improving the efficiency and applicability of a single wearable device. Attached Figure Description

[0064] Figure 1 This is a schematic diagram of the structure of the cross-scenario reusable multimodal physiological intelligent monitoring device provided by the present invention;

[0065] Figure 2 This is a schematic diagram illustrating the use of the multimodal physiological intelligent monitoring device that can be reused across different scenarios provided by the present invention;

[0066] Figure 3 This is a schematic diagram of the structure of the cross-scenario reusable multimodal physiological intelligent monitoring system provided by the present invention. Detailed Implementation

[0067] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0068] Example 1

[0069] This invention provides a reusable, multimodal physiological intelligent monitoring device that can be used across different scenarios, such as... Figure 1 As shown, it includes a multimodal physiological acquisition sensor array, an acquisition control module, a placement detection module, an adaptive parameter adjustment module, and a signal processing module.

[0070] A multimodal physiological data acquisition sensor array includes several arrayed sensing units. These sensing units include ambient light sensors, photoplethysmography (PPG) sensors, temperature sensors, sensing electrodes, pressure electrodes, sound acquisition units, inertial measurement units, etc. During data acquisition, the data is directly applied to the skin surface of the measurement site via a retractable flexible strap or adhesive electrode pads, or connected to external channels such as metal contacts or embedded electrode sockets, such as electrode wires or blood oxygen signal acquisition heads, which can be connected to the human body for signal transmission. Specifically, for example... Figure 2As shown, the multimodal physiological data acquisition sensor array, through flexible straps, headbands, wristbands, cuffs, patch structures, or support structures, possesses stretchability and adjustability, allowing it to conform to the skin surface of different monitoring sites such as the head, chest, abdomen, or limbs. This enables stable acquisition of various signals, including electroencephalograms (EEG), electrocardiograms (ECG), respiratory rate monitoring (RIP), electromyography (EMG), and electrical conduction analysis (EDA / impedance). Specifically, the sensing electrodes can be carriers of electrical signals, such as ECG and EEG; or they can be photoelectric, piezoelectric, or electrochemical sensors used to acquire physiological parameters such as respiration, skin conduction activity, and body surface temperature, as well as environmental parameters such as ambient temperature and ambient light. The sound acquisition device, such as a microphone, can acquire sound, such as pulse sounds, breathing sounds, or heartbeat sounds. The inertial measurement unit, consisting of a three-axis accelerometer and a gyroscope, can simultaneously record the user's body movement and posture changes, thereby providing more comprehensive multi-dimensional data support for subsequent physiological state analysis.

[0071] The acquisition and control module is used to control the sensing electrodes to collect spontaneously generated physiological electrical signals from the human body, and to control the pressure electrodes to generate an excitation current that is input into the human skin. The sensing electrodes then collect the physiological electrical signals generated by the human body under this excitation current. Electrocardiogram (ECG), electroencephalogram (EEG), and electromyogram (EMG) signals are spontaneously generated physiological electrical signals that can be directly acquired through the sensing electrodes. Human inhalation and exhalation cause changes in chest cavity movement, resulting in increases and decreases in impedance. These changes in impedance cause corresponding changes in the electrical signal response under the excitation current; this change in electrical signal is the respiratory waveform (RIP). Skin conductance essentially monitors changes in skin resistance / conductivity through the responding electrical signal. Therefore, respiratory signals and skin conductance signals are physiological electrical signals generated by the human body under an excitation current.

[0072] The placement detection module is used to calculate the comprehensive confidence level of each physiological signal acquired by the multimodal physiological acquisition sensor array, thereby determining the physiological signal type to which the current physiological electrical signal belongs. These physiological electrical signal types include electrocardiogram (ECG), electroencephalogram (EEG), electromyography (EMG), electrodermal conductance (EDC), and respiratory signals.

[0073] The placement detection module specifically includes a preprocessing unit, a feature calculation unit, a position determination unit, a contact determination unit, a waveform feature determination unit, a sound determination unit, and a type determination unit.

[0074] The preprocessing unit is used to preprocess the physiological electrical signals collected by the sensing electrodes in the multimodal physiological acquisition sensor array, the temperature signals collected by the temperature sensor, the sound signals collected by the sound collector, the ambient light signals collected by the ambient light sensor, and the inertial signals collected by the inertial measurement unit. Specifically, it includes DC removal and bandpass filtering.

[0075] The feature calculation unit is used to perform feature calculations on each signal within a preset window time (short-time calibration window, T = 5 s) after preprocessing. Specifically, it includes: calculating the waveform characteristics and contact impedance of the physiological electrical signal based on the physiological electrical signal; calculating the skin temperature or ambient temperature based on the temperature signal; calculating the energy spectrum and envelope of the sound signal based on the sound signal; calculating the light intensity based on the ambient light signal; and calculating the attitude angle of the current position based on the inertial signal. The methods used for feature calculation include short-time energy / envelope calculation, power spectrum estimation (Welch), instantaneous peak detection, and short-time cross-correlation / phase delay estimation.

[0076] The position determination unit is used to determine the current acquisition position based on the attitude angle, and then determine the position confidence score of the current physiological electrical signal belonging to each physiological signal type based on the current acquisition position. Specifically, the attitude angle includes pitch, roll, and yaw. Pitch is the pitch angle, roll is the roll angle, and yaw is the yaw angle. For example, if the pitch is between 60° and 120°, the current acquisition position can be determined to be the chest, and it is highly likely that an electrocardiogram (ECG) or respiratory signal is being acquired. Therefore, the position confidence score for each physiological signal type can be set as posescore. ECG = 0.6, posescore RIP = 0.1, posescore EEG =0.1, posescore EMG = 0.1, posescore EDA = 0.1; If the pitch is below 30°, the current acquisition location is determined to be the head, and there is a high probability of acquiring EEG data. In this case, the position confidence score for each physiological signal type can be set as posescore. ECG = 0.1, posescore RIP = 0.1, posescore EEG = 0.6, posescore EMG = 0.1, posescore EDA = 0.1; If the pitch is between 30° and 60°, the current acquisition location is determined to be a limb or wrist, and there is a high probability of acquiring electromyography (EMG) or electrodermal activity (EDA). In this case, the position confidence score for each physiological signal type can be set as posescore. ECG = 0.1, posescore RIP = 0.1, posescore EEG = 0.1, posescore EMG = 0.35, posescore EDA= 0.35. In addition, the location can be determined by ambient light / temperature. For example, if the light intensity is higher than a preset threshold and the temperature is within a preset range, the location can be determined to be on the head. In this case, the confidence level of the EEG signal can be set to the highest.

[0077] The contact determination unit is used to determine the contact confidence level of the current physiological electrical signal according to the contact impedance. For example, if the contact impedance is less than 5kΩ, it indicates that the electrode and skin have formed an excellent conductive path, and the position confidence level for each physiological signal type is set as cscore. ECG = 0.25, cscore RIP = 0.04, cscore EEG = 0.5, cscore EMG = 0.2, cscore EDA = 0.01; If the contact impedance is not less than 100kΩ, then set the location confidence score for each physiological signal type as cscore. ECG = 0.05, cscore RIP = 0.15, cscore EEG = 0.05, cscore EMG = 0.05, cscore EDA = 0.7. Furthermore, contact determination can be based on temperature and light intensity. For example, if the skin temperature / surface temperature T_skin collected by the temperature sensor is higher than the skin temperature threshold (e.g., T_skin_threshold = 30 ℃) and the ambient light intensity is lower than the occlusion threshold (e.g., Ambient_Lux_threshold = 10 lux, indicating the device is covered or in close contact with the skin), then the user can be advised of good contact, possibly indicated by electromyography (EMG), respiratory signals, or electrocardiogram (ECG) signals. If the impedance is high and fluctuates significantly over time, or the temperature is lower than the skin temperature threshold, or the ambient light intensity is high (indicating no contact / exposure), then poor contact quality is indicated, possibly indicated by EEG signals from the head area.

[0078] The waveform feature determination unit is used to determine the signal attribute confidence level of a physiological electrical signal belonging to various physiological signal types based on the waveform characteristics of the physiological electrical signal. Specifically, the waveform feature determination unit includes an electrocardiogram (ECG) signal determination subunit, an electroencephalogram (EEG) signal determination subunit, an electromyography (EMG) signal determination subunit, an electrodermal signal determination subunit, and a respiratory signal determination subunit. The ECG signal determination subunit is used to set the signal attribute confidence level of the physiological electrical signal to be the highest when a periodic narrow QRS complex is detected in the waveform of the physiological electrical signal, and an SCG peak appears after the R wave in the QRS complex. For example, it sets the confidence level to be the highest, such as `contscore`. ECG =0.8, contscoreRIP = 0.05, contscore EEG = 0.05, contscore EMG = 0.05, contscore EDA =0.05; The EEG signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as an EEG signal to be the highest when the waveform of the physiological electrical signal does not have QRS peaks, and the low-frequency energy is dominant with an amplitude in the μV range. For example, contscore ECG =0.05, contscore RIP = 0.1, contscore EEG = 0.7, contscore EMG = 0.1, contscore EDA =0.05; The electromyography (EMG) signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as an EMG signal to be the highest when the high-frequency energy of the physiological electrical signal is dominant and appears in the form of a short burst. For example, contscore ECG = 0.06, contscore RIP = 0.02, contscore EEG = 0.14, contscore EMG = 0.76, contscore EDA = 0.02; The skin conductance signal determination subunit is used to measure skin impedance. It is used to set the signal attribute of the physiological electrical signal belonging to the skin conductance signal with the highest confidence level when the contact impedance signal spectrum energy is mainly concentrated in the low-frequency part and the signal consists of a slowly changing baseline (skin conductance level, SCL) and a rapid transient response (skin conductance response, SCR). For example, contscore. ECG =0.08, contscore RIP = 0.14, contscore EEG = 0.07, contscore EMG = 0.05, contscore EDA =0.66; The respiratory signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal belonging to the respiratory signal to the maximum when the low-frequency envelope of the contact impedance is synchronized with the sound signal envelope and has periodicity in the respiratory frequency band. For example, contscore ECG = 0.05, contscore RIP = 0.75, contscore EEG = 0.06, contscore EMG= 0.02, contscore EDA = 0.12. In addition, the pulse wave PPG signal can be combined to determine whether it is a respiratory signal. For example, when the low-frequency envelope of the contact impedance is synchronized with the PPG envelope (the delay can be corrected) and the respiratory frequency band (0.05–0.5 Hz) has obvious periodicity, the signal attribute that is determined to be a respiratory signal has the highest confidence.

[0079] The sound determination unit is used to determine whether a sound signal belongs to the heart sound component or the pulse component based on the sound signal energy spectrum and envelope, and to determine the sound attribute confidence level of the current physiological electrical signal belonging to each physiological signal type based on the sound signal attributes. The sound determination unit includes a first determination subunit and a second determination subunit. The first determination subunit is used to determine that the current sound signal belongs to the heart sound component when the sound signal energy spectrum detects an S1 / S2 energy peak and the sound signal envelope detects a short-term heart sound envelope. In this case, the signal attribute confidence level of the physiological electrical signal belonging to the ECG and respiratory signals is set to the highest, for example, audioscore. ECG = 0.52, audioscore RIP = 0.45, audioscore EEG = 0.01, audioscore EMG = 0.01, audioscore EDA = 0.01; The second determination subunit is used to determine that the current sound signal belongs to the pulse component when the sound signal envelope detects the pulse envelope (corresponding to the PPG peak). It sets the signal attribute of the physiological electrical signal belonging to electromyography (EMG) and electrodermal (ED) signals to have the highest confidence level, for example, audioscore. ECG = 0.05, audioscore RIP =0.04, audioscore EEG = 0.01, audioscore EMG = 0.48, audioscore EDA = 0.42.

[0080] The type determination unit is used to weightedly fuse location confidence, contact confidence, signal attribute confidence, and sound attribute confidence to obtain a comprehensive confidence score. The type with the highest comprehensive confidence score is selected as the corresponding physiological signal type. Specifically, the fusion can be performed according to the following formula:

[0081] ,

[0082] In the formula, , , These represent the weights of the corresponding types; specifically, they can be set. =0.3, =0.2, =0.4, =0.1. The above weights are exemplary values ​​for the example implementation and can be adjusted according to the specific implementation.

[0083] The adaptive parameter adjustment module is used to generate the sampling rate, amplification factor, filter bandwidth, and ADC bit depth corresponding to the current physiological signal type. The adaptive parameter adjustment module specifically includes:

[0084] The mode determination unit is used to determine the current acquisition mode M* based on the comprehensive confidence level of the current physiological signal. The acquisition modes include ECG acquisition mode, EEG acquisition mode, EMG acquisition mode, SCADA acquisition mode, and respiratory impedance acquisition mode. In addition, the acquisition mode may also include pulse wave acquisition mode, heart sound acquisition mode, and breath sound acquisition mode. The specific rules for determining the pulse wave acquisition mode, heart sound acquisition mode, and breath sound acquisition mode are as follows: when the sound signal acquired by the sound acquisition device is a heart sound, the current acquisition mode is determined to be the heart sound acquisition mode; when the sound signal acquired by the sound acquisition device is a breath sound, the current acquisition mode is determined to be the breath sound acquisition mode; when the photoplethysmography (PPG) pulse wave sensor acquires a pulse wave signal, and the signal acquired by the sensing electrode is not significant, it is determined to be the pulse wave acquisition mode.

[0085] The short-time observation unit is used to measure and estimate the following observations based on physiological electrical signals: short-time signal amplitude estimation. rms Motion index M (normalized short-time acceleration variance calculated by the IMU) and available bandwidth BW avail ;

[0086] The target bandwidth and sampling rate calculation unit is used to determine the filtering bandwidth based on the nominal bandwidth of the current acquisition mode M*, the overall confidence level, and the motion index M. :

[0087]

[0088]

[0089] in, As a bandwidth security factor, it is determined by querying the preset parameters. The mapping table between M and bandwidth security factor is obtained. This mapping table is existing technology and will not be described in detail here. This represents the overall confidence level of the current acquisition mode M*. Indicates the sampling rate. The nominal bandwidth of the dominant mode M* is determined by the inherent spectral width occupied by the physiological signal corresponding to the dominant mode M*. Indicates the oversampling factor;

[0090] Amplification factor and ADC bit depth configuration unit, used to estimate signal based on short-time amplitude. rms Estimating short-time peak-to-peak value and based on and ADC full amplitude Calculate the amplification factor G, and calculate the ADC bit depth b based on the target signal-to-noise ratio:

[0091]

[0092]

[0093] in, This refers to the clip function. Indicates the target occupancy percentage. , Indicates the minimum and maximum magnification; Indicates the target signal-to-noise ratio. It represents rounding up;

[0094] The encoding and compression strategy setting unit is used to calculate the original code rate based on the ADC bit depth and sampling rate. And based on available bandwidth With energy budget EB, the original bitrate R raw Mapped to target bitrate This allows us to determine the required compression ratio. The system selects a compression strategy with a compression ratio of CR. If CR≤2, it adopts reversible predictive coding + entropy coding; if CR>2 and≤6, it adopts lossy prediction or wavelet compression; if CR>6, it adopts hierarchical summarization + event-triggered reporting strategy.

[0095] The parameter sending unit is used to send the magnification factor G and sampling rate. The ADC bit depth (b) and encoding compression strategy are respectively distributed to the amplifier, analog-to-digital converter, and encoder.

[0096] The short-time acknowledgment and quality check unit is used to calculate the real-time signal-to-noise ratio (SNR) within a short-time acknowledgment window (e.g., 2–5 seconds). est Key feature detection rates (e.g., R-peak detection rate of ECG, respiratory correlation coefficient of RIP, etc.) and reconstruction error ε, if SNR est If the values ​​are greater than the threshold and the key feature detection rate is greater than the threshold, then accept the current parameter configuration and continue running; otherwise, execute the closed-loop adjustment unit.

[0097] The closed-loop adjustment unit is used to adjust according to the following priority order:

[0098] (1) Increase the magnification factor (if it does not cause saturation);

[0099] (2) Increase the sampling rate (if hardware and bandwidth allow);

[0100] (3) Reduce compression intensity or switch to near-lossless / lossless encoding;

[0101] (4) If the above measures fail to improve the situation, the user should be prompted to reposition the device or proceed with manual calibration.

[0102] The event-driven switching unit is used to immediately switch the device to event-driven mode when a predefined abnormal event (such as ventricular premature beats, epileptic seizure-like waveforms, or sudden high-amplitude EMG) is identified during normal data processing. This temporarily increases the sampling rate and reduces the compression intensity to record lossless or near-lossless data for the event period. After the event ends, the system returns to normal parameters.

[0103] The signal processing module amplifies the signal according to the stated amplification factor, performs analog-to-digital conversion according to the stated sampling rate and ADC bit depth, filters the signal according to the stated filtering bandwidth, and encodes and compresses the signal according to the stated encoding and compression strategy to obtain the processed physiological electrical signal. This module includes, but is not limited to, a preamplifier circuit, a filtering circuit, and a high-precision analog-to-digital converter (ADC). The preamplifier circuit amplifies the weak physiological signal initially, increasing the signal amplitude; the filtering circuit removes noise interference from the signal, such as power frequency interference, electromyography interference, and motion artifacts, thereby improving signal quality; the high-precision ADC converts the amplified and filtered analog signal into a digital signal, and the ADC sampling rate can be adjusted according to different needs, achieving high-fidelity digital acquisition of multimodal physiological signals such as EEG, ECG, EMG, and ESK. Through this module, the system can stably acquire high-quality physiological data, providing a reliable data foundation for subsequent signal synchronization, analysis, and regulation.

[0104] This embodiment may also include a signal synchronization and alignment module, used to achieve synchronous acquisition of multi-channel physiological signals between the host and multiple slave devices, or other external physiological monitoring devices. In this embodiment, this module achieves high-precision time alignment of multimodal physiological signals from different devices through timestamp synchronization, external trigger signals, or master-slave communication protocols, thereby ensuring the consistency of each signal channel in the time dimension and providing a reliable time reference for subsequent multimodal data fusion and joint analysis. This module supports multiple communication interfaces (such as USB, SPI, I²C, Bluetooth, Wi-Fi, etc.) and synchronization mechanisms, is compatible with data access methods of various external devices, and achieves collaborative acquisition and signal synchronization among multiple devices based on a unified time reference or external trigger signals, improving the overall performance and analytical accuracy of the system in cross-device and cross-modal signal processing.

[0105] This embodiment may also include a physiological state regulation module, used to intervene in the target physiological system through a specific form of energy medium to achieve active and adaptive regulation of the user's physiological state. In this embodiment, the module can automatically or manually trigger corresponding regulation mechanisms based on the analysis results of the collected physiological signals. During the regulation process, the module outputs energy in the form of light, electricity, magnetism, sound, heat, etc., to have a directional effect on the nervous system, circulatory system, muscle tissue, etc., thereby achieving, for example, relaxation training by guiding breathing frequency through light stimulation, improving attention by regulating brainwave activity using sound waves and low-frequency electrical stimulation, improving sleep quality by combining light therapy and breathing rhythm intervention, or relieving local muscle pain by combining electrical stimulation and thermotherapy. In addition, the module can also control the parameters of other regulation devices through a signal synchronization alignment module to achieve better regulation effects. This module can be linked with the intelligent analysis platform in the system to form a closed-loop control mechanism of "monitoring-analysis-feedback-intervention", significantly improving the practicality and intelligence level of the device in application scenarios such as health management, neurofeedback training, personalized rehabilitation, and adaptive regulation of exercise efficiency.

[0106] Example 2

[0107] This invention also provides a reusable multimodal physiological intelligent monitoring system that can be used across different scenarios, such as... Figure 3 As shown, it includes:

[0108] The aforementioned multimodal physiological intelligent monitoring equipment; and

[0109] The intelligent physiological signal analysis platform provides an external program interface to perform personalized and custom analysis and application of processed physiological signals.

[0110] The intelligent physiological signal analysis platform includes:

[0111] The host connection and control platform is used for bidirectional communication with multimodal physiological intelligent monitoring equipment to acquire data and send commands;

[0112] A signal decoding platform used to reconstruct physiological electrical signals through specific decoding methods;

[0113] The middle layer is used for alignment, standardization, and resource scheduling, enabling the unified management and use of physiological electrical signals of different frequencies and characteristics.

[0114] The signal analysis plugin platform is used to upload and download user-defined external programs. These external programs call physiological electrical signals through an intermediate layer to achieve targeted analysis and personalized applications of various physiological signals.

[0115] The host connection and control platform enables bidirectional communication between the intelligent analysis platform and the physiological measurement device host, supporting real-time acquisition of physiological data and issuance of control commands. This ensures efficient collaboration between the system and the data acquisition, analysis, and feedback regulation processes. In this embodiment, the platform uses multiple communication protocols to achieve data interaction with the host, supporting various connection methods including wireless communication (such as Bluetooth, Wi-Fi, Zigbee) and wired or dedicated communication interfaces (such as USB, UART, ANT+). It can flexibly select communication channels based on usage scenarios and device types to improve system compatibility and connection stability. Through this platform, the intelligent analysis system can actively acquire collected multimodal physiological signals and send control commands to the host based on the analysis results. These commands dynamically adjust signal acquisition parameters, activate physiological state regulation modules, or trigger specific intervention mechanisms, thereby achieving closed-loop management from data acquisition to intelligent regulation and significantly improving the system's real-time performance and intelligence level.

[0116] The signal decoding platform is used to parse and restore physiological signal data received from the host and transmitted via wireless or wired communication protocols, converting it into raw electrophysiological signals suitable for subsequent analysis. In this embodiment, the platform is compatible with various communication transmission methods, supporting efficient decoding of physiological data transmitted via protocols such as Bluetooth, Wi-Fi, Zigbee, USB, and UART. The platform integrates multiple protocol parsing engines, automatically identifying and loading corresponding decoding algorithms based on the communication interface type, ensuring that the decoded signals meet the requirements of subsequent analysis in terms of time synchronization, sampling accuracy, and channel consistency. Through this platform, the system can uniformly convert physiological signals from heterogeneous sources into a structured and standardized data format, providing stable and reliable data input for the signal analysis plug-in platform, thereby achieving efficient fusion and in-depth analysis of multimodal signals.

[0117] The signal analysis plug-in platform is used to build an open and extensible physiological signal analysis ecosystem, which supports users, developers or third-party institutions to upload, download, deploy and run custom physiological signal analysis plug-ins, so as to achieve personalized analysis of multi-modal physiological signals and expansion of diverse application scenarios. The platform provides standardized plug-in interfaces and operating environments, allowing users to develop and deploy plug-in programs based on various collected signals (such as electroencephalogram, electrocardiogram, electromyogram, galvanic skin response, respiratory signal, acceleration signal, etc.) to implement personalized analysis tasks including signal processing, feature extraction, pattern recognition, state modeling, linkage control, etc. Users are not limited to the analysis logic preset by the system, but can freely construct analysis processes according to their own needs to meet the application requirements in different fields and scenarios. Further, the signal analysis plug-in platform also supports the application market model and builds an open plug-in sharing and distribution ecosystem. Users can upload the self-developed analysis plug-ins to the platform, and after security verification and version management, publish them to the plug-in market. Other users can browse, download and deploy them according to their needs to achieve the sharing and reuse of plug-in resources. This application market model not only supports individual developers to publish plug-ins, but also supports scientific research institutions, medical institutions, enterprises and third-party platforms to access, forming an open ecological system covering multiple fields such as physiological signal analysis, brain-computer interface, emotion recognition, sports training, and health management. The platform can also provide functions such as plug-in scoring, version update, permission management, usage statistics, etc., to enhance the manageability of plug-ins and user experience. Through this plug-in platform, the system realizes a full-link open architecture from signal acquisition, data transmission, signal decoding, intelligent analysis to function expansion, builds a technical ecosystem that supports multi-dimensional physiological data calling and multi-scenario intelligent application development, and significantly improves the flexibility, functional scalability and long-term use value of the system.

[0118] The present invention is based on the innovative idea of "1 + n + ∞" (1 host + n accessories / scenarios + ∞ analysis algorithms). By multiplexing hardware and algorithms, it integrates the main functions of wearable physiological monitoring devices such as smart bracelets, smart watches, smart headbands, portable electrocardiographs, and respiratory straps, solves the problems of single product form and limited application of current wearable products, effectively expands the usage scenarios of single devices, and meets the personalized development and customization requirements of each link of physiological signal acquisition, analysis, feedback, and regulation.

[0119] It should be understood that the above embodiments and the descriptions in the specification only illustrate the principles, main features and advantages of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the protection scope of the present invention.

Claims

1. A multimodal physiological intelligent monitoring device that can be reused across different scenarios, characterized in that, include: A multimodal physiological data acquisition sensor array includes several arrayed sensing units, each of which includes sensing electrodes, pressure electrodes, a sound acquisition unit, and an inertial measurement unit. The acquisition and control module is used to control the sensing electrodes to acquire the physiological electrical signals spontaneously generated by the human body, and to control the pressure electrodes to generate an excitation current that is input into the human skin, and to acquire the physiological electrical signals generated by the human body under the excitation current through the sensing electrodes. The placement detection module is used to calculate the comprehensive confidence level of the current physiological electrical signal belonging to each physiological signal type from the signals collected by the multimodal physiological acquisition sensor array, thereby determining the physiological signal type to which the physiological electrical signal collected by the sensing electrode belongs. The physiological electrical signal types include electrocardiogram (ECG), electroencephalogram (EEG), electromyogram (EMG), electrodermal signal, and respiratory signal. Respiratory signal and EMG are physiological electrical signals obtained by the human body under excitation current, while ECG, EEG, and EMG are physiological electrical signals spontaneously generated by the human body. An adaptive parameter adjustment module is used to generate the sampling rate, amplification factor, filter bandwidth, and ADC bit depth corresponding to the current physiological signal type; The signal processing module is used to amplify the signal according to the amplification factor, perform analog-to-digital conversion according to the sampling rate and ADC bit depth, filter the signal according to the filtering bandwidth, and encode and compress the signal according to the encoding and compression strategy to obtain the processed physiological electrical signal.

2. The cross-scenario reusable multimodal physiological intelligent monitoring system according to claim 1, characterized in that, The multimodal physiological data acquisition sensor array acquires data by directly covering the skin surface of the measurement site on the human body with a stretchable flexible strap or adhesive electrode pad, or through a signal transmission entity that can be connected to the human body.

3. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 1, characterized in that, The placement detection module includes: The preprocessing unit is used to preprocess the physiological electrical signals collected by the sensing electrodes in the multimodal physiological acquisition sensor array, the sound signals collected by the sound acquisition unit, and the inertial signals collected by the inertial measurement unit. The feature calculation unit is used to perform feature calculations on each signal within a preset window time after preprocessing. Specifically, it includes: calculating the waveform characteristics and contact impedance of the physiological electrical signal based on the physiological electrical signal; calculating the energy spectrum and envelope of the sound signal based on the sound signal; and calculating the attitude angle of the current position based on the inertial signal. The position determination unit is used to determine the current acquisition position based on the attitude angle, and to determine the position confidence of the current physiological electrical signal belonging to each physiological signal type based on the current acquisition position; The contact determination unit is used to determine the contact confidence level of the current physiological electrical signal belonging to each physiological signal type based on the contact impedance. The waveform feature determination unit is used to determine the signal attribute confidence level of the physiological electrical signal belonging to each physiological signal type based on the waveform features of the physiological electrical signal. The sound determination unit is used to determine whether a sound signal belongs to the heart sound component or the pulse component based on the energy spectrum and envelope of the sound signal, and to determine the sound attribute confidence of the current physiological electrical signal belonging to each physiological signal type based on the attributes of the sound signal. The type determination unit is used to perform weighted fusion of location confidence, contact confidence, signal attribute confidence and sound attribute confidence to obtain a comprehensive confidence score, and select the type with the highest comprehensive confidence score as the physiological signal type.

4. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 3, characterized in that, The waveform feature determination unit specifically includes: The ECG signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal to be the ECG signal when a periodic narrow QRS peak is detected in the waveform of the physiological electrical signal, and an SCG peak appears after the R wave in the QRS peak. The EEG signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal to be an EEG signal when the waveform of the physiological electrical signal does not have QRS peaks, and the low-frequency energy is dominant and the amplitude is in the μV range. The electromyography signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as an electromyography signal to be the highest when the high-frequency energy of the physiological electrical signal is dominant and appears in the form of a short burst. The skin conductance signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as a skin conductance signal when the spectral energy of the contact impedance signal is mainly concentrated in the low-frequency part and the signal consists of a slowly changing baseline and a rapid transient response. The respiratory signal determination subunit is used to set the signal attribute confidence of the physiological electrical signal as a respiratory signal to be the highest when the low-frequency envelope of the contact impedance is synchronized with the envelope of the sound signal and has periodicity in the respiratory frequency band.

5. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 3, characterized in that, The sound determination unit specifically includes: The first determination subunit is used to determine that the current sound signal belongs to the heart sound component when the energy spectrum of the sound signal detects the S1 / S2 energy peak and the sound signal envelope detects the short-term heart sound envelope. Then, the confidence level of the physiological electrical signal belonging to the signal attribute of the electrocardiogram signal and the respiratory signal is set to be the highest. The second determination subunit is used to determine that the current sound signal belongs to the pulse component when the sound signal envelope detects the pulse envelope, and to set the signal attribute of the physiological electrical signal to have the highest confidence level as belonging to electromyography and electrodermal signals.

6. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 1, characterized in that, The adaptive sampling adjustment module specifically includes: The modality determination unit is used to determine the current acquisition modality M* based on the comprehensive confidence level of the current physiological signal. The acquisition modality includes ECG acquisition modality, EEG acquisition modality, EMG acquisition modality, SCTA acquisition modality, and respiratory impedance acquisition modality. The short-time observation unit is used to measure and estimate the following observations based on physiological signals: short-time signal amplitude estimation. rms Motion index M and available bandwidth BW avail ; The target bandwidth and sampling rate calculation unit is used to determine the filtering bandwidth based on the nominal bandwidth of the current acquisition mode M*, the overall confidence level, and the motion index M. : ， ， in, As a bandwidth security factor, it is determined by querying the preset parameters. The mapping table between M and bandwidth security factor is obtained. This represents the overall confidence level of the current acquisition mode M*. Indicates the sampling rate. The nominal bandwidth of the dominant mode M* is determined by the inherent spectral width occupied by the physiological signal corresponding to the dominant mode M*. Indicates the oversampling factor; Amplification factor and ADC bit depth configuration unit, used to estimate signal based on short-time amplitude. rms Estimating short-time peak-to-peak value and based on and ADC full amplitude Calculate the amplification factor G, and calculate the ADC bit depth b based on the target signal-to-noise ratio: ， ， in, This refers to the clip function. Indicates the target occupancy percentage. , Indicates the minimum and maximum magnification; Indicates the target signal-to-noise ratio. It represents rounding up; The encoding and compression strategy setting unit is used to calculate the original code rate based on the ADC bit depth and sampling rate. And based on available bandwidth With energy budget EB, the original bitrate R raw Mapped to target bitrate This allows us to determine the required compression ratio. And select an encoding compression strategy with a compression ratio of CR; The parameter sending unit is used to send the magnification factor G and sampling rate. The ADC bit depth (b) and encoding compression strategy are respectively distributed to the amplifier, analog-to-digital converter, and encoder.

7. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 6, characterized in that, The sensing unit also includes a photoplethysmography (PPG) sensor, which is used to acquire pulse wave signals. The acquisition modes also include pulse wave acquisition mode, heart sound acquisition mode, and breath sound acquisition mode. When the photoplethysmography pulse wave sensor acquires a pulse wave signal and the physiological electrical signal is lower than a preset threshold, the current acquisition mode is determined to be pulse wave acquisition mode. When the sound acquisition device acquires a sound signal that is a heart sound, the current acquisition mode is determined to be heart sound acquisition mode. When the sound acquisition device acquires a sound signal that is a breath sound, the current acquisition mode is determined to be breath sound acquisition mode.

8. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 1, characterized in that, The sensing unit also includes a temperature sensor and an ambient light sensor. The temperature sensor is used to collect skin temperature or ambient temperature, and the ambient light sensor is used to collect the intensity of external light.

9. The multimodal physiological intelligent monitoring device that can be reused across scenarios according to claim 1, characterized in that, The adaptive sampling adjustment module further includes: The short-time acknowledgment and quality inspection unit is used to calculate the real-time signal-to-noise ratio (SNR) within the short-time acknowledgment window. est Key feature detection rate and reconstruction error ε, if SNR est If the values ​​are greater than the threshold and the key feature detection rate is greater than the threshold, then accept the current parameter configuration and continue running; otherwise, execute the closed-loop adjustment unit. The closed-loop adjustment unit is used to adjust according to the following priority order: (1) Increase the magnification; (2) Increase the sampling rate; (3) Reduce compression intensity or switch to near-lossless / lossless encoding; (4) If the above measures fail to improve the situation, the user should be prompted to reposition the device or proceed with manual calibration. The event-driven switching unit is used to immediately switch the device to event-driven mode when a predefined abnormal event is identified during normal data processing. This temporarily increases the sampling rate and reduces the compression intensity to record lossless or near-lossless data for the event period. After the event ends, the system returns to normal parameters.

10. A multimodal physiological intelligent monitoring system that can be reused across different scenarios, characterized in that, include: The multimodal physiological intelligent monitoring device as described in claim 1; A physiological signal intelligent analysis platform is provided to offer an external program interface for personalized and customizable analysis and application of processed physiological signals. The physiological signal intelligent analysis platform includes: The host connection and control platform is used for bidirectional communication with multimodal physiological intelligent monitoring equipment to acquire data and send commands; A signal decoding platform used to reconstruct physiological electrical signals through specific decoding methods; The middle layer is used for alignment, standardization, and resource scheduling, enabling the unified management and use of physiological electrical signals of different frequencies and characteristics. The signal analysis plugin platform is used to upload and download user-defined external programs. These external programs call physiological electrical signals through an intermediate layer to achieve targeted analysis and personalized applications of various physiological signals.

Images