A multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system and method

Abstract

The application discloses a multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system, comprising a processor and a memory. When the executable instructions of the processor are executed by the processor, the processor performs the following operations: in a reference calibration period, a first reference response and a second reference response are acquired; a first correction parameter is determined according to the first reference response, and a second correction parameter is determined according to the second reference response; in an object measurement period, a first object impedance response and a second object impedance response are acquired; the first object impedance response and the second object impedance response are respectively subjected to synchronous demodulation to obtain a first impedance change signal and a second impedance change signal; the first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter; and an output parameter is generated according to the corrected first impedance change signal and the corrected second impedance change signal.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a real-time blood pressure monitoring system and method that integrates multi-frequency impedance and pulse wave fusion. Background Technology

[0002] With the development of continuous health monitoring, telemedicine, and wearable medical devices, real-time, cuffless, or low-burden blood pressure monitoring technology has gradually become a focus of research and application. Traditional cuff-based blood pressure measurement typically uses intermittent compression, making it difficult to continuously acquire blood pressure changes over long periods and potentially causing discomfort. Therefore, existing technologies are increasingly employing signals such as pulse waves, electrocardiogram (ECG) signals, photoplethysmography (PPG), and bioimpedance to continuously estimate blood pressure. Among these, blood pressure monitoring methods based on pulse wave transit time, pulse arrival time, or pulse wave morphology parameters typically establish a correspondence between pulse wave characteristics and blood pressure through individualized calibration models. For example, CN1931088A discloses a technical solution for detecting and calibrating blood pressure monitors using pulse wave transit time and Korotkoff sound delay time.

[0003] Bioimpedance measurement technology, by applying AC excitation to a subject and detecting its impedance response, can acquire information related to tissue electrical properties, blood volume changes, and vascular pulsation. Compared to a single excitation frequency, multi-frequency impedance measurement can obtain the frequency response of the subject at different excitation frequencies and extract richer physiological characteristics through resistance spectrum analysis or impedance modeling. Existing technologies disclose processing procedures such as multi-frequency impedance scanning, resistance spectrum analysis, and tissue impedance model construction. For example, CN117814775B discloses a scheme for tissue impedance analysis and physiological parameter processing using multi-frequency impedance data. In relevant blood pressure monitoring systems, fusing multi-frequency impedance change signals with pulse wave signals is beneficial for utilizing the complementarity of different signals in terms of blood flow changes, vascular elasticity, and tissue electrical characteristics, thereby improving the continuity and stability of blood pressure estimation.

[0004] However, multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring systems typically include multiple excitation, acquisition, amplification, filtering, analog-to-digital conversion, and digital processing channels. Due to differences in the electronic device parameters, gain, phase response, sampling clock, and signal transmission path of each channel, the impedance response acquired at different excitation frequencies may exhibit amplitude deviations, phase deviations, and channel delays. If impedance change signals at different frequencies are directly fused, channel differences may be mistakenly interpreted as physiological changes in the subject, leading to pulse wave morphology distortion, feature extraction errors, and fluctuations in blood pressure estimation results. Existing bioimpedance devices typically use reference impedance or measurement standards for calibration. For example, US12048519B2 discloses a technical solution for calibrating bioimpedance measurement devices using measurement standards and generating correction data based on the deviation between the measurement results and the standard values. However, this type of calibration usually focuses on overall calibration before device manufacturing or use, making it difficult to dynamically compensate for amplitude, phase, and time differences between different frequency channels during real-time measurement.

[0005] Furthermore, the actual excitation current may fluctuate due to changes in power supply, excitation circuit status, or load, causing the amplitude of the impedance response to vary with the measurement time. Without amplitude normalization by combining the actual excitation current and the reference response, impedance change signals obtained at different measurement periods may lack comparability. The contact state between the object and the electrode, as well as the object's impedance baseline, may also be affected by the tightness of the fit, skin condition, movement posture, and electrode position, resulting in contact impedance errors and baseline drift. CN104095631B points out that inconsistent contact areas between the electrode and the measured object may introduce contact impedance errors and affect the accuracy of bioimpedance testing. Therefore, relying solely on initial calibration parameters is insufficient to guarantee measurement consistency during long-term real-time monitoring.

[0006] On the other hand, some existing physiological signal monitoring technologies validate the measurement results and remove abnormal data when validation fails. For example, CN115916044A discloses a technical solution for validating impedance measurement results and retaining or removing relevant measurement results based on the validation results. However, existing solutions typically struggle to evaluate the quality of different frequency channels and dynamically determine fusion weights based on changes in reference response between adjacent reference calibration periods; they also struggle to promptly reacquire the reference response when there are significant changes in the contact state between the object and the electrode or when the object's impedance baseline changes. In cases of reduced consistency between different frequency channels, existing systems may continue to output blood pressure parameters or only save processed feature data, making it impossible to trace the object's impedance response before and after the abnormal event, and making it difficult to distinguish whether the abnormality was caused by physiological changes in the object, decreased channel performance, changes in actual excitation current, or abnormal electrode contact.

[0007] Therefore, existing technologies still require a multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system that can use reference impedance response to correct the amplitude, phase, and time of different frequency channels during real-time blood pressure monitoring, combine actual excitation current to achieve amplitude normalization, and dynamically evaluate measurement quality, adjust fusion weights, trigger recalibration, and save original data related to abnormal events based on changes in reference response and consistency between channels. Summary of the Invention

[0008] The purpose of this invention is to provide a multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system. During the reference calibration period, the amplitude correction parameters, phase correction parameters, and channel delay parameters for each frequency channel are determined using pre-determined reference impedance responses at different excitation frequencies. During the object measurement period, the object impedance responses acquired at different frequencies are synchronously demodulated, and amplitude correction, phase correction, and time compensation are performed. Furthermore, amplitude normalization is performed based on the actual excitation current and reference response; channel quality and corresponding weights are determined based on response changes between adjacent reference calibration periods; and the corrected impedance change signals are weighted and combined. Recalibration is performed when the contact state parameter or object impedance baseline change exceeds a threshold, and abnormal events are identified and the relevant original responses are saved when the consistency between channels is below a threshold.

[0009] The above-mentioned technical means can compensate for the amplitude, phase and time differences between different frequency channels, reduce the impact of actual excitation current changes, electrode contact changes and baseline drift on monitoring results, and improve the consistency, stability and reliability of multi-frequency impedance signals and blood pressure parameter estimation. At the same time, channel quality weighting reduces the impact of abnormal or low-quality channels on output results and retains the original data before and after abnormal events, enhancing the ability of anomaly analysis and measurement result traceability.

[0010] This invention provides a real-time blood pressure monitoring system that integrates multi-frequency impedance and pulse wave fusion, comprising: The processor and memory; the memory stores processor-executable instructions, which, when executed by the processor, cause the processor to perform the following operations: During the reference calibration period, a first reference response and a second reference response are acquired. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of a predetermined reference impedance at a second excitation frequency. The first excitation frequency and the second excitation frequency are different. A first correction parameter is determined based on a first reference response, and a second correction parameter is determined based on a second reference response; During the object measurement period, a first object impedance response and a second object impedance response are acquired. The first object impedance response corresponds to a first excitation frequency, and the second object impedance response corresponds to a second excitation frequency. The impedance responses of the first and second objects are synchronously demodulated to obtain the first impedance change signal and the second impedance change signal. The first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter. Output parameters are generated based on the corrected first impedance change signal and the corrected second impedance change signal.

[0011] In a preferred embodiment, the first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter, and the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter; The correction of the first impedance change signal includes: The amplitude correction, phase correction, and time compensation are performed on the first impedance change signal; Correcting the second impedance change signal includes: The amplitude, phase, and time of the second impedance change signal are corrected.

[0012] In a preferred embodiment, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: The quality of the first frequency channel and the quality of the second frequency channel are determined based on the change in the first reference response and the change in the second reference response between adjacent reference calibration periods, respectively. The first weight and the second weight are determined based on the quality of the first frequency channel and the quality of the second frequency channel, respectively. Based on the corrected first impedance change signal and the corrected second impedance change signal, output parameters are generated, including: Based on the first weight and the second weight, the corrected first impedance change signal and the corrected second impedance change signal are weighted and combined to generate output parameters.

[0013] In a preferred embodiment, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: Obtain contact state parameters between the object and the electrode or baseline parameters of the object impedance; When the change in the contact state parameter or the object impedance baseline parameter exceeds the first preset threshold, the first reference response and the second reference response are reacquired.

[0014] In a preferred embodiment, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: The first actual excitation current at the first excitation frequency and the second actual excitation current at the second excitation frequency are obtained respectively. The amplitude of the first impedance change signal and the second impedance change signal is normalized based on the first actual excitation current, the first reference response, the second actual excitation current and the second reference response.

[0015] In a preferred embodiment, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: Store the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter; When the consistency between the first impedance change signal and the second impedance change signal is detected to be lower than a second preset threshold, an abnormal event is determined to have occurred. The first and second object impedance responses are acquired within preset time periods before and after the occurrence of the abnormal event.

[0016] This invention also provides a method for real-time blood pressure monitoring using multi-frequency impedance and pulse wave fusion, comprising: During the reference calibration period, a first reference response and a second reference response are acquired. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of a predetermined reference impedance at a second excitation frequency. The first excitation frequency and the second excitation frequency are different. A first correction parameter is determined based on a first reference response, and a second correction parameter is determined based on a second reference response; During the object measurement period, a first object impedance response and a second object impedance response are acquired. The first object impedance response corresponds to a first excitation frequency, and the second object impedance response corresponds to a second excitation frequency. The impedance responses of the first and second objects are synchronously demodulated to obtain the first impedance change signal and the second impedance change signal. The first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter. Output parameters are generated based on the corrected first impedance change signal and the corrected second impedance change signal.

[0017] In a preferred embodiment, the first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter, and the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter; The correction of the first impedance change signal includes: The amplitude correction, phase correction, and time compensation are performed on the first impedance change signal; Correcting the second impedance change signal includes: The amplitude, phase, and time of the second impedance change signal are corrected.

[0018] In a preferred embodiment, it further includes: The quality of the first frequency channel and the quality of the second frequency channel are determined based on the change in the first reference response and the change in the second reference response between adjacent reference calibration periods, respectively. The first weight and the second weight are determined based on the quality of the first frequency channel and the quality of the second frequency channel, respectively. Based on the corrected first impedance change signal and the corrected second impedance change signal, output parameters are generated, including: Based on the first weight and the second weight, the corrected first impedance change signal and the corrected second impedance change signal are weighted and combined to generate output parameters.

[0019] In a preferred embodiment, it further includes: Obtain contact state parameters between the object and the electrode or baseline parameters of the object impedance; When the change in the contact state parameter or the object impedance baseline parameter exceeds the first preset threshold, the first reference response and the second reference response are reacquired.

[0020] Compared with the prior art, the present invention has the following advantages: The present invention can compensate for the amplitude, phase and time differences between different frequency channels, reduce the impact of actual excitation current changes, electrode contact changes and baseline drift on monitoring results, and improve the consistency, stability and reliability of multi-frequency impedance signals and blood pressure parameter estimation; at the same time, by reducing the impact of abnormal or low-quality channels on output results through channel quality weighting, and retaining the original data before and after abnormal events, it enhances the ability of anomaly analysis and measurement result traceability. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the application scenario of the multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system provided in this embodiment of the invention.

[0022] Figure 2 The logic block diagram of the multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring device provided in the embodiments of the present invention is shown.

[0023] Figure 3 A block diagram of a computer system provided in an embodiment of the present invention.

[0024] Figure 4 This is a flowchart of a method provided in an embodiment of the present invention.

[0025] Figure 5 This is a schematic view of the system operation process of the present invention. Detailed Implementation

[0026] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0027] like Figure 1 As shown, the multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system of this embodiment is applied to a scenario of real-time, continuous blood pressure monitoring of a monitoring subject 105. The monitoring subject 105 wears a wrist-worn blood pressure monitoring device 110. The wrist-worn blood pressure monitoring device 110 has an impedance measurement electrode group 115 and a pulse wave sensor 120 disposed on the side facing the wrist skin. A processing module 125 is disposed inside the device housing. The processing module 125 includes a processor and a memory. The memory stores processor-executable instructions, and the processor and memory can perform some functions of this invention. The impedance measurement electrode group 115 contacts the wrist skin of the monitoring subject 105 to apply an excitation signal to the wrist tissue and acquire the corresponding impedance response. The pulse wave sensor 120 is, for example, a photoplethysmography pulse wave sensor, which emits detection light (such as...) Figure 1 (As shown by the dashed line) The light is injected into the wrist tissue and the reflected light is received, thereby obtaining the pulse wave signal.

[0028] During operation, the processing module 125 first controls the impedance measurement electrode group 115 to measure the predetermined reference impedance at the first excitation frequency and the second excitation frequency during the reference calibration period, respectively, to obtain the first reference response and the second reference response. Based on these responses, the processing module 125 determines the first correction parameter and the second correction parameter, which include amplitude correction parameter, phase correction parameter, and channel delay parameter. Subsequently, during the object measurement period, the processing module 125 obtains the first object impedance response at the first excitation frequency and the second object impedance response at the second excitation frequency, and performs synchronous demodulation on both responses to obtain the first impedance change signal and the second impedance change signal. Then, the processing module 125 uses the first correction parameter and the second correction parameter to perform amplitude correction, phase correction, and time compensation on the two impedance change signals, respectively. Optionally, the processing module 125 also obtains the actual excitation current at the two excitation frequencies and combines it with the corresponding reference response to normalize the amplitude of the two impedance change signals to eliminate errors introduced by excitation source fluctuations.

[0029] Furthermore, the processing module 125 evaluates the quality of the first frequency channel and the second frequency channel respectively based on the changes in the two reference responses between adjacent reference calibration periods, and determines the first weight and the second weight accordingly. It then performs a weighted combination of the two corrected impedance change signals and fuses the weighted combination result with the pulse wave signal collected by the pulse wave sensor 120. For example, it calculates blood pressure based on impedance change characteristics and pulse wave conduction characteristics to generate output parameters, which may include systolic blood pressure, diastolic blood pressure, mean arterial pressure, and pulse rate. The wrist-worn blood pressure monitoring device 110 transmits the output parameters to the smart terminal 135 in real time via a wireless transmission signal 130 (e.g., Bluetooth, Wi-Fi). The screen of the smart terminal 135 displays the blood pressure value and pulse waveform. The smart terminal 135 can also upload the monitoring data to the hospital workstation 140 for long-term storage, trend analysis, and remote monitoring.

[0030] In addition, the processing module 125 also acquires the contact state parameters or object impedance baseline parameters between the monitored object 105 and the impedance measurement electrode group 115 in real time. When the change in the above parameters exceeds the first preset threshold (e.g., due to loosening of the fit or changes in sweat causing a change in the contact state), it automatically re-performs the reference calibration and re-acquires the first reference response and the second reference response to ensure measurement accuracy. The processing module 125 continuously stores the two impedance change signals and the corresponding correction parameters. When the consistency between the first impedance change signal and the second impedance change signal is detected to be lower than the second preset threshold, an abnormal event is determined to have occurred, and the first object impedance response and the second object impedance response acquired in each preset time period before and after the occurrence of the abnormal event are stored for subsequent traceability analysis.

[0031] like Figure 2 As shown, the multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring device 200 of this embodiment includes the following logical functions: a reference calibration module 205, a calibration parameter determination module 210, an impedance acquisition module 215, a synchronous demodulation module 220, a signal correction module 225, a channel quality assessment module 230, a weighted fusion module 235, an amplitude normalization module 240, a calibration trigger module 245, an anomaly detection module 250, and a data storage module 255. This device 200 can be implemented as follows: Figure 1 The logical functional architecture of the processing module 125 in the wrist-worn blood pressure monitoring device 110 shown.

[0032] The reference calibration module 205 is used to acquire a first reference response and a second reference response during the reference calibration period. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of the same reference impedance at a second excitation frequency, wherein the first excitation frequency and the second excitation frequency are different. The correction parameter determination module 210 is used to determine a first correction parameter based on the first reference response and a second correction parameter based on the second reference response; wherein the first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter, and the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter.

[0033] Impedance acquisition module 215 is used to acquire a first object impedance response corresponding to a first excitation frequency and a second object impedance response corresponding to a second excitation frequency during the object measurement period. Synchronous demodulation module 220 is connected to impedance acquisition module 215 and is used to synchronously demodulate the first and second object impedance responses to obtain a first impedance change signal and a second impedance change signal. Signal correction module 225 is connected to synchronous demodulation module 220 and correction parameter determination module 210 respectively, and is used to perform amplitude correction, phase correction, and time compensation on the first impedance change signal according to a first correction parameter, and perform amplitude correction, phase correction, and time compensation on the second impedance change signal according to a second correction parameter.

[0034] The channel quality assessment module 230 is connected to the reference calibration module 205 and is used to determine the quality of the first frequency channel and the quality of the second frequency channel based on the changes in the first reference response and the second reference response between adjacent reference calibration periods, respectively. The weighted fusion module 235 is connected to both the signal correction module 225 and the channel quality assessment module 230. It is used to determine the first weight and the second weight based on the quality of the two frequency channels, and to perform a weighted combination of the two corrected impedance change signals according to the weights to generate output parameters. Optionally, the weighted fusion module 235 further fuses the weighted combination result with the pulse wave signal to calculate blood pressure parameters such as systolic blood pressure and diastolic blood pressure.

[0035] The amplitude normalization module 240 is used to acquire the first actual excitation current at the first excitation frequency and the second actual excitation current at the second excitation frequency, and to normalize the amplitude of the two impedance change signals based on the first actual excitation current, the first reference response, the second actual excitation current, and the second reference response, so as to eliminate the influence of excitation current fluctuations on the measurement results. The calibration trigger module 245 is connected to the reference calibration module 205 and is used to acquire the contact state parameters between the object and the electrode or the object impedance baseline parameters. When the change in the above parameters exceeds a first preset threshold, the reference calibration module 205 is triggered to reacquire the first reference response and the second reference response to realize the automatic update of calibration parameters.

[0036] The anomaly detection module 250 is used to detect the consistency between the first impedance change signal and the second impedance change signal, and determines that an anomaly event has occurred when the consistency is lower than a second preset threshold. The data storage module 255 is connected to the above modules and is used to store the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter; and when the anomaly detection module 250 determines that an anomaly event has occurred, it stores the first object impedance response and the second object impedance response acquired in preset time periods before and after the time of the anomaly event, for subsequent traceability and analysis.

[0037] It should be noted that the division of modules in the above-mentioned device 200 is only a logical functional division. In actual implementation, each module can be implemented as hardware, software, firmware, or any combination thereof. For example, each module can be implemented as a hardware circuit or a functional unit therein, such as an Application-Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), or a Microcontroller (MCU); it can also be implemented as computer program instructions stored in memory and executed by a processor; it can also be implemented as firmware embedded in hardware; or it can be implemented by several combinations of the above forms. The above-mentioned modules can be integrated into the same processor or the same chip, or they can be distributed and implemented in multiple physical entities; each module can be further merged into one module, or any module can be split into multiple sub-modules. Those skilled in the art should understand that, as long as they do not depart from the concept defined by the claims of this invention, all variations in the physical implementation, integration level, and division granularity of the above-mentioned modules should fall within the protection scope of this invention.

[0038] like Figure 3 As shown, this embodiment provides a computer system 300, which includes the processing module 125 of the wrist-worn blood pressure monitoring device 110 in the aforementioned embodiment, or... Figure 2All the logical functions of the illustrated device 200 can be implemented on the computer system 300. The computer system 300 includes a bus 310, and interconnected components via the bus 310, including a central processing unit 305, main memory 315, read-only memory 320, mass storage device 325, impedance measurement interface 330, and pulse wave sensor interface 335. The computer system 300 can also be coupled via the bus 310 to peripheral interfaces such as a wireless communication interface 340 and a display interface 345. The bus 310 can be a bus of various industry standard architectures, used for transmitting data, instructions, and control signals between the aforementioned components.

[0039] The central processing unit 305 executes processor-executable instructions stored in the memory to implement the various operations in the foregoing embodiments, including: acquiring the first and second reference responses during the reference calibration period, determining the calibration parameters, synchronously demodulating the object impedance response, amplitude correction of the impedance change signal, phase correction and time compensation, frequency channel quality assessment and weighted combination, amplitude normalization, recalibration triggering, and abnormal event detection. The central processing unit 305 can be, for example, a microcontroller (MCU), a digital signal processor (DSP), a general-purpose microprocessor, or a combination of the above devices.

[0040] Main memory 315, such as static random access memory (SRAM) or dynamic random access memory (DRAM), is used to temporarily store intermediate data during system operation, such as the first impedance change signal and the second impedance change signal obtained through synchronous demodulation. Read-only memory 320, such as flash memory or electrically erasable programmable read-only memory (EEPROM), is used to store the boot program and the fixed signal processing and calibration program. Mass storage device 325, such as flash memory, is used to store the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter; and when an abnormal event is determined to occur, it stores the first object impedance response and the second object impedance response obtained within each preset time period before and after the occurrence of the abnormal event for subsequent tracing and analysis.

[0041] Impedance measurement interface 330 is used to connect to impedance measurement electrode group 115, which includes an excitation signal generation circuit, a demodulation circuit, and an analog-to-digital conversion circuit. It is used to apply excitation current to the monitored object at a first excitation frequency and a second excitation frequency, respectively, to acquire the reference response during the reference calibration period and the object impedance response during the object measurement period, and to obtain the actual excitation current at each excitation frequency, as well as the contact state parameters between the object and the electrodes or the object impedance baseline parameters. Pulse wave sensor interface 335 is used to connect to pulse wave sensor 120, to acquire pulse wave signals, and to transmit them to bus 310.

[0042] The wireless communication interface 340, such as Bluetooth, Wi-Fi, or a cellular communication interface, is used to send the generated output parameters to the smart terminal 135 in real time, and optionally upload them to the hospital workstation 140 via the smart terminal 135. The display interface 345 is used to drive the display screen to display blood pressure parameters such as systolic and diastolic pressure, as well as pulse waveforms locally.

[0043] It should be noted that the aforementioned computer system 300 can be integrated entirely into the wrist-worn blood pressure monitoring device 110, or some of its functions can be distributed and deployed on the smart terminal 135 or the hospital workstation 140; the types of the aforementioned memory, the architecture of the bus, and the number and types of peripheral interfaces can all be increased, decreased, or replaced according to actual applications. Those skilled in the art should understand that, as long as they do not depart from the concept defined by the claims of this invention, all the above changes should fall within the protection scope of this invention.

[0044] Figure 4 A flowchart illustrating a method according to an embodiment of the present invention. As shown in the figure, the method of the present invention includes the following steps: Step 1: During the reference calibration period, obtain the first reference response and the second reference response. The first reference response corresponds to the response of a predetermined reference impedance at the first excitation frequency, and the second reference response corresponds to the response of a predetermined reference impedance at the second excitation frequency. The first excitation frequency and the second excitation frequency are different. Step 2: Determine the first correction parameter based on the first reference response, and determine the second correction parameter based on the second reference response; Step 3: During the object measurement period, acquire the first object impedance response and the second object impedance response. The first object impedance response corresponds to the first excitation frequency, and the second object impedance response corresponds to the second excitation frequency. Step 4: Simultaneously demodulate the impedance responses of the first and second objects to obtain the first impedance change signal and the second impedance change signal; Step 5: Correct the first impedance change signal according to the first correction parameter, and correct the second impedance change signal according to the second correction parameter; Step 6: Generate output parameters based on the corrected first impedance change signal and the corrected second impedance change signal.

[0045] The multi-frequency impedance and pulse wave fusion real-time blood pressure monitoring system of the present invention includes: a processor and a memory, wherein the memory stores processor-executable instructions, and when executed by the processor, the processor performs the following operations: During the reference calibration period, a first reference response and a second reference response are acquired. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of a predetermined reference impedance at a second excitation frequency. The first excitation frequency and the second excitation frequency are different. In one example, the present invention employs a wrist-worn blood pressure monitoring device, which includes a device housing, a processor, a memory, a multi-frequency impedance measurement circuit, a pulse wave sensor, a power management circuit, a wireless communication circuit, and a contact state detection circuit.

[0046] The wrist-worn blood pressure monitoring device features an impedance measurement electrode assembly and a pulse wave sensor on the side facing the wrist skin. The impedance measurement electrode assembly includes a first excitation electrode, a second excitation electrode, a first detection electrode, and a second detection electrode. The first and second excitation electrodes inject an alternating current into the wrist tissue, while the first and second detection electrodes acquire the alternating voltage across the wrist tissue. The four electrodes can be made of 316L gold-plated stainless steel, with each electrode having a skin contact area of ​​approximately 4 mm × 6 mm. The center-to-center distance between the excitation and detection electrodes is 3 mm to 6 mm. The multi-frequency impedance measurement circuit includes an AD5941 bioimpedance analog front-end, an ADG704 analog switch, an OPA2192 low-noise operational amplifier, and a 100 Ω precision current sensing resistor. The AD5941 generates the alternating current excitation, acquires the electrode voltage, and performs analog-to-digital conversion; the ADG704 switches between the human impedance path and the internal reference impedance path; and the OPA2192 buffers and adjusts the gain of the detection voltage. The processor can be an nRF52840 chip with a clock frequency of 64 MHz, featuring an internal 12-bit analog-to-digital converter and a high-speed SPI interface. External memory can be an 8 MB W25Q64 serial flash memory for storing reference responses, calibration parameters, impedance change signals, abnormal event data, and device operation logs. The device can also be equipped with a 32.768 kHz crystal oscillator to provide a reference for channel time synchronization and timestamps. The pulse wave sensor can be a MAX86141 photoplethysmography (PPG) sensor, configured with a 525 nm wavelength green LED and photodetector, with a sampling rate of 400 Hz. The pulse wave sensor shares the same time reference with the impedance measurement electrode group, enabling feature extraction of impedance change signals and pulse wave signals on the same time axis. The contact state detection circuit can use a thin-film pressure sensor located below the electrode mounting bracket, or utilize the impedance measurement electrode group to obtain the electrode-skin interface impedance. The power management circuit uses a rechargeable lithium battery, a BQ24075 charging management chip, and a low-noise, low-dropout regulator, with an analog front-end supply voltage of 3.3 V.

[0047] In a specific example, the wrist-worn blood pressure monitoring device measures 42 mm × 38 mm × 11 mm and weighs approximately 38 g; the wristband is 20 mm wide; the device fits snugly against the skin near the radial artery in the wrist, with the pulse wave sensor positioned opposite the radial artery projection area, and impedance measurement electrodes located on either side of the pulse wave sensor. The processor transmits the measurement results to a hospital workstation via Bluetooth Low Energy. The hospital workstation uses a standard commercial computer configured with an Intel Core i5 processor, 16 GB of memory, and a Windows or Linux operating system.

[0048] In one example, the reference calibration period refers to the time during which the reference response is continuously acquired after the device is powered on, the wearer's condition stabilizes, and the device connects its internal reference impedance to the multi-frequency impedance measurement circuit. The duration of the reference calibration period is set to 8 seconds, with the first 2 seconds used to establish a stable simulated front-end state and the last 6 seconds used to obtain valid reference data.

[0049] The first reference calibration can be performed after the device is started, and subsequent reference calibrations are performed every 10 minutes. When the contact status parameters or the object impedance baseline parameters change significantly, the device immediately initiates a new 8-second reference calibration period.

[0050] The predetermined reference impedance is a 220 Ω metal film precision resistor with an accuracy of 0.1% and a temperature coefficient not exceeding 25 ppm / ℃. The reasons for choosing 220 Ω are as follows: First, the equivalent measurement impedance of wrist skin and superficial tissue in the 50 kHz to 100 kHz range is typically between approximately 150 Ω and 350 Ω, with 220 Ω falling in the middle of this range; second, the detection voltage corresponding to 220 Ω is approximately 4.4 mV, which allows the analog front-end to operate within a good linear input range; and third, the 220 Ω precision resistor facilitates the establishment of a stable and repeatable electrical reference within the device.

[0051] In a specific example, the first excitation frequency is 50 kHz and the second excitation frequency is 100 kHz, both of which use a sinusoidal excitation current with an effective value of 20 μA.

[0052] At 50 kHz, the theoretical response of the internal reference impedance is 220 Ω with a phase of 0°. The first reference response acquired by the device is converted to 216.8 Ω with a phase of -2.4°, corresponding to a detection voltage of approximately 4.336 mV.

[0053] At 100 kHz, the theoretical response of the internal reference impedance is 220 Ω with a phase of 0°. The second reference response acquired by the device is converted to 218.9 Ω with a phase of -1.1°, corresponding to a detection voltage of approximately 4.378 mV.

[0054] A first correction parameter is determined based on a first reference response, and a second correction parameter is determined based on a second reference response; During the object measurement period, a first object impedance response and a second object impedance response are acquired. The first object impedance response corresponds to a first excitation frequency, and the second object impedance response corresponds to a second excitation frequency. In one example, the object measurement period refers to the time during which the impedance measurement electrode set remains in contact with the skin of the human wrist, continuously acquiring the human impedance response and pulse wave signal. A single object measurement period is set to 30 seconds. 30 seconds covers at least 30 cardiac cycles, suitable for extracting pulse amplitude, pulse interval, pulse rise time, and impedance variation characteristics.

[0055] The process of acquiring the object impedance response includes: the processor controls the AD5941 to inject alternating current between the first excitation electrode and the second excitation electrode at 50 kHz and 100 kHz respectively; the first detection electrode and the second detection electrode collect the alternating voltage across the human tissue; the analog front end converts the alternating voltage into digital sampled values; and the processor demodulates the digital sampled values ​​according to the corresponding excitation frequency.

[0056] In a specific example, after the device is worn on a human body, the average impedance measured on the 50 kHz channel is approximately 223.4 Ω, the peak-to-peak impedance change caused by the cardiac cycle is approximately 1.66 Ω, and the standard deviation of the channel noise is approximately 0.06 Ω. Taking the peak systolic phase as an example, the instantaneous impedance corresponding to the first object's impedance response is 224.23 Ω, and the impedance change relative to the local baseline is approximately 0.83 Ω.

[0057] In one specific example, the average impedance measured on the 100 kHz channel is approximately 219.7 Ω, the peak-to-peak impedance change due to the cardiac cycle is approximately 1.14 Ω, and the standard deviation of the channel noise is approximately 0.08 Ω. Taking the same systolic peak as an example, the instantaneous impedance corresponding to the impedance response of the second object is 220.27 Ω, and the impedance change relative to the local baseline is approximately 0.57 Ω.

[0058] The impedance responses of the first and second objects are synchronously demodulated to obtain the first impedance change signal and the second impedance change signal. In one specific embodiment, the first object impedance response and the second object impedance response are synchronously demodulated according to their respective excitation frequencies.

[0059] The processor uses a digital phase accumulator sourced from the same excitation signal to generate a first reference sine sequence and a first reference cosine sequence. The impedance response of the first object is multiplied by the first reference sine sequence and the first reference cosine sequence, respectively, to obtain a first in-phase component and a first quadrature component. The first in-phase component and the first quadrature component are filtered by a fourth-order low-pass filter with a cutoff frequency of 20 Hz and output at a sampling rate of 200 Hz.

[0060] For the impedance response of the second object, the processor performs the same processing using the second reference sine sequence and the second reference cosine sequence corresponding to 100 kHz to obtain the second in-phase component and the second quadrature component. The processor calculates the impedance magnitude, phase, and impedance change relative to the local baseline based on the in-phase component and the quadrature component.

[0061] In a specific example, the AC voltage change of the first object impedance response near the peak of systole is approximately 16.6 μV. Dividing this by the 20 μA excitation current yields the original first impedance change signal of approximately 0.83 Ω. After removing the 2-second sliding median baseline, the first impedance change signal exhibits a pulse-related waveform that first rises rapidly and then falls slowly over one cardiac cycle.

[0062] In a specific example, the AC voltage change of the second object impedance response near the peak of the same systolic phase is approximately 11.4 μV. Dividing this by the 20 μA excitation current yields the original second impedance change signal of approximately 0.57 Ω. The second impedance change signal has the same cardiac cycle as the first impedance change signal, but its amplitude and phase differ due to frequency characteristics.

[0063] The first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter. Output parameters are generated based on the corrected first impedance change signal and the corrected second impedance change signal.

[0064] In one example, the output parameters include at least the weighted impedance change, and may also include systolic blood pressure estimates, diastolic blood pressure estimates, pulse rate, and signal quality indicators. The pulse wave sensor provides a photoplethysmography (PPG) pulse wave, from which the processor extracts the pulse arrival time, pulse interval, and rise time, and fuses the weighted impedance change with the pulse wave characteristics.

[0065] At the start of object measurement, the device has not yet obtained the reference response change between adjacent reference calibration periods, so a pre-set initial first weight and initial second weight are used. Since the 50 kHz channel typically has a high pulse change amplitude in wrist measurements, the initial first weight is set to 0.60, the initial second weight is set to 0.40, and the sum of the two initial weights is 1.

[0066] In a specific example, after amplitude correction, phase correction, and time compensation, the peak value of the contraction period of the first impedance change signal is 0.837 Ω, and the peak value of the contraction period of the second impedance change signal is 0.553 Ω. Using an initial first weight of 0.60 and an initial second weight of 0.40, a weighted combination is obtained, yielding a weighted impedance change of approximately 0.723 Ω.

[0067] During the same measurement period, the pulse wave sensor measured the pulse arrival time as 182 ms. The processor inputs the weighted impedance change of 0.723Ω and the pulse arrival time of 182 ms into a pre-stored blood pressure estimation model, obtaining a systolic blood pressure of approximately 129 mmHg, a diastolic blood pressure of approximately 78 mmHg, and a pulse rate of approximately 76 beats / min. The output parameters can be displayed on a screen or transmitted to a hospital workstation via Bluetooth.

[0068] In some preferred embodiments, the first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter, and the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter; The correction of the first impedance change signal includes: The amplitude correction, phase correction, and time compensation are performed on the first impedance change signal; Correcting the second impedance change signal includes: The amplitude, phase, and time of the second impedance change signal are corrected.

[0069] In one specific embodiment, the processor determines the first and second correction parameters according to the following steps: First, the reference voltage acquired during the reference calibration period is synchronously demodulated to obtain the amplitude and phase of the reference response. Second, the amplitude of the reference response is compared with the theoretical amplitude of 220 Ω to obtain the amplitude correction parameter. The amplitude correction parameter is equal to the ratio of the theoretical amplitude to the actual measured amplitude. Third, the phase of the reference response is compared with the theoretical phase of 0° to obtain the phase correction parameter. The phase correction parameter is used to compensate for the phase shift introduced by the analog front-end, analog switches, electrode connection lines, and analog-to-digital conversion channels. Finally, the channel delay parameter is calculated based on the phase shift of the reference response and the corresponding excitation frequency. The channel delay parameter is used for fractional delay compensation of the digital signal.

[0070] In a specific example, the amplitude of the first reference response is 216.8 Ω, and the theoretical amplitude is 220 Ω; therefore, the first amplitude correction parameter is 1.0148. The phase of the first reference response is -2.4°; therefore, the first phase correction parameter is +2.4°. At 50 kHz, a 2.4° phase shift corresponds to a channel delay of approximately 0.13 μs; therefore, the first channel delay parameter is set to 0.13 μs, and the first channel signal is advanced by 0.13 μs during time compensation.

[0071] In a specific example, the amplitude of the second reference response is 218.9 Ω, and the theoretical amplitude is 220 Ω; therefore, the second amplitude correction parameter is 1.0050. The phase of the second reference response is -1.1°; therefore, the second phase correction parameter is +1.1°. At 100 kHz, a 1.1° phase shift corresponds to a channel delay of approximately 0.03 μs; therefore, the second channel delay parameter is set to 0.03 μs, and the second channel signal is advanced by 0.03 μs during time compensation.

[0072] The first and second correction parameters can be stored in external flash memory as floating-point numbers and associated with a reference calibration timestamp.

[0073] In one example, the amplitude correction of the first impedance change signal is performed by multiplication using a first amplitude correction parameter of 1.0148; the phase correction is performed by complex rotation of the first in-phase component and the first quadrature component using a first phase correction parameter of +2.4°; and the time compensation is performed by advancing the first channel waveform by 0.13 μs using a first channel delay parameter of 0.13 μs.

[0074] In a specific example, the first impedance change signal is 0.825 Ω after amplitude normalization. After the first amplitude correction, the amplitude becomes 0.837 Ω; after a +2.4° phase rotation, the phase of the first channel is aligned with the reference phase; after using a 0.13 μs fractional delay filter, the waveforms of the first channel and the second channel are aligned on the same rising edge of the pulse.

[0075] The amplitude correction of the second impedance change signal is performed by multiplication using the second amplitude correction parameter of 1.0050; the phase correction is performed using the second phase correction parameter of +1.1°; and the time compensation is performed using the second channel delay parameter of 0.03 μs.

[0076] In a specific example, the second impedance change signal is 0.550 Ω after amplitude normalization. After the second amplitude correction, the amplitude becomes 0.553 Ω; after a +1.1° phase rotation, the phase of the second channel is corrected; after using a 0.03 μs fractional delay filter, the rising edge of the second channel waveform is synchronized with the pulse of the first channel waveform.

[0077] In some preferred embodiments, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: The quality of the first frequency channel and the quality of the second frequency channel are determined based on the change in the first reference response and the change in the second reference response between adjacent reference calibration periods, respectively. The first weight and the second weight are determined based on the quality of the first frequency channel and the quality of the second frequency channel, respectively. Based on the corrected first impedance change signal and the corrected second impedance change signal, output parameters are generated, including: Based on the first weight and the second weight, the corrected first impedance change signal and the corrected second impedance change signal are weighted and combined to generate output parameters.

[0078] In one example, frequency channel quality is used to represent the stability of a given excitation frequency channel between adjacent reference calibration periods. Frequency channel quality can be determined based on the amplitude and phase variations of the reference response. The smaller the amplitude and phase variations, the higher the frequency channel quality.

[0079] In one specific embodiment, the processor normalizes the amplitude change relative to 220 Ω and the phase change relative to 5° to obtain the channel drift, and then converts the channel drift into a quality value between 0 and 1. The quality value can be calculated by dividing 1 by the sum of 1 and the channel drift.

[0080] After the first reference calibration, the first reference response of the 50 kHz channel was 216.8 Ω and -2.4°; the adjacent reference calibration result 10 minutes later was 217.6 Ω and -2.1°. The amplitude change between the two reference calibration periods was 0.8 Ω and the phase change was 0.3°, resulting in a calculated first frequency channel quality of approximately 0.94.

[0081] After the first reference calibration, the second reference response of the 100 kHz channel was 218.9 Ω and -1.1°; the adjacent reference calibration result 10 minutes later was 221.5 Ω and -2.0°. The amplitude change between the two reference calibration periods was 2.6 Ω and the phase change was 0.9°, resulting in a calculated second frequency channel quality of approximately 0.84.

[0082] The processor adjusts the initial first weight and initial second weight based on the quality of the first frequency channel and the quality of the second frequency channel. Specifically, the initial first weight is multiplied by the quality of the first frequency channel to obtain the first channel adjustment value; the initial second weight is multiplied by the quality of the second frequency channel to obtain the second channel adjustment value; and the first channel adjustment value and the second channel adjustment value are then normalized to obtain the updated first weight and second weight.

[0083] In a specific example, the initial first weight is 0.60, and the initial second weight is 0.40. The first channel correction value is 0.60 × 0.94, which is 0.564; the second channel correction value is 0.40 × 0.84, which is 0.336. The first weight is 0.564 divided by 0.900, which is approximately 0.627; the second weight is 0.336 divided by 0.900, which is approximately 0.373. The sum of the two updated weights is 1.

[0084] Using updated first weights of 0.627 and second weights of 0.373, the 0.837 Ω first impedance change signal and the 0.553 Ω second impedance change signal are weighted and combined to obtain a weighted impedance change of approximately 0.731 Ω. Because the drift of the second frequency channel between adjacent reference calibration periods is greater than that of the first frequency channel, the second weight is reduced from the initial value of 0.40 to 0.373, and the first weight is increased from the initial value of 0.60 to 0.627.

[0085] The purpose of dynamically re-determining the weights is to adjust the channel contribution based on the actual changes in the reference response, while preserving the initial reliability of each frequency channel. When the reference response of a frequency channel drifts significantly, its contribution to the output parameters is reduced; when a frequency channel remains stable, its contribution to the output parameters is increased. This reduces the impact of electrode contact, skin moisture content, ambient temperature, and changes in the simulation front-end gain on the weighted combination results, thereby improving the stability of continuous blood pressure monitoring.

[0086] In some preferred embodiments, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: Obtain contact state parameters between the object and the electrode or baseline parameters of the object impedance; When the change in the contact state parameter or the object impedance baseline parameter exceeds the first preset threshold, the first reference response and the second reference response are reacquired.

[0087] In one example, the contact state parameter is used to represent the degree of contact between the impedance measurement electrode assembly and human skin. The contact state parameter can be determined by the electrode-skin interface impedance, the output value of the thin-film pressure sensor, and the high-frequency contact impedance between the detection electrodes.

[0088] In this embodiment, the electrode-skin interface impedance is normalized as a contact state parameter. When the device is first worn and the wristband is properly tightened, the interface impedance is approximately 8.5 kΩ, which is defined as the contact state baseline value of 1.00. When the interface impedance increases to 10.4 kΩ, the contact state parameter is 1.224, indicating a decrease in the effective contact between the electrode and the skin.

[0089] The object impedance baseline parameter represents the stable average value of human tissue impedance over a short time window. The processor takes the median of the real impedance over 2 consecutive seconds on a 50 kHz channel to obtain the object impedance baseline parameter. Under normal wearing conditions, the object impedance baseline parameter is approximately 223 Ω.

[0090] The first preset threshold can be set to a relative change of 20%. When the change in the contact state parameter relative to the initial value is greater than 20%, or the change in the object impedance baseline parameter relative to the initial value is greater than 20%, the device re-performs reference calibration.

[0091] In one specific example, the device's initial contact state parameter was 1.00. During the measurement, due to the wristband loosening, it rose to 1.224, a change of 22.4%, exceeding the first preset threshold of 20%. The processor stopped using the current calibration parameters, controlled the analog switch to connect to the internal 220 Ω reference impedance, and re-executed the 8-second reference calibration.

[0092] In another specific example, the initial baseline impedance parameter of the device was 223 Ω. After human movement, the baseline impedance parameter increased to 278 Ω, a change of approximately 24.7%, exceeding the first preset threshold of 20%. The processor also reacquired the first and second reference responses. After recalibration, the 50 kHz channel obtained a first reference response of 217.1 Ω and -2.2°, and the 100 kHz channel obtained a second reference response of 219.0 Ω and -1.0°. The first and second correction parameters were updated based on the new reference responses.

[0093] The purpose of setting contact state parameters and object impedance baseline parameters is to identify changes in wearing conditions. Loosening of the garment, wet skin, changes in wrist posture, or partial detachment of the electrodes may render the original calibration parameters inapplicable. Recalibrating after exceeding a first preset threshold can reduce amplitude and phase errors caused by contact changes.

[0094] In some preferred embodiments, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: The first actual excitation current at the first excitation frequency and the second actual excitation current at the second excitation frequency are obtained respectively. The amplitude of the first impedance change signal and the second impedance change signal is normalized based on the first actual excitation current, the first reference response, the second actual excitation current and the second reference response.

[0095] In one example, the actual excitation current is acquired after synchronous demodulation is completed and before amplitude correction is performed using the first and second correction parameters. The processor calculates the actual excitation current for the 50 kHz and 100 kHz channels by detecting the voltage across a 100 Ω precision current sensing resistor.

[0096] In a specific example, the processor sets 20 μA as the nominal excitation current and 220 Ω as the nominal reference impedance. The first actual excitation current measured at the first excitation frequency is 19.6 μA RMS, and the second actual excitation current measured at the second excitation frequency is 19.2 μA RMS. During the reference calibration period, the amplitude of the first reference response is 216.8 Ω, and the amplitude of the second reference response is 218.9 Ω.

[0097] The processor calculates the amplitude normalization ratio for each of the two frequency channels based on the ratio of the actual excitation current to the nominal excitation current, and the ratio of the nominal reference impedance to the corresponding reference response amplitude. The amplitude normalization ratio for the first channel is calculated as follows: First channel amplitude normalization ratio =(19.6 μA÷20.0 μA)×(220Ω÷216.8Ω) =0.98 × 1.0148 =0.9945, rounded to three decimal places, we get 0.994.

[0098] The amplitude normalization ratio of the second channel is calculated as follows: Second channel amplitude normalization ratio =(19.2 μA÷20.0 μA)×(220Ω÷218.9Ω) =0.96 × 1.0050 =0.9648, rounded to three decimal places, we get 0.965.

[0099] Therefore, the amplitude normalization ratio for the first channel is 0.994, and the amplitude normalization ratio for the second channel is 0.965. The first term in the above calculation is used to compensate for the deviation of the actual excitation current from the nominal excitation current, and the second term is used to compensate for the deviation of the corresponding reference response amplitude from the 220 Ω nominal reference impedance.

[0100] In a specific example, the original first impedance change signal is 0.83 Ω, which is multiplied by the first channel normalization ratio of 0.994 to obtain 0.825 Ω; the original second impedance change signal is 0.57 Ω, which is multiplied by the second channel normalization ratio of 0.965 to obtain 0.550 Ω. Subsequently, multiplying by the first amplitude correction parameter of 1.0148 and the second amplitude correction parameter of 1.0050 respectively, we obtain amplitude correction results of 0.837 Ω and 0.553 Ω.

[0101] The purpose of using the actual excitation current for amplitude normalization is to eliminate amplitude deviations caused by current source output errors, battery voltage changes, and analog switch on-resistance changes, thereby making the impedance change signals obtained at different times and under different wearing conditions more comparable.

[0102] In some preferred embodiments, when the processor-executable instructions are executed by the processor, the processor also causes the processor to perform the following operations: Store the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter; When the consistency between the first impedance change signal and the second impedance change signal is detected to be lower than a second preset threshold, an abnormal event is determined to have occurred. The first and second object impedance responses are acquired within preset time periods before and after the occurrence of the abnormal event.

[0103] In one example, the consistency between the first and second impedance change signals is used to indicate whether the responses of the two frequency channels to the same cardiac event are consistent. Consistency can be determined jointly by the time-aligned correlation coefficient, the difference in pulse peak position, and the amplitude ratio.

[0104] This embodiment first calculates the Pearson correlation coefficient between the two frequency channels within a 2-second sliding window, and then calculates the time difference and amplitude ratio of the pulse peak positions of the two channels. If the correlation coefficient, peak position difference, and amplitude ratio are all within the allowable range, the consistency is considered to be high.

[0105] The second preset threshold can be determined by collecting multiple sets of calibration data under normal wear, slight hand movements, and different skin contact conditions. Based on 40 sets of normal operation data, a threshold that makes the false alarm rate of normal data less than 5% is selected as the second preset threshold. In this embodiment, the second preset threshold is set to 0.65.

[0106] In a specific example, the correlation coefficient between the first and second impedance change signals within a 2-second window is 0.42, the pulse peak delay of the second channel relative to the first channel is 86 ms, and the amplitude ratio of the two channels is 0.58. Using the smaller of the correlation coefficient and the amplitude ratio as the consistency evaluation value, a consistency score of 0.42 is obtained, which is lower than the second preset threshold of 0.65. The processor determines that an abnormal event has occurred.

[0107] Abnormal events can be caused by rapid human movement, brief electrode detachment, obstruction of the pulse wave sensor, or saturation of a certain frequency channel. The processor continuously stores the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter. After detecting an abnormal event, it saves the first object impedance response and the second object impedance response collected within 5 seconds before and after the occurrence of the abnormal event.

[0108] In a specific example, the abnormal event occurred at 14:32:18.450. The device saves the raw object impedance response from 14:32:13.450 to 14:32:23.450, with both the 50 kHz and 100 kHz channels saved at a 1 kHz sampling rate, for a total of 10,000 samples per channel. It also saves the timestamps before and after the event, the actual excitation current, contact state parameters, and the object impedance baseline parameters.

[0109] The purpose of setting up consistency judgment and abnormal event caching is to avoid using the waveform of a single abnormal channel for blood pressure estimation, while retaining the original data before and after the abnormality occurred, so as to facilitate hospital workstations to review, wear status analysis and equipment maintenance.

[0110] In one example, a combined software and hardware simulation was conducted in a simulated deployment environment of the outpatient and inpatient wards of a top-tier general hospital in Beijing. The simulation hardware included an nRF52840 processor, an AD5941 impedance simulation front-end, a MAX86141 pulse wave sensor, a W25Q64 external memory, a four-electrode impedance measurement electrode group 115, and a wristband pressure sensor. The host computer used an Intel Core i7-12700 processor, 32 GB of memory, and an Ubuntu 22.04 operating system. The algorithm development environment employed Python 3.11, NumPy 2.0, and C language embedded firmware.

[0111] The simulation data includes 48 sets of human wrist impedance and pulse wave data, each lasting 30 seconds. The simulated systolic blood pressure ranges from 90 mmHg to 170 mmHg, diastolic blood pressure from 55 mmHg to 105 mmHg, and pulse rate from 55 beats / min to 110 beats / min. Simulation conditions include excitation current variation within 6% of the nominal value, contact interface impedance variation of 10% to 25%, single-frequency channel amplitude drift of 1% to 5%, and 20% of the measurement window including hand movement interference.

[0112] The comparison method uses only a 50 kHz channel, fixed amplitude correction parameters, and fixed weights in a single-frequency approach. This implementation uses dual-frequency channels (50 kHz and 100 kHz), normalized actual excitation current, dynamic calibration of the reference response, dynamic weights, and anomaly detection.

[0113] In the simulation, "decision efficiency" is defined as the proportion of the number of measurement windows that complete the judgment of valid signals, output blood pressure estimates and correctly mark abnormal windows within 3 seconds to the total number of measurement windows.

[0114] Simulation results show that the single-frequency method achieves an effective decision rate of 86.5% within 3 seconds, while this implementation achieves 93.4%, representing an improvement in decision efficiency of approximately 8.0%. The mean absolute error of systolic blood pressure is 8.7 mmHg for the single-frequency method, compared to 7.4 mmHg for this implementation, a reduction of approximately 14.9%. The F1 score for abnormal window identification is 0.82 for the single-frequency method, compared to 0.89 for this implementation, an improvement of approximately 8.5%. The median processing delay is 1.74 seconds for the single-frequency method, compared to 1.52 seconds for this implementation, a reduction of approximately 12.6%.

[0115] The simulation results above demonstrate that calibration between multi-frequency impedance channels, normalization of actual excitation current, dynamic weight adjustment, and consistency anomaly detection can improve the effective decision ratio and anomaly identification capability of wrist-worn blood pressure monitoring devices without significantly increasing hardware resources, and reduce the impact of impedance channel drift on blood pressure estimation results.

[0116] The above embodiments are merely illustrative of preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, various modifications, equivalent substitutions, combinations, splits, or variations can be made to the above embodiments without departing from the technical concept and essence of the present invention. All such modifications, substitutions, combinations, splits, or variations falling within the scope of the technical concept of the present invention should be included within the scope of protection of the present invention.

[0117] The descriptions of specific embodiments, parameter values, step sequences, module divisions, and application scenarios in this specification are for the purpose of facilitating understanding and illustrating the technical solutions of the present invention, and should not be construed as limiting the specific implementation of the present invention. Any equivalent changes, equivalent substitutions, or conventional improvements made based on the claims and specification of the present invention should fall within the protection scope of the present invention. The protection scope of the present invention is defined by the appended claims.

Claims

1. A real-time blood pressure monitoring system that integrates multi-frequency impedance and pulse wave fusion, characterized in that, include: A processor and a memory, wherein the memory stores processor-executable instructions that, when executed by the processor, cause the processor to perform the following operations: During the reference calibration period, a first reference response and a second reference response are acquired. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of the predetermined reference impedance at a second excitation frequency. The first excitation frequency and the second excitation frequency are different. A first correction parameter is determined based on the first reference response, and a second correction parameter is determined based on the second reference response; During the object measurement period, a first object impedance response and a second object impedance response are acquired, wherein the first object impedance response corresponds to the first excitation frequency and the second object impedance response corresponds to the second excitation frequency. The impedance responses of the first object and the second object are synchronously demodulated to obtain the first impedance change signal and the second impedance change signal. The first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter; Output parameters are generated based on the corrected first impedance change signal and the corrected second impedance change signal.

2. The system according to claim 1, characterized in that, The first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter; the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter. The correction of the first impedance change signal includes: The first impedance change signal is subjected to amplitude correction, phase correction and time compensation; Correcting the second impedance change signal includes: The second impedance change signal is subjected to amplitude correction, phase correction and time compensation.

3. The system according to claim 1, characterized in that, When executed by the processor, the processor-executable instructions also cause the processor to perform the following operations: The quality of the first frequency channel and the quality of the second frequency channel are determined based on the changes in the first reference response and the second reference response between adjacent reference calibration periods, respectively. The first weight and the second weight are determined based on the quality of the first frequency channel and the quality of the second frequency channel, respectively. Based on the corrected first impedance change signal and the corrected second impedance change signal, output parameters are generated, including: Based on the first weight and the second weight, the corrected first impedance change signal and the corrected second impedance change signal are weighted and combined to generate output parameters.

4. The system according to claim 2, characterized in that, When executed by the processor, the processor-executable instructions also cause the processor to perform the following operations: Obtain contact state parameters between the object and the electrode or baseline parameters of the object impedance; When the change in the contact state parameter or the object impedance baseline parameter exceeds a first preset threshold, the first reference response and the second reference response are reacquired.

5. The system according to claim 2, characterized in that, When executed by the processor, the processor-executable instructions also cause the processor to perform the following operations: The first actual excitation current at the first excitation frequency and the second actual excitation current at the second excitation frequency are obtained respectively. The amplitude of the first impedance change signal and the second impedance change signal is normalized based on the first actual excitation current, the first reference response, the second actual excitation current, and the second reference response.

6. The system according to claim 2, characterized in that, When executed by the processor, the processor-executable instructions also cause the processor to perform the following operations: Store the first impedance change signal, the second impedance change signal, the first correction parameter, and the second correction parameter; When the consistency between the first impedance change signal and the second impedance change signal is detected to be lower than a second preset threshold, an abnormal event is determined to have occurred. The impedance responses of the first object and the second object are stored within preset time periods before and after the occurrence of the abnormal event.

7. A real-time blood pressure monitoring method using multi-frequency impedance and pulse wave fusion, characterized in that, include: During the reference calibration period, a first reference response and a second reference response are acquired. The first reference response corresponds to the response of a predetermined reference impedance at a first excitation frequency, and the second reference response corresponds to the response of the predetermined reference impedance at a second excitation frequency. The first excitation frequency and the second excitation frequency are different. A first correction parameter is determined based on the first reference response, and a second correction parameter is determined based on the second reference response; During the object measurement period, a first object impedance response and a second object impedance response are acquired, wherein the first object impedance response corresponds to the first excitation frequency and the second object impedance response corresponds to the second excitation frequency. The impedance responses of the first object and the second object are synchronously demodulated to obtain the first impedance change signal and the second impedance change signal. The first impedance change signal is corrected according to the first correction parameter, and the second impedance change signal is corrected according to the second correction parameter; Output parameters are generated based on the corrected first impedance change signal and the corrected second impedance change signal.

8. The method according to claim 7, characterized in that, The first correction parameter includes a first amplitude correction parameter, a first phase correction parameter, and a first channel delay parameter; the second correction parameter includes a second amplitude correction parameter, a second phase correction parameter, and a second channel delay parameter. The correction of the first impedance change signal includes: The first impedance change signal is subjected to amplitude correction, phase correction and time compensation; Correcting the second impedance change signal includes: The second impedance change signal is subjected to amplitude correction, phase correction and time compensation.

9. The method according to claim 7, characterized in that, Also includes: The quality of the first frequency channel and the quality of the second frequency channel are determined based on the changes in the first reference response and the second reference response between adjacent reference calibration periods, respectively. The first weight and the second weight are determined based on the quality of the first frequency channel and the quality of the second frequency channel, respectively. Based on the corrected first impedance change signal and the corrected second impedance change signal, output parameters are generated, including: Based on the first weight and the second weight, the corrected first impedance change signal and the corrected second impedance change signal are weighted and combined to generate output parameters.

10. The method according to claim 8, characterized in that, Also includes: Obtain contact state parameters between the object and the electrode or baseline parameters of the object impedance; When the change in the contact state parameter or the object impedance baseline parameter exceeds a first preset threshold, the first reference response and the second reference response are reacquired.

Images