A multi-functional distributed multi-vital sign non-contact detection radar system

Abstract

The application discloses a kind of multi-functional distributed multi-vital sign non-contact detection radar systems, by central processing module, line distribution module and front-end detection module constitute.Front-end detection module is made of multiple vital sign sensing access module, emits optical signal or microwave signal to detection area and receives echo, obtains original detection signal.Central processing module includes signal generation and processing unit, pulse extraction unit, blood pressure estimation unit, dynamic tracking and fall prevention early warning unit, realize the generation and processing of detection signal, pulse extraction, blood pressure estimation and based on range image and micro-doppler analysis fall prevention monitoring and early warning processing and other functions.Line distribution module connects sensing access module and central processing module and adds delay in each link, is made of multiple optical splitters and optical delay. The application realizes non-contact, breathing and pulse detection, obtains human pulse information, quickly estimates human systolic pressure, prevents fall and dynamic tracking continuous measurement function.

Description

Technical Field

[0001] This invention relates to the fields of radar and vital sign detection technology, and in particular to a multifunctional distributed multi-vital sign non-contact radar system. Background Technology

[0002] Vital signs monitoring is a core technical indicator for assessing the physiological state of the human body, mainly including parameters such as respiratory rate, pulse rate, body temperature, and blood pressure. In the fields of medical monitoring and public safety, achieving continuous and real-time monitoring of these vital signs parameters has significant clinical application value and social significance. Traditional contact monitoring technologies have obvious limitations: on the one hand, their sensors need to maintain close contact with the skin, making them unsuitable for special populations with limited skin conditions, such as burn patients; on the other hand, long-term contact measurement can easily cause user discomfort, making it difficult to meet the clinical application needs of continuous monitoring. These technical shortcomings have prompted non-contact vital signs monitoring technology to become an important research direction in the field of medical monitoring.

[0003] Currently, non-contact vital sign detection technology is mainly based on visual imaging monitoring systems. These systems capture subtle changes in the human skin (such as skin color fluctuations or chest movement) through cameras and extract signals such as heart rate and respiratory rate using image processing algorithms. Their advantages are that the equipment is widely available and can monitor multiple targets, but they are dependent on lighting conditions and are easily affected by motion.

[0004] Furthermore, traditional pulse diagnosis relies on the physician's tactile sense, which is highly subjective and difficult to quantify, record, and pass on. Most existing electronic pulse oximeters require contact pressure sensors, which can easily cause user discomfort, or lead to significant errors in test results due to the inability of some users to cooperate correctly. In addition, traditional blood pressure measurement relies on contact methods, such as cuff-based oscillometric electronic blood pressure monitors. The inflatable cuff compresses the arm, which can cause pain or discomfort, especially for patients requiring frequent measurements. Moreover, traditional cuff-based measurements require tens of seconds per measurement, with a 1-2 minute interval between measurements, making it impossible to capture instantaneous changes in blood pressure (such as sudden hypertension or hypotension), limiting its application in acute and critical care scenarios. Summary of the Invention

[0005] This invention provides a multifunctional distributed multi-vital sign non-contact detection radar system, which realizes non-contact detection of respiration and pulse, simultaneous acquisition of human pulse information, rapid estimation of human systolic blood pressure, fall prevention, and dynamic tracking and continuous measurement functions.

[0006] This invention provides a multifunctional distributed multi-vital sign non-contact detection radar system, comprising: a central processing module, a line distribution module, and a front-end detection module; the central processing module includes: an optical spectrum manipulation module, a second-stage 1×2 beam splitter 1, an optical circulator, a 2×1 optical coupler, photodetectors N+1, N×1 optical couplers, photodetectors N+2, a digital-to-analog converter, a data processing module, a pulse extraction unit, a blood pressure estimation unit, and a dynamic tracking and fall prevention early warning unit; the line distribution module includes: a 1×N beam splitter 1, optical delay units 1~N, a 1×N beam splitter 2, a third-stage 1×2 beam splitter 1~N, and optical delay units N+1~2N; the front-end detection module includes: an optical sensing access module 1. ~N and RF sensing access modules 1~N; the first optical signal output terminal of the optical domain spectrum manipulation module is connected to the optical signal input terminal of the second-stage 1×2 beam splitter 1, the first optical signal output terminal of the second-stage 1×2 beam splitter 1 is connected to the optical signal input terminal of the optical circulator, and the optical circulator is also connected to the bidirectional optical path of the 1×N beam splitter 1; the 1×N beam splitter 1 is also connected to the bidirectional optical path of each optical delay unit in the optical delay units 1~N; each optical delay unit in the optical delay units 1~N is also connected to the bidirectional optical path of each optical sensing access module in the optical sensing access modules 1~N; the second optical signal output terminal of the second-stage 1×2 beam splitter 1 is connected to the first optical signal input terminal of the 2×1 optical coupler. The optical signal output terminal of the optical circulator is connected to the second optical signal input terminal of the 2×1 optical coupler; the optical signal output terminal of the 2×1 optical coupler is connected to the optical signal input terminal of the photodetector N+1, and the electrical signal output terminal of the photodetector N+1 is connected to the first electrical signal input terminal of the digital-to-analog converter; the second optical signal output terminal of the optical domain spectrum manipulation module is connected to the optical signal input terminal of the 1×N beam splitter 2, and the optical signal output terminal of the 1×N beam splitter 2 is connected to the optical signal input terminals of each beam splitter in the third-stage 1×2 beam splitter 1~N; the first signal output terminal of each beam splitter in the third-stage 1×2 beam splitter 1~N is connected to the optical delay unit N+1~2 The optical signal input terminals of each optical delay unit in N are connected one-to-one; the optical signal output terminals of each optical delay unit in N+1~2N are connected one-to-one with the optical signal input terminals of each radio frequency sensor access module in RF sensor access modules 1~N; the second signal output terminals of each beam splitter in the third-stage 1×2 beam splitter 1~N are connected one-to-one with the optical signal input terminals of each radio frequency sensor access module in RF sensor access modules 1~N; the optical signal output terminals of each radio frequency sensor access module in RF sensor access modules 1~N are connected to the optical signal input terminals of the N×1 optical coupler; the optical signal output terminal of the N×1 optical coupler is connected to the optical signal input terminal of the photodetector N+2.The electrical signal output terminal of the photodetector N+2 is connected to the second electrical signal input terminal of the digital-to-analog converter; the electrical signal output terminal of the digital-to-analog converter is connected to the electrical signal input terminal of the data processing module; the signal output terminal of the data processing module is connected to the signal input terminals of the pulse extraction unit, the blood pressure estimation unit, and the dynamic tracking and fall prevention warning unit; the data processing module is used to perform format conversion and digital filtering on the received data to obtain respiratory detection data, pulse detection data, and human target motion data; the pulse extraction unit is used to generate a pulse map based on the obtained pulse detection data; the blood pressure estimation unit is used to substitute the obtained pulse detection data into the pre-constructed fitting expression SBP_i=k; Amp_i+c is used to obtain the blood pressure value SBP_i; where k and c are preset coefficients, and Amp_i is the pulse detection data; the dynamic tracking and fall prevention warning unit is used to obtain distance image sequences of different parts of the human body based on the human target motion data, and simultaneously extract micro-Doppler features; calculate the position coordinates of the human body based on the distance image sequence; obtain the coordinates of a specific physiological part based on the position coordinates of the human body, and send control commands to the steering mechanism of the optical sensing access module to control the detection beam of the optical sensing access module to turn towards the coordinates of the specific physiological part; input the distance image sequence and the micro-Doppler features into the fall recognition model to obtain the fall recognition result.

[0007] Specifically, the data processing module includes:

[0008] The format conversion submodule is used to convert the format of the received data, remove outliers by setting a threshold, and remove points that exceed the threshold.

[0009] The digital filtering submodule is used to split the signal output by the format conversion submodule into two paths. One path uses different preset center frequencies to digitally filter the signal to obtain the intermediate frequency digital signal of respiratory detection and the intermediate frequency digital signal of pulse detection. The other path performs a fast Fourier transform in the distance dimension to obtain human target motion data.

[0010] The time-frequency analysis submodule is used to perform short-time Fourier transform on the intermediate frequency digital signals of the respiratory detection and the pulse detection respectively to obtain the corresponding respiratory signal time-frequency curve and pulse signal time-frequency curve;

[0011] The respiratory sign extraction submodule is used to smooth the time-frequency curve of the respiratory signal, filter out noise, and calculate the oscillation frequency of the curve by taking the occurrence of a maximum and minimum value of the time-frequency curve of the respiratory signal as a cycle, and then extract respiratory sign information.

[0012] The pulse sign extraction submodule is used to smooth the pulse signal time-frequency curve, filter out noise, and calculate the curve oscillation frequency by taking the occurrence of a maximum and minimum value of the pulse signal time-frequency curve as a cycle, thereby extracting pulse sign information.

[0013] Specifically, the pulse extraction unit includes:

[0014] The waveform extraction subunit is used to extract the envelope of the pulse signal time-frequency curve output by the time-frequency analysis submodule to obtain the pulse signal waveform at each pulse position.

[0015] The waveform fusion subunit is used to align the obtained pulse signal waveforms on the time axis after delay compensation to generate the final pulse image.

[0016] Specifically, the dynamic tracking and fall prevention early warning unit includes:

[0017] The range dimension resolution subunit is used to perform range dimension resolution on the human target motion data through the inverse synthetic aperture radar imaging algorithm to obtain range image sequences of different parts of the human body.

[0018] A position coordinate calculation subunit is used to calculate the position coordinates of the human body based on the distance image sequence;

[0019] The physiological part coordinate acquisition subunit is used to input the position coordinates of the human body into a preset human body model to obtain the coordinates of the specific physiological part.

[0020] The dynamic tracking command output subunit is used to send control commands to the steering mechanism of the optical sensing access module according to the coordinates of the specific physiological part, so as to control the detection beam of the optical sensing access module to turn towards the coordinates of the specific physiological part.

[0021] A micro-Doppler feature extraction subunit is used to synchronously extract micro-Doppler features from the human target motion data;

[0022] The fall recognition subunit is used to input the distance image sequence and the micro-Doppler features into the fall recognition model to obtain the fall recognition result.

[0023] Specifically, the position coordinate calculation subunit is used to process the one-dimensional distance image output by each of the radio frequency sensing access modules to extract the distance information of multiple scattering points on the human body; to associate the distances of different radio frequency sensing access modules corresponding to the same scattering point, and to obtain the three-dimensional coordinates of the scattering point through triangulation; to repeat the above process for all scattering points to obtain the three-dimensional coordinates of multiple points on the human body, thereby reconstructing the posture and position of the human body.

[0024] Specifically, the dynamic tracking and fall prevention warning unit also includes:

[0025] The echo signal strength analysis subunit is used to analyze the strength of the received echo signal.

[0026] The beam pointing command fine-tuning subunit is used to fine-tune the control command to maximize the strength of the received echo signal.

[0027] Specifically, the dynamic tracking and fall prevention warning unit also includes:

[0028] The fall warning subunit is used to send a collaborative confirmation command to the front-end detection modules of other perspectives for collaborative detection if the fall recognition subunit identifies a fall; and to issue an alarm if the fall recognition results based on the collaborative detection of M preset front-end detection modules are all falls.

[0029] Specifically, the optical domain spectrum manipulation module includes: a laser, a phase modulator 1, a phase modulator 2, a microwave signal source 1, a microwave signal source 2, an optical filter 1, an optical filter 2, and a first-stage 1×2 beam splitter; the signal output terminal of the laser is connected to the first signal input terminal of the phase modulator 1, and the signal output terminal of the microwave signal source 1 is connected to the second signal input terminal of the phase modulator 1; the signal output terminal of the phase modulator 1 is connected to the signal input terminal of the optical filter 1; the signal output terminal of the optical filter 1 is connected to the first signal input terminal of the phase modulator 2, and the signal output terminal of the microwave signal source 2 is connected to the second signal input terminal of the phase modulator 2; the signal output terminal of the phase modulator 2 is connected to the signal input terminal of the optical filter 2; the signal output terminal of the optical filter 2 is connected to the signal input terminal of the first-stage 1×2 beam splitter, and the first signal output terminal of the first-stage 1×2 beam splitter is connected to the signal input terminals of the second-stage 1×2 beam splitter 1 and the 1×N beam splitter 2.

[0030] Specifically, the radio frequency sensing access module includes: a photodetector N, a noise amplifier N, an antenna front-end N, and an intensity modulator N; the signal input terminal of the photodetector N is connected to the signal output terminal of the splitter in the third-stage 1×2 beam splitter 1~N, the signal output terminal of the photodetector N is connected to the signal input terminal of the noise amplifier N, the signal output terminal of the noise amplifier N is connected to the signal input terminal of the antenna front-end N; the signal output terminal of the antenna front-end N is connected to the first signal input terminal of the intensity modulator N, the second signal input terminal of the intensity modulator N is connected to the signal output terminal of the optical delay unit in the optical delay unit N+1~2N; and the signal output terminal of the intensity modulator N is connected to the signal input terminal of the N×1 optical coupler.

[0031] Specifically, it also includes: an intelligent diagnostic unit; the signal input terminal of the intelligent diagnostic unit is connected to the signal input terminals of the data processing module, the pulse extraction unit, and the blood pressure estimation unit, and is used to acquire the detection data output by the data processing module, the pulse extraction unit, and the blood pressure estimation unit, and extract human feature parameters from the detection data; the human feature parameters are input into a pre-trained deep learning diagnostic model, and the output results are syndrome classification, health risk warning, and physiological state assessment conclusions.

[0032] One or more technical solutions provided in this invention have at least the following technical effects or advantages:

[0033] 1. This system consists of a central processing module, a line distribution module, and a front-end detection module. The front-end detection module comprises multiple vital sign detection sensor access modules, including optical sensor access modules 1-N and radio frequency sensor access modules 1-N. These sensor access modules, based on microwave photonics technology, emit optical or microwave signals into the detection area and receive their echoes to obtain the raw detection signals. The central processing module includes a signal generation and processing unit, a pulse extraction unit, a blood pressure estimation unit, and a dynamic tracking and fall prevention early warning unit. It primarily performs the functions of generating and processing detection signals, pulse extraction, blood pressure estimation, and fall prevention monitoring and early warning processing based on high-precision distance imaging and micro-Doppler analysis. The line distribution module mainly connects each sensor access module to the central processing module and adds appropriate delays to each link. It consists of multiple beam splitters and optical delayers. This module is constructed of fiber optic devices, resulting in low loss and facilitating long-distance deployment. This system can be flexibly deployed according to actual conditions. The front-end detection modules can be distributed in various rooms for detecting vital signs; the wiring modules can be laid separately or utilize existing optical fibers and devices in existing optical communication networks; the central processing module can be deployed in the control center for controlling signal generation and data monitoring. This invention achieves contactless, respiratory and pulse detection, simultaneous acquisition of human pulse information, rapid estimation of systolic blood pressure, fall prevention, and dynamic tracking continuous measurement functions.

[0034] 2. Through a distributed node network, large-scale, multi-target vital sign monitoring is achieved. Fiber optic devices are used for distributed deployment, resulting in low deployment (modification) costs.

[0035] 3. Utilizing the "ultra-wideband radar positioning + optical beam scanning" architecture, the front-end detection module integrates an optical phased array or a two-dimensional mechanical steering structure, enabling its detection beam to have dynamic tracking and scanning capabilities. This breaks through the traditional monitoring requirements for a static state and achieves seamless, continuous vital sign and unobtrusive safety monitoring of users in a free-moving state.

[0036] 4. Through multimodal data fusion and intelligent diagnostic models, consistent and repeatable auxiliary diagnostic results can be provided, which is expected to realize intelligent diagnosis of traditional Chinese medicine and help to "accurately perceive" and "proactively warn" sub-health status and early disease.

[0037] In summary, this invention provides a multifunctional distributed multi-vital sign non-contact detection radar system that integrates distributed detection, multi-vital sign monitoring, intelligent diagnosis, fall prevention, and dynamic tracking continuous measurement. This system greatly improves the cost-effectiveness and application potential of the system and can be widely used in fields such as smart elderly care, clinical monitoring, health management, and home healthcare. Attached Figure Description

[0038] Figure 1 This is an overall structural block diagram of a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention;

[0039] Figure 2 This is a structural block diagram of the optical domain spectrum manipulation module in the multifunctional distributed multi-vital sign non-contact detection radar system provided in the embodiments of the present invention;

[0040] Figure 3 This is a structural block diagram of the radio frequency sensing access module in the multifunctional distributed multi-vital sign non-contact detection radar system provided in the embodiments of the present invention;

[0041] Figure 4 A schematic diagram illustrating the application of the multifunctional distributed multi-vital sign non-contact detection radar system provided in this embodiment of the invention to hospital ward monitoring;

[0042] Figure 5 A flowchart illustrating the processing of respiratory frequency and pulse frequency in a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention.

[0043] Figure 6 This is a schematic diagram of pulse measurement in a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention;

[0044] Figure 7 This is a flowchart of pulse signal data processing in a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention;

[0045] Figure 8 This is a flowchart of the systolic blood pressure measurement data processing in a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention;

[0046] Figure 9 A schematic diagram of pulse waves in a multifunctional distributed multi-vital sign non-contact detection radar system provided in an embodiment of the present invention;

[0047] Figure 10 A flowchart illustrating the dynamic tracking and fall prevention early warning process in a multifunctional distributed multi-vital sign non-contact detection radar system provided in this embodiment of the invention.

[0048] Figure 11 A flowchart illustrating the intelligent diagnostic process in a multifunctional distributed multi-vital sign non-contact detection radar system provided in this embodiment of the invention. Detailed Implementation

[0049] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0050] like Figure 1As shown, the multifunctional distributed multi-vital sign non-contact detection radar system provided in this embodiment of the invention includes: a central processing module, a line distribution module, and a front-end detection module; the central processing module includes: an optical domain spectrum manipulation module, a second-stage 1×2 beam splitter 1, an optical circulator, a 2×1 optical coupler, photodetectors N+1, N×1 optical couplers, photodetectors N+2, a digital-to-analog converter, a data processing module, a pulse extraction unit, a blood pressure estimation unit, and a dynamic tracking and fall prevention early warning unit; wherein, the optical domain spectrum manipulation module, the second-stage 1×2 beam splitter 1, the optical circulator, the 2×1 optical coupler, photodetectors N+1, N×1 optical couplers, photodetectors N+2, the digital-to-analog converter, and the data processing module constitute the signal generation and processing unit.The line distribution module includes: a 1×N beam splitter 1, optical delay units 1~N, a 1×N beam splitter 2, a third-stage 1×2 beam splitter 1~N, and optical delay units N+1~2N; the front-end detection module includes: optical sensor access modules 1~N and radio frequency sensor access modules 1~N; the first optical signal output terminal of the optical domain spectrum manipulation module is connected to the optical signal input terminal of the second-stage 1×2 beam splitter 1, the first optical signal output terminal of the second-stage 1×2 beam splitter 1 is connected to the optical signal input terminal of the optical circulator, and the optical circulator is also connected to the bidirectional optical path of the 1×N beam splitter 1; the 1×N beam splitter 1 is also connected to the bidirectional optical path of each optical delay unit in the optical delay units 1~N; each optical delay unit in the optical delay units 1~N is also connected to the optical sensor access modules 1~N. Each optical sensing access module is connected to a corresponding bidirectional optical path; the second optical signal output of the second-stage 1×2 beam splitter 1 is connected to the first optical signal input of the 2×1 optical coupler; the optical signal output of the optical circulator is connected to the second optical signal input of the 2×1 optical coupler; the optical signal output of the 2×1 optical coupler is connected to the optical signal input of photodetector N+1, and the electrical signal output of photodetector N+1 is connected to the first electrical signal input of the digital-to-analog converter; the second optical signal output of the optical domain spectrum manipulation module is connected to the optical signal input of the 1×N beam splitter 2, and the optical signal output of the 1×N beam splitter 2 is connected to the optical signal input of each beam splitter in the third-stage 1×2 beam splitter 1~N; the third-stage 1×2 beam splitter... The first signal output terminal of each beam splitter in optical units 1 to N is connected one-to-one with the optical signal input terminal of each optical delay unit in optical delay units N+1 to 2N; the optical signal output terminal of each optical delay unit in optical delay units N+1 to 2N is connected one-to-one with the optical signal input terminal of each radio frequency sensor access module in radio frequency sensor access modules 1 to N; the second signal output terminal of each beam splitter in the third-stage 1×2 beam splitter 1 to N is connected one-to-one with the optical signal input terminal of each radio frequency sensor access module in radio frequency sensor access modules 1 to N; the optical signal output terminal of each radio frequency sensor access module in radio frequency sensor access modules 1 to N is connected to the optical signal input terminal of an N×1 optical coupler; the optical signal output terminal of the N×1 optical coupler is connected to the photodetector. The optical signal input terminal of the photodetector N+2 is connected; the electrical signal output terminal of the photodetector N+2 is connected to the second electrical signal input terminal of the digital-to-analog converter; the electrical signal output terminal of the digital-to-analog converter is connected to the electrical signal input terminal of the data processing module; the signal output terminal of the data processing module is connected to the signal input terminals of the pulse extraction unit, the blood pressure estimation unit, and the dynamic tracking and fall prevention warning unit; the data processing module is used to perform format conversion and digital filtering on the received data to obtain respiratory detection data, pulse detection data, and human target motion data; the pulse extraction unit is used to fuse and generate a pulse map based on the obtained pulse detection data; the blood pressure estimation unit is used to substitute the obtained pulse detection data into the pre-constructed fitting expression SBP_i=k. Amp_i+c yields the blood pressure value SBP_i; where k and c are preset coefficients, and Amp_i is the pulse detection data. The dynamic tracking and fall prevention warning unit is used to obtain distance image sequences of different parts of the human body based on human target motion data, and simultaneously extract micro-Doppler features; the position coordinates of the human body are calculated based on the distance image sequence; the coordinates of specific physiological parts are obtained based on the position coordinates of the human body, and control commands are sent to the steering mechanism of the optical sensing access module to control the detection beam of the optical sensing access module to point to the coordinates of the specific physiological parts; the distance image sequence and micro-Doppler features are input into the fall recognition model to obtain the fall recognition result. When pulse signal detection is performed, the target light sideband generated by the optical domain spectrum manipulation module passes through the second-stage 1×2 beam splitter 1, one path is input at port 1 and output at port 2 of the optical circulator, and then divided into N branches by a 1×N beam splitter in the line distribution module. Each branch is further delayed by an optical delay unit before the optical sensing access module emits an optical signal for pulse signal detection. The other path is sent to the 2×1 optical coupler as the local oscillator optical signal. The received echo optical signal is input at port 2 and output at port 3 of the optical circulator to the 2×1 optical coupler. The local oscillator optical signal and the echo optical signal are coupled by the 2×1 optical coupler and then sent to the photodetector N+1 beat frequency to obtain the intermediate frequency electrical signal. After being converted into a digital signal by the analog-to-digital converter, it is sent to the data processing module for data processing. When detecting respiratory signals, the target light sideband generated by the optical spectrum manipulation module is split into N branches by a 1×N beam splitter. Each branch is then split into two paths by a 1×2 beam splitter. One path is directly sent to the RF sensor access module via port 1. After being converted into a microwave signal by the RF sensor access module, it is emitted outward for respiratory signal detection. The other path is sent to the RF sensor access module via an appropriate delay via an optical delay unit as a local oscillator signal via port 2. The echo signal received by the RF sensor access module is modulated into an optical signal and output via port 3. After passing through an N×1 optical coupler, it is sent to the photodetector for N+2 beat frequency to complete optical deskewing and obtain an intermediate frequency electrical signal. After being converted into a digital signal by an analog-to-digital converter, it is sent to the data processing module for data processing.

[0051] It should be noted that this system's function of rapidly estimating systolic blood pressure involves two stages: a model establishment stage and a routine monitoring stage. In the model establishment stage, the system repeatedly probes the subject's pulse wave signal and simultaneously extracts the amplitude of the first peak of the pulse wave. Simultaneously, a standard blood pressure measurement device (such as a cuff-type electronic blood pressure monitor) is used to simultaneously measure the subject's systolic blood pressure. After extracting the pulse waveform using the pulse measurement data processing workflow, the peak-to-peak value of the first peak in each pulse cycle is read and divided by 2 to obtain the amplitude value of that pulse wave. Multiple blood pressure measurements are performed, and the systolic blood pressure SBP_i and the average amplitude value Amp_i of multiple pulse waves within each blood pressure measurement cycle are read. Linear fitting is then performed on the data to obtain the fitting expression SBP_i=k. Amp_i+c, where k and c are storage coefficients. During routine monitoring, the system uses an optical sensor access module to perform pulse wave testing on the subject and extract the pulse wave amplitude value. This amplitude value is then substituted into the fitting expression to calculate and output the estimated systolic blood pressure value. This method can be applied to any optical sensor access module in the system.

[0052] The structure of the data processing module is described in detail. The data processing module includes:

[0053] The format conversion submodule is used to convert the format of the received data, remove outliers by setting a threshold, and remove points that exceed the threshold.

[0054] The digital filtering submodule is used to split the signal output by the format conversion submodule into two paths. One path uses different preset center frequencies to digitally filter the signal to obtain the intermediate frequency digital signal of respiratory detection and the intermediate frequency digital signal of pulse detection. The other path performs a fast Fourier transform in the distance dimension to obtain human target motion data.

[0055] The time-frequency analysis submodule is used to perform short-time Fourier transform on the intermediate frequency digital signals of respiratory detection and pulse detection from N branches, respectively, to obtain the corresponding respiratory signal time-frequency curve and pulse signal time-frequency curve.

[0056] The respiratory sign extraction submodule is used to smooth the time-frequency curve of the respiratory signal, filter out noise, and calculate the oscillation frequency of the curve by taking the occurrence of a maximum and minimum value of the time-frequency curve of the respiratory signal as a cycle, and then extract the respiratory sign information.

[0057] The pulse sign extraction submodule is used to smooth the pulse signal time-frequency curve, filter out noise, and calculate the oscillation frequency of the curve by taking the occurrence of a maximum and minimum value of the pulse signal time-frequency curve as a cycle, thereby extracting pulse sign information.

[0058] The structure of the pulse extraction unit is described in detail. The pulse extraction unit includes:

[0059] The waveform extraction subunit is used to extract the envelope of the pulse signal time-frequency curve output by the time-frequency analysis submodule to obtain the pulse signal waveform at each pulse position.

[0060] The waveform fusion subunit is used to align the obtained pulse signal waveforms on the time axis after delay compensation to generate the final pulse image.

[0061] It should be noted that when implementing the pulse image extraction function, at least three (the specific number depends on the actual detection needs) optical sensing access modules in this system are respectively aligned with any position (such as the "cun", "guan", and "chi" positions of the radial artery in the human wrist). The signals of each optical sensing access module are processed, and the pulse image extraction unit separates and extracts the pulse waves of their respective independent positions (such as "cun", "guan", and "chi") to form a pulse image.

[0062] The structure of the dynamic tracking and fall prevention warning unit is described in detail. The dynamic tracking and fall prevention warning unit includes:

[0063] The range dimension resolution subunit is used to perform range dimension resolution on human target motion data through inverse synthetic aperture radar imaging algorithm to obtain range image sequences of different parts of the human body (such as head, torso, and limbs).

[0064] The position coordinate calculation subunit is used to calculate the position coordinates of the human body based on the distance image sequence;

[0065] Specifically, the position coordinate calculation subunit is used to process the one-dimensional distance image output by each radio frequency sensor access module, extract the distance information of multiple scattering points on the human body (such as head, torso, limbs, etc.); associate the distances of different radio frequency sensor access modules corresponding to the same scattering point, and obtain the three-dimensional coordinates of the scattering point through triangulation (spherical intersection); repeat the above process for all scattering points to obtain the three-dimensional coordinates of multiple points on the human body, thereby reconstructing the posture and position of the human body.

[0066] The physiological location coordinate acquisition sub-unit is used to input the position coordinates of the human body into a preset human body model to obtain the coordinates of a specific physiological location (such as the radial artery in the wrist).

[0067] The dynamic tracking command output subunit is used to send control commands to the steering mechanism of the optical sensing access module according to the coordinates of a specific physiological part, so as to control the detection beam of the optical sensing access module to point to the coordinates of the specific physiological part and realize seamless continuous measurement.

[0068] The micro-Doppler feature extraction subunit is used to simultaneously extract micro-Doppler features from human target motion data;

[0069] The fall detection subunit is used to input distance image sequences and micro-Doppler features into the fall detection model to obtain fall detection results. This model identifies fall events by analyzing the abrupt collapse in the height direction of the human body, the incoordination of limb movements, and the micro-Doppler spectral features of the impact.

[0070] To ensure the accuracy of dynamic tracking, the dynamic tracking and fall prevention warning unit also includes:

[0071] The echo signal strength analysis subunit is used to analyze the strength of the received echo signal.

[0072] The beam pointing command fine-tuning subunit is used to fine-tune the control commands to maximize the strength of the received echo signal.

[0073] To improve the accuracy of fall warnings, the dynamic tracking and fall prevention warning unit also includes:

[0074] The fall warning subunit is used to send a collaborative confirmation command to other front-end detection modules for collaborative detection if the fall recognition subunit identifies a fall. If the fall recognition results based on collaborative detection by M preset front-end detection modules all indicate a fall, an alarm is issued. Specifically, when a single node identifies a suspected fall event, it requests multi-view collaborative confirmation from neighboring nodes through a distributed network. Only if multiple nodes determine it to be high-risk is a high-level alarm triggered.

[0075] The structure of the optical domain spectrum manipulation module will be described in detail, such as... Figure 2As shown, the optical domain spectrum manipulation module includes: a laser, phase modulator 1, phase modulator 2, microwave signal source 1, microwave signal source 2, optical filter 1, optical filter 2, and a first-stage 1×2 beam splitter; the signal output terminal of the laser is connected to the first signal input terminal of phase modulator 1, and the signal output terminal of microwave signal source 1 is connected to the second signal input terminal of phase modulator 1; the signal output terminal of phase modulator 1 is connected to the signal input terminal of optical filter 1; the signal output terminal of optical filter 1 is connected to the first signal input terminal of phase modulator 2, and the signal output terminal of microwave signal source 2 is connected to the second signal input terminal of phase modulator 2; the signal output terminal of phase modulator 2 is connected to the signal input terminal of optical filter 2; the signal output terminal of optical filter 2 is connected to the signal input terminal of the first-stage 1×2 beam splitter, and the first signal output terminal of the first-stage 1×2 beam splitter is connected to the signal input terminals of the second-stage 1×2 beam splitter 1 and the 1×N beam splitter 2. Its working principle is as follows: The laser generates single-frequency continuous light as the carrier of phase modulator 1. This light is phase-modulated by the microwave signal generated by microwave signal source 1 through phase modulator 1. The modulated optical signal is then sent to dual-bandpass filter 1 to filter out the desired optical sideband pairs. This optical sideband signal serves as the carrier of phase modulator 2, which is also phase-modulated by the microwave signal generated by microwave signal source 2. The modulated optical signal then passes through bandpass filter 2, filtering out the desired optical sidebands before entering the first stage. The beam splitters send the beams to the second-stage 1×2 beam splitter and 1×N beam splitter, respectively.

[0076] In this embodiment, microwave signal source 1 and microwave signal source 2 can be implemented by independent signal sources or by a single multi-output signal source. The laser is any one of a semiconductor laser, a fiber laser, or other single-frequency laser. The dual-bandpass filter and the bandpass filter are any one of a fiber optic grating, a programmable optical filter, or other optical filters.

[0077] The structure of the radio frequency sensing access module is described in detail, such as... Figure 3As shown, the RF sensing access module includes: a photodetector N, a noise amplifier N, an antenna front-end N, and an intensity modulator N. The signal input terminal of the photodetector N is connected to the signal output terminal of the splitter in the third-stage 1×2 beam splitter 1~N; the signal output terminal of the photodetector N is connected to the signal input terminal of the noise amplifier N; the signal output terminal of the noise amplifier N is connected to the signal input terminal of the antenna front-end N; the signal output terminal of the antenna front-end N is connected to the first signal input terminal of the intensity modulator N; the second signal input terminal of the intensity modulator N is connected to the signal output terminal of the optical delay unit in the optical delay unit N+1~2N; and the signal output terminal of the intensity modulator N is connected to the signal input terminal of the N×1 optical coupler. The optical signal directly entering the RF sensing access module is converted into an electrical signal by the photodetector N and then radiated outward through the low-noise amplifier N and the antenna front-end N for breathing detection. The echo signal is received by the antenna front-end N and input to the intensity modulator N to modulate the local oscillator signal.

[0078] In this embodiment, the microwave antenna in the antenna front end can be a single element antenna or an array antenna.

[0079] In order to realize the intelligent diagnostic function, it also includes: intelligent diagnostic unit; the signal input end of intelligent diagnostic unit is connected to the signal input end of data processing module, pulse extraction unit and blood pressure estimation unit, and is used to obtain the detection data output by data processing module, pulse extraction unit and blood pressure estimation unit, and extract human feature parameters from the detection data; the human feature parameters are input into the pre-trained deep learning diagnostic model, and the output results are syndrome classification, health risk warning and physiological state assessment conclusion. Specifically, a set of multi-dimensional feature vectors are extracted from the obtained respiratory rate, pulse rate, cun / guan / chi pulse waveforms and estimated blood pressure data. The vector includes, but is not limited to, respiratory rate, respiratory amplitude, pulse rate, pulse amplitude, pulse rhythm, pulse shape, pulse momentum, heart rate variability, blood pressure and other human feature parameters extracted from respiratory waves and pulse waves. The vector is input into the pre-trained deep learning diagnostic model, and the model output is at least one of the following diagnostic results: (1) TCM syndrome classification, such as the probability distribution of syndromes such as "liver yang hyperactivity", "qi and blood deficiency" and "phlegm and dampness accumulation"; (2) health risk warning. For example, risk labels such as "increased cardiovascular risk level" and "possibility of sleep disorder" are identified; (3) Physiological status assessment, such as quantitative scores of "stress level" and "fatigue level". The training process of this deep learning diagnostic model is as follows: (1) Constructing an expert-annotated dataset. This system collects a large amount of multimodal data (pulse, respiration, blood pressure, etc.) from patients, and an expert committee (composed of several senior TCM doctors) gives a unified syndrome diagnosis label for each set of data. (2) Model training. Using a deep learning framework (such as TensorFlow, PyTorch), the feature vectors corresponding to the multimodal data are used as input, and the syndrome labels annotated by experts are used as the target output to train the network parameters until the model can accurately map from the data to the diagnostic conclusion. (3) Model verification and deployment. The diagnostic accuracy of the model is verified on an independent test set. After the performance meets the standard, the model is deployed in the intelligent diagnostic unit of the system.

[0080] To achieve system integration, by combining optoelectronic hybrid integration technology, lasers, phase modulators, optical filters, optical delayers, beam splitters, optical couplers, photodetectors, RF amplifiers, and signal sources can be partially or completely integrated into a single module, thereby reducing the system size.

[0081] In this embodiment, the optical sensing access module can be a single fiber collimator, an array of multiple fiber collimators, or other optical antenna front-ends. The optical delay unit can be a programmable optical delay unit, a manually adjustable optical delay unit, or a delay fiber.

[0082] To facilitate public understanding, the technical solution of this invention will be further explained in detail theoretically, taking the generation of two symmetrical chirped optical signals by the optical domain spectrum manipulation module as an example.

[0083] The optical spectrum manipulation module outputs a frequency-centrically symmetrical double-chirped optical signal, which can be represented as:

[0084] (1)

[0085] in, This indicates the amplitude of the output double-chirped signal. , , Indicates the center frequency of the laser. This represents the frequency difference between the two sidebands of a double-chirped signal. This indicates the frequency modulation slope. The above-mentioned dual-chirped optical signal is generated by the first stage... The optical splitter is divided into two channels for pulse signal measurement and respiratory signal measurement, respectively.

[0086] When measuring the pulse signal, the aforementioned dual-chirped optical signal is transmitted to the nth optical sensing access module for outward transmission. The transmitted signal can be represented as follows:

[0087] (2)

[0088] in, Indicates the amplitude of the emitted light signal. The link delay between the second-stage 1×2 beam splitter 1 and the nth optical sensor access module is given. The local oscillator optical signal is sent to the 2×1 coupler. Assume the link delay from the second-stage 1×2 beam splitter 1 to the 2×1 coupler is... The local oscillator optical signal input from the 2×1 coupler can be expressed as:

[0089] (3)

[0090] The human pulse causes periodic micro-movements in the skin. When the optical sensor interface is close to the skin near the pulse point, the echo signal reflected from the skin received by the nth optical sensor interface module can be represented as...

[0091] (4)

[0092] in, Indicates the amplitude of the echo signal. This indicates the phase change caused by skin reflection. The time delay is caused by the periodic movement of the skin near the pulse, and can be expressed as...

[0093] (5)

[0094] in, This represents the distance between the optical sensing module and the skin, assuming no micro-movements. This indicates periodic displacement caused by micro-movements of the skin. Represents the speed of light. Distance resolution. Where B is the bandwidth of the optical signal. For optical signals with bandwidths on the order of tens of GHz, the distance resolution is on the order of centimeters to decimeters. The skin twitches caused by a human pulse are generally less than 1 millimeter. Therefore, the displacement caused by skin twitches... It is much smaller than the distance resolution, so the second term in formula (5) can be ignored. It is the Doppler frequency caused by skin micro-movements, which can be expressed as

[0095] (6)

[0096] in, The radial velocity representing the micro-movement of the skin. The frequency of the optical signal used for pulse detection is approximately 193.5 THz. Compared to microwave signals, which typically have frequencies in the GHz range, the Doppler frequency shift generated by using an optical signal as the detection signal is significantly enhanced. The echo optical signal and the local oscillator optical signal are coupled via a 2×1 optical coupler and mixed at point N+1 of the photodetector. The output intermediate frequency signal can be expressed as...

[0097] (7)

[0098] in, This indicates the amplitude of the signal after descrambling. and Let and represent the phases of the echo signal, respectively. From the above equation, the frequency of the intermediate frequency signal finally obtained by the nth optical sensor access module can be expressed as:

[0099] (8)

[0100] make Formula (8) can be written as

[0101] (9)

[0102] After data processing, the above signals yield the frequency variation curve of the intermediate frequency (IF) signal of the pulse signal over time. Based on the above analysis, the IF signal frequency of the pulse signal will be delayed... Symmetry. The periodic skin micro-movements caused by the pulse produce periodic Doppler frequency shifts, both with the same period. Therefore, the periodicity of the intermediate frequency signal is the same as that of the pulse signal. The pulse rate can be extracted by observing the frequency changes of the intermediate frequency signal detected by the pulse. By appropriately setting the delay amounts of optical delay units 1 to N, the frequencies of the intermediate frequency signals detected by N branches when detecting the pulse can be distinguished in the frequency domain, thereby enabling simultaneous measurement of pulse signals from multiple individuals (multiple test points). Assuming the delay amounts of optical delay units 1 to N differ by... The center frequencies of each branch differ by a factor of 1.

[0103] (10)

[0104] To ensure complete separation of signals in the frequency domain, the following must be met:

[0105] (11)

[0106] When measuring respiratory signals, the dual-chirped optical signal output from the aforementioned optical domain spectrum manipulation module is split into N branches by a 1×N beam splitter. Each branch is further split into two paths by a third-stage 1×2 beam splitter. One path is sent to the intensity modulator in the RF sensor access module via an optical delay unit as the local oscillator carrier signal. The other path is sent to the photodetector in the RF sensor access module, where the beat frequency is converted into an RF signal. This signal is then radiated outward through a low-noise amplifier and the antenna front end to detect human respiratory signals. The RF signal radiated outward by the nth RF sensor access module can be expressed as...

[0107] (12)

[0108] in, This represents the time delay caused by link delay from the 1×N beam splitter 2 to the front end of the nth branch antenna. When a human body performs respiratory movements, the echo signal caused by the periodic displacement of the chest cavity can be represented as...

[0109] (13)

[0110] in, Indicates the amplitude of the echo signal. It is the time delay difference between the radio frequency signal transmitted by the antenna front end and the received radio frequency signal. It is the Doppler frequency shift caused by the periodic movement of the human chest cavity. This is the phase change caused by signal reflection. The received radio frequency signal is used as a modulation signal and input to the intensity modulator to modulate the local oscillator optical carrier. The output of the intensity modulator in the nth branch can be expressed as...

[0111] (14)

[0112] in, It is the delay from the third-stage 1×2 beam splitter of the nth branch to the input of the intensity modulator. It is the half-wave voltage of the intensity modulator. The intensity modulator operates with quadrature bias, retaining ±1st order sidebands. The frequency of the intermediate frequency signal obtained after three beats by the photodetector can be expressed as:

[0113] (15)

[0114] Microwave signals are used to detect human respiratory movements. The frequency of microwave signals is on the order of tens of GHz, while the Doppler frequency is on the order of Hz. It can be ignored. The periodic changes that occur with the movement of the chest cavity during respiration can be represented as follows:

[0115] (16)

[0116] in, This indicates the displacement between the antenna tip and the human chest cavity when there is no breathing. This represents the displacement caused by the periodic movement of the chest cavity during respiration. This system utilizes ultra-wideband microwave signals for respiration detection, achieving a range resolution on the order of millimeters (e.g., a range resolution of 7.5 mm corresponding to a 20 GHz bandwidth), sufficient to resolve the approximately 1-2 cm fluctuation displacement of the chest cavity during respiration. Therefore, the second term in equation (16) cannot be ignored. Let... Formula (14) can be simplified to

[0117] (17)

[0118] After data processing, the above signals yield the frequency change curve of the intermediate frequency signal of the respiratory signal over time. The periodic chest displacement caused by respiration leads to a periodic change in the frequency of the respiratory intermediate frequency signal. Since the two have the same period, the respiratory rate can be extracted by the frequency change of the intermediate frequency signal detected by respiration.

[0119] For the same respiratory detection branch, the distance between different individuals and the antenna front end is different, i.e. Differences lead to The differences are reflected in the different frequencies of the intermediate frequency (IF) signals. Therefore, the IF signals of respiration from different individuals detected by the same branch can be distinguished in the frequency domain, thus enabling simultaneous measurement of respiration signals from multiple individuals. In different respiration detection branches, by rationally designing the delay amount of the optical delay unit from N+1 to 2N, the IF signals from each branch can be... Different causes Because they are different, respiratory signals detected by different branches can be distinguished in the frequency domain.

[0120] The following specific embodiments illustrate the multifunctionality achievable by this system:

[0121] Example 1: A multifunctional distributed multi-vital sign non-contact radar system for monitoring hospital wards

[0122] like Figure 4As shown, the central processing module of this system is located at the nurses' station or central monitoring station, equipped with a vital signs monitoring result display interface. Medical staff can monitor the vital signs of all patients in the ward in real time through the display interface. The system can be configured with an alarm function. By setting a threshold, when the monitored data exceeds the threshold, the system will issue an alarm, facilitating timely treatment and rescue by medical staff. The front-end detection modules in this system are deployed in each ward and bed to achieve coverage of the entire space. The front-end detection modules transmit and receive vital signs detection signals, and the signals received by each sensor access module are uniformly transmitted to the central processing module for data processing and result display. Among them, only one radio frequency sensor access module needs to be deployed in each ward, which can monitor the respiratory signals of multiple patients in the ward. One optical sensor access module is deployed in each bed, and one optical sensor access module can monitor the pulse and blood pressure of one patient. The front-end detection modules and the central processing module are connected by a line distribution module, which is composed of optical fiber and optical fiber devices, with low loss and easy long-distance laying. The line distribution module can be laid independently as needed, or it can be multiplexed with the optical fiber devices in the existing optical communication network.

[0123] Specifically, the processing flow for respiratory rate and pulse rate is as follows: Figure 5 As shown:

[0124] (1) Data preprocessing: The format of the output data of the analog-to-digital converter is converted, and outliers are removed by setting a threshold to remove points that exceed the threshold;

[0125] (2) Digital filtering: The preprocessed data is first digitally filtered using digital bandpass filters with different center frequencies set in advance, and the filtered data are filtered out respectively. The intermediate frequency digital signals corresponding to the respiratory and pulse detection of each branch;

[0126] (3) Time-frequency analysis: for the filtered... The respiratory and pulse signals of each branch were subjected to short-time Fourier transforms to obtain... The respiratory signal time-frequency curve and pulse signal time-frequency curve of each branch;

[0127] (4) Parameter extraction: The time-frequency curve after time-frequency analysis is smoothed and noise is filtered out. The oscillation frequency of the curve is calculated by taking the occurrence of a maximum and minimum value of the time-frequency curve as a period, and vital sign information is extracted. The respiratory rate is obtained after processing the respiratory signal, and the pulse rate is obtained after processing the pulse signal.

[0128] (5) Display and store: Display or store respiratory rate and pulse rate as needed.

[0129] Traditional care solutions require the purchase of multiple independent monitoring devices, and typically one device can only be used by one patient at a time. This invention, however, uses a single hardware system (distributed detection module + central processing module) to simultaneously monitor respiration, pulse, and blood pressure, significantly reducing equipment procurement costs. This system can continuously and automatically monitor vital signs, changing the inefficient traditional model that relies on nurses' regular rounds and manual measurement and recording. One system requires only one or a small number of personnel to monitor the entire area via a central screen, eliminating the need for frequent room visits and disruptions to rest, greatly improving per capita care efficiency and reducing repetitive workload for nursing staff.

[0130] In addition to monitoring in hospital wards, this system can also be applied to nursing homes, community health monitoring, and other scenarios.

[0131] Example 2: Pulse Diagnosis Acquisition Function

[0132] like Figure 6 As shown, when pulse information of the subject needs to be obtained, three (or six, for both hands) optical sensor access modules (1, 2, 3) in the system are precisely aligned with the radial artery at the "cun," "guan," and "chi" points of one wrist of the subject, respectively. These optical sensor access modules can be fixedly installed or mounted on a finely adjustable robotic arm to accommodate anatomical differences between individuals. According to the system's measurement principle, each optical sensor access module emits a laser signal to the corresponding wrist position and receives the micro-Doppler effect echo signal generated by the skin epidermis near the radial artery pulsating with the pulse. In the central processing module, according to... Figure 7 The signal processing flow shown performs data processing and ultimately generates a comprehensive pulse map.

[0133] Example 3: Rapid Systolic Blood Pressure Estimation Function

[0134] like Figure 8 As shown, the blood pressure estimation module operates in two phases:

[0135] A. Model building phase:

[0136] The user sits still, the optical sensor access module is aligned with the radial artery of the subject, and a standard cuff blood pressure monitor is worn simultaneously. The system continuously acquires pulse waves from the radial artery (or other aortic location) of the subject, and simultaneously performs multiple blood pressure measurements using the standard blood pressure monitor. The central processing module records the average pulse wave amplitude (Amp_i) and the corresponding systolic blood pressure value (SBP_i) for each blood pressure measurement cycle. Multiple tests are repeated to collect multiple sets of (Amp_i, SBP_i) data pairs. Linear fitting is used to fit a personalized model for the user: SBP = k Amp + c. Parameters k and c are stored.

[0137] B. Routine monitoring phase:

[0138] A blood pressure monitor is not required during routine blood pressure monitoring. The optical sensor module of this system is aligned with the radial artery (maintaining the same artery position as during the model establishment phase) to perform pulse wave testing and obtain data such as... Figure 9 The waveform shown represents a complete cycle. The pulse wave amplitude is extracted. The pulse wave amplitude is then input into the model established above, using SBP = k... Press Amp + c to instantly calculate the current estimated systolic blood pressure (SBP). This system can display or store the estimated systolic blood pressure value in real time, or plot its trend over time. It also allows setting threshold alarms to issue a warning when the estimated value exceeds the safe range.

[0139] Example 4: Dynamic Tracking and Fall Prevention Early Warning Function

[0140] Both dynamic tracking and fall prevention warning functions rely on acquiring high-precision distance images. This is achieved by using an RF sensor access module to radiate an ultra-wideband linear frequency modulated (LFM) RF signal, receiving the echo, and then using de-slant reception to obtain a high-resolution distance image. This distance image can distinguish the distance between different parts of the human body and the RF sensor access module. For example... Figure 10 As shown, based on the acquisition of high-precision distance images, dynamic tracking and fall prevention warning functions are implemented respectively.

[0141] The fall prevention warning system, based on high-precision distance imaging, tracks the instantaneous rate of change of the human body's center of mass height in real time along the distance dimension. Under normal activity, height changes are gradual, but during a fall, a phenomenon known as "height collapse" occurs, where height drops rapidly within a very short period. Once the echo data from a certain radio frequency (RF) sensor access module is identified by the system as "suspected fall," it immediately sends a coordination request through the central processing module to neighboring RF sensor access modules covering the same area. These distributed RF sensor access modules in the same area independently extract and judge features from different angles. When a majority (which can be set according to accuracy requirements, such as 80%) of the echo data from multiple RF sensor access modules is identified as a fall, the target user's fall is confirmed, and an alarm is immediately triggered (audio-visual, push notification to guardian or emergency center), simultaneously providing the precise location of the fall.

[0142] Based on high-precision distance imaging, dynamic tracking extracts the target user's three-dimensional coordinates (x, y, z) in real time with an accuracy of 1 cm (corresponding to a radio frequency signal bandwidth of 15 GHz). According to the human skeletal model built into the dynamic tracking and fall prevention warning unit, the orientation and limb swing models of the human body are considered. The theoretical spatial coordinates (x_target, y_target, z_target) of the target physiological location (such as the radial artery "guan" point on the left wrist) are calculated in real time. Based on the predicted target coordinates (x_target, y_target, z_target) and the coordinates of the optical sensing access module itself, the required beam pointing angles (azimuth θ and elevation φ) are calculated. For optical sensing access modules using an optical phased array antenna front end, the system controls an optical phase shifter to achieve beam scanning, instantly pointing the beam to the target position. For collimators using mechanical steering devices or other optical antenna front ends, the system drives a servo motor to physically turn the antenna and align it with the target position. After the optical sensing access module is aligned with the target position, the system begins to acquire the pulse wave signal of the target location. By analyzing the quality of the received signal (such as signal-to-noise ratio), the beam direction is fine-tuned to ensure it is always locked onto the position with the strongest signal. During this process, physiological parameters such as pulse and blood pressure are continuously measured and timestamped with the current spatial location and behavioral state (such as stationary or walking), and stored in the database.

[0143] Example 5: Intelligent Diagnostic Function

[0144] like Figure 11 As shown, the workflow of intelligent diagnosis is as follows:

[0145] (1) Synchronous acquisition of multimodal data. Using this system, respiratory signal waves and pulse waveforms of distributed nodes are acquired synchronously.

[0146] (2) Deep Feature Extraction. The respiratory rate, respiratory amplitude, pulse rate, and pulse amplitude of each distributed node are extracted in the data processing module. Pulse parameters and blood pressure parameters are further extracted in the pulse extraction unit and blood pressure estimation unit. A multi-dimensional deep feature vector is constructed using these parameters.

[0147] (3) Intelligent diagnostic model inference. The deep feature vector is input into the pre-trained deep learning diagnostic model. The model outputs a probability distribution of TCM syndromes and a health risk index and physiological state score.

[0148] (4) Diagnostic results output and interpretation. The diagnostic results of the model are presented to the user in a visual form, and confidence level and key evidence can be attached.

[0149] The training process of the deep learning diagnostic model in the intelligent diagnostic unit is as follows:

[0150] (1) Construct an expert-annotated dataset. This system collects a large amount of multimodal data (pulse, respiration, blood pressure, etc.) from patients, and an expert committee (which may consist of several senior TCM doctors) gives a unified syndrome diagnosis label for each set of data.

[0151] (2) Model training. Using a deep learning framework, the deep feature vectors corresponding to the multimodal data are used as input, and the syndrome labels labeled by experts are used as the target output. The network parameters are trained until the model can accurately map from the data to the diagnostic conclusion.

[0152] (3) Verify the diagnostic accuracy of the model on an independent test set. Once the performance meets the standard, deploy the model in the central processing unit of the system.

[0153] In summary, this invention provides a distributed multi-vital sign non-contact health monitoring system that can acquire human pulse information without contact and simultaneously, and can quickly estimate systolic blood pressure based on pulse wave amplitude, prevent falls, dynamically track continuous measurement, and provide intelligent diagnosis. It can simultaneously monitor and distinguish targets in multiple different areas and is particularly suitable for complex scenarios requiring group care, such as communities, hospitals, and nursing homes.

[0154] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0155] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0156] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0157] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0158] Any aspects of this invention not described in detail in the embodiments are well-known techniques to those skilled in the art. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this invention and not to limit it. Although this invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this invention without departing from the spirit and scope of this invention, and all such modifications and substitutions should be covered within the scope of the claims of this invention.

Claims

1. A multifunctional distributed multi-vital sign non-contact detection radar system, characterized in that, include: Central processing module, line distribution module, and front-end detection module; The central processing module includes: an optical spectrum manipulation module, a second-stage 1×2 beam splitter 1, an optical circulator, a 2×1 optical coupler, photodetectors N+1, N×1 optical couplers, photodetectors N+2, a digital-to-analog converter, a data processing module, a pulse extraction unit, a blood pressure estimation unit, and a dynamic tracking and fall prevention early warning unit; the line distribution module includes: a 1×N beam splitter 1, optical delay units 1~N, a 1×N beam splitter 2, a third-stage 1×2 beam splitter 1~N, and optical delay units N+1~2N; the front-end detection module includes: optical sensor access modules 1~N and radio frequency sensor access modules 1~N; the first optical signal output terminal of the optical spectrum manipulation module is connected to the optical signal input terminal of the second-stage 1×2 beam splitter 1. The first optical signal output terminal of the second-stage 1×2 beam splitter 1 is connected to the optical signal input terminal of the optical circulator. The optical circulator is also connected to the bidirectional optical path of the 1×N beam splitter 1. The 1×N beam splitter 1 is also connected to the bidirectional optical path of each optical delay unit among the optical delay units 1 to N. Each optical delay unit among the optical delay units 1 to N is also connected to a corresponding bidirectional optical path of each optical sensing access module among the optical sensing access modules 1 to N. The second optical signal output terminal of the second-stage 1×2 beam splitter 1 is connected to the first optical signal input terminal of the 2×1 optical coupler. The optical signal output terminal of the optical circulator is connected to the second optical signal input terminal of the 2×1 optical coupler. The optical signal output terminal of the 2×1 optical coupler is connected to the... The optical signal input terminal of photodetector N+1 is connected, and the electrical signal output terminal of photodetector N+1 is connected to the first electrical signal input terminal of the digital-to-analog converter; the second optical signal output terminal of the optical domain spectrum manipulation module is connected to the optical signal input terminal of the 1×N beam splitter 2, and the optical signal output terminal of the 1×N beam splitter 2 is connected to the optical signal input terminals of each beam splitter in the third-stage 1×2 beam splitter 1~N; the first signal output terminal of each beam splitter in the third-stage 1×2 beam splitter 1~N is connected one-to-one with the optical signal input terminal of each optical delay unit in the optical delay units N+1~2N; the optical signal output terminal of each optical delay unit in the optical delay units N+1~2N is connected to each of the RF sensing access modules 1~N. The optical signal input terminals of the RF sensor access modules are connected one-to-one; the second signal output terminals of each of the third-stage 1×2 beam splitters 1~N are connected one-to-one with the optical signal input terminals of each of the RF sensor access modules 1~N; the optical signal output terminals of each of the RF sensor access modules 1~N are connected to the optical signal input terminals of the N×1 optocoupler; the optical signal output terminal of the N×1 optocoupler is connected to the optical signal input terminal of the photodetector N+2; the electrical signal output terminal of the photodetector N+2 is connected to the second electrical signal input terminal of the digital-to-analog converter; and the electrical signal output terminal of the digital-to-analog converter is connected to the electrical signal input terminal of the data processing module.The signal output terminal of the data processing module is connected to the signal input terminals of the pulse extraction unit, the blood pressure estimation unit, and the dynamic tracking and fall prevention warning unit; the data processing module is used to perform format conversion and digital filtering on the received data to obtain respiratory detection data, pulse detection data, and human target motion data; the pulse extraction unit is used to generate a pulse map based on the obtained pulse detection data; the blood pressure estimation unit is used to substitute the obtained pulse detection data into a pre-constructed fitting expression SBP_i=k; Amp_i+c is used to obtain the blood pressure value SBP_i; where k and c are preset coefficients, and Amp_i is the pulse detection data; the dynamic tracking and fall prevention warning unit is used to obtain distance image sequences of different parts of the human body based on the human target motion data, and simultaneously extract micro-Doppler features; calculate the position coordinates of the human body based on the distance image sequence; obtain the coordinates of a specific physiological part based on the position coordinates of the human body, and send control commands to the steering mechanism of the optical sensing access module to control the detection beam of the optical sensing access module to turn towards the coordinates of the specific physiological part; input the distance image sequence and the micro-Doppler features into the fall recognition model to obtain the fall recognition result.

2. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 1, characterized in that, The data processing module includes: The format conversion submodule is used to convert the format of the received data, remove outliers by setting a threshold, and remove points that exceed the threshold. The digital filtering submodule is used to split the signal output by the format conversion submodule into two paths. One path uses different preset center frequencies to digitally filter the signal to obtain the intermediate frequency digital signal of respiratory detection and the intermediate frequency digital signal of pulse detection. The other path performs a fast Fourier transform in the distance dimension to obtain human target motion data. The time-frequency analysis submodule is used to perform short-time Fourier transform on the intermediate frequency digital signals of the respiratory detection and the pulse detection respectively to obtain the corresponding respiratory signal time-frequency curve and pulse signal time-frequency curve; The respiratory sign extraction submodule is used to smooth the time-frequency curve of the respiratory signal, filter out noise, and calculate the oscillation frequency of the curve by taking the occurrence of a maximum and minimum value of the time-frequency curve of the respiratory signal as a cycle, and then extract respiratory sign information. The pulse sign extraction submodule is used to smooth the pulse signal time-frequency curve, filter out noise, and calculate the curve oscillation frequency by taking the occurrence of a maximum and minimum value of the pulse signal time-frequency curve as a cycle, thereby extracting pulse sign information.

3. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 2, characterized in that, The pulse extraction unit includes: The waveform extraction subunit is used to extract the envelope of the pulse signal time-frequency curve output by the time-frequency analysis submodule to obtain the pulse signal waveform at each pulse position. The waveform fusion subunit is used to align the obtained pulse signal waveforms on the time axis after delay compensation to generate the final pulse image.

4. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 1, characterized in that, The dynamic tracking and fall prevention early warning unit includes: The range dimension resolution subunit is used to perform range dimension resolution on the human target motion data through the inverse synthetic aperture radar imaging algorithm to obtain range image sequences of different parts of the human body. A position coordinate calculation subunit is used to calculate the position coordinates of the human body based on the distance image sequence; The physiological part coordinate acquisition subunit is used to input the position coordinates of the human body into a preset human body model to obtain the coordinates of the specific physiological part. The dynamic tracking command output subunit is used to send control commands to the steering mechanism of the optical sensing access module according to the coordinates of the specific physiological part, so as to control the detection beam of the optical sensing access module to turn towards the coordinates of the specific physiological part. A micro-Doppler feature extraction subunit is used to synchronously extract micro-Doppler features from the human target motion data; The fall recognition subunit is used to input the distance image sequence and the micro-Doppler features into the fall recognition model to obtain the fall recognition result.

5. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 4, characterized in that, The position coordinate calculation subunit is specifically used to process the one-dimensional distance image output by each of the radio frequency sensing access modules and extract the distance information of multiple scattering points on the human body; By associating the distances of different radio frequency sensor access modules to the same scattering point, the three-dimensional coordinates of the scattering point are obtained through triangulation. The above process is repeated for all scattering points to obtain the three-dimensional coordinates of multiple points on the human body, thereby reconstructing the posture and position of the human body.

6. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 4, characterized in that, The dynamic tracking and fall prevention early warning unit also includes: The echo signal strength analysis subunit is used to analyze the strength of the received echo signal. The beam pointing command fine-tuning subunit is used to fine-tune the control command to maximize the strength of the received echo signal.

7. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 4, characterized in that, The dynamic tracking and fall prevention early warning unit also includes: The fall warning subunit is used to send a collaborative confirmation command to the front-end detection modules of other perspectives for collaborative detection if the fall recognition subunit identifies a fall; and to issue an alarm if the fall recognition results based on the collaborative detection of M preset front-end detection modules are all falls.

8. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 1, characterized in that, The optical spectrum manipulation module includes: a laser, a phase modulator 1, a phase modulator 2, a microwave signal source 1, a microwave signal source 2, an optical filter 1, an optical filter 2, and a first-stage 1×2 beam splitter; the signal output terminal of the laser is connected to the first signal input terminal of the phase modulator 1, and the signal output terminal of the microwave signal source 1 is connected to the second signal input terminal of the phase modulator 1; the signal output terminal of the phase modulator 1 is connected to the signal input terminal of the optical filter 1; the signal output terminal of the optical filter 1 is connected to the first signal input terminal of the phase modulator 2, and the signal output terminal of the microwave signal source 2 is connected to the second signal input terminal of the phase modulator 2; the signal output terminal of the phase modulator 2 is connected to the signal input terminal of the optical filter 2; the signal output terminal of the optical filter 2 is connected to the signal input terminal of the first-stage 1×2 beam splitter, and the first signal output terminal of the first-stage 1×2 beam splitter is connected to the signal input terminals of the second-stage 1×2 beam splitter 1 and the 1×N beam splitter 2.

9. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 1, characterized in that, The radio frequency sensing access module includes: a photodetector N, a noise amplifier N, an antenna front-end N, and an intensity modulator N; the signal input terminal of the photodetector N is connected to the signal output terminal of the splitter in the third-stage 1×2 beam splitter 1~N, the signal output terminal of the photodetector N is connected to the signal input terminal of the noise amplifier N, the signal output terminal of the noise amplifier N is connected to the signal input terminal of the antenna front-end N; the signal output terminal of the antenna front-end N is connected to the first signal input terminal of the intensity modulator N, the second signal input terminal of the intensity modulator N is connected to the signal output terminal of the optical delay unit in the optical delay unit N+1~2N; and the signal output terminal of the intensity modulator N is connected to the signal input terminal of the N×1 optical coupler.

10. The multifunctional distributed multi-vital sign non-contact detection radar system as described in claim 1, characterized in that, Also includes: Intelligent diagnostic unit; The signal input terminal of the intelligent diagnostic unit is connected to the signal input terminals of the data processing module, the pulse extraction unit, and the blood pressure estimation unit. It is used to acquire the detection data output by the data processing module, the pulse extraction unit, and the blood pressure estimation unit, and extract human feature parameters from the detection data. The human feature parameters are then input into a pre-trained deep learning diagnostic model to output syndrome classification, health risk warning, and physiological state assessment conclusions.

Images