Non-contact vital sign monitoring system and method based on mechanically programmable metasurface

Abstract

The invention discloses a non-contact vital sign monitoring system and method based on a mechanical programmable metasurface, and belongs to the technical field of intelligent medical monitoring. According to the system, 360-degree continuous phase regulation and control are achieved through the mechanical programmable metasurface, the three-dimensional dynamic beam focusing technology and the improved variational mode decomposition algorithm are combined, limb movement noise is effectively restrained, and high-precision multi-person non-contact respiration monitoring is achieved. The system adopts a modular detachable metasurface architecture, supports the fast reconstruction of wavefront coding, and can reduce the average error of respiratory rate monitoring of a human body in a double-arm swinging / walking state to 0.5 times / minute without continuous power supply. Experiments show that when the system is used for synchronously monitoring breathing signals of the motion state of two persons in a complex indoor environment, the signal-to-noise is improved by 14 dB or above compared with that of a traditional scheme, and the manufacturing cost is reduced by 77.5%. The method is suitable for the fields of intelligent old-age care, infectious disease isolation monitoring, sleep breathing disorder screening and the like, and the clinical requirements of medical privacy protection and long-term health monitoring are met.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent medical monitoring, and particularly relates to a non-contact vital sign monitoring system based on a mechanically programmable metasurface, which combines three-dimensional dynamic beam focusing technology and an improved signal processing algorithm, and is applicable to multi-target respiration monitoring in a complex indoor environment. Background Art

[0002] Non-invasive continuous monitoring of vital signs is of great value in disease early warning, postoperative rehabilitation, and elderly care. Traditional contact sensors are prone to causing skin irritation and restricting patient movement. Although existing non-contact electromagnetic sensing technologies can alleviate this problem, they face two major bottlenecks: First, electrically controlled metasurfaces rely on PIN / varactor diodes to achieve beam control, and the power consumption of a single unit reaches hundreds of milliwatts, resulting in excessive energy consumption of large-scale systems and making it difficult to meet the long-term monitoring requirements; Second, the clutter generated by human limb movements and the physiological echoes of the chest are highly coupled in the frequency domain, and existing sensing systems with insufficient spatial resolution are difficult to effectively separate signals, resulting in a monitoring error of the respiration rate exceeding 4 times per minute during movement.

[0003] In recent years, wavefront manipulation technology based on programmable metasurfaces has provided new ideas for the above problems. However, its electrical control mechanism has problems such as discontinuous phase regulation and high manufacturing cost (more than 30 yuan for a single electrically controlled unit), and it cannot meet the real-time beam reconstruction requirements in dynamic scenarios. In addition, existing algorithms mostly focus on static targets and lack effective means to suppress multi-target motion interference in complex environments, seriously restricting clinical practicability. Summary of the Invention

[0004] Technical Problems

[0005] The purpose of the present invention is to break through the motion sensitivity and energy consumption limitations of traditional non-contact monitoring technologies, and propose a robust vital sign monitoring system based on a mechanically programmable metasurface (BioMeta). This system collaborates with three-dimensional dynamic beam focusing and an improved signal processing algorithm to solve key technical problems such as limb motion noise suppression, signal separation in multi-person scenarios, and low-power long-term monitoring.

[0006] Technical Solution

[0007] A non-contact vital sign monitoring system based on a mechanically programmable metasurface, comprising the following modules:

[0008] Mechanically programmable metasurface module: composed of multiple detachable mechanically rotating superatoms, each superatom supports multiple preset rotation angles, and realizes 360° continuous phase regulation through the drive of a stepper motor;

[0009] Signal transceiver module: Uses a general software radio peripheral as the signal source, paired with a pair of high-gain horn antennas, responsible for transmitting continuous wave signals (CW) and receiving reflected echoes; Optimizes the metasurface coding matrix using the Gerchberg-Saxton (GS) algorithm, dynamically adjusts the rotation angle of the meta-atoms, and focuses electromagnetic energy on the target chest area;

[0010] Mechanical control unit: Calibrates the phase-angle relationship of each meta-atom through a preset test signal, establishes a phase codebook, calculates the required phase distribution according to the target position, and drives the corresponding meta-atom to rotate to the specified angle;

[0011] Data processing module: Combines an improved variational mode decomposition algorithm to separate limb movement noise and physiological echo signals and extract the breathing frequency.

[0012] Preferably, the steps of the Gerchberg-Saxton (GS) algorithm include:

[0013] S1. Initialize the electric field distribution in the target focusing area and initialize the metasurface coding matrix;

[0014] S2. According to the current metasurface coding matrix, calculate the phase distribution of the electromagnetic wave on the metasurface, and propagate the phase distribution to the target area through Fourier transform to obtain the electric field distribution in the target area;

[0015] S3. Compare the calculated electric field distribution in the target area with the preset target field, apply constraints in the target area, retain the amplitude information of the target field, but use the calculated phase information;

[0016] S4. Propagate the corrected electric field distribution in the target area back to the metasurface through inverse Fourier transform to obtain a new metasurface phase distribution;

[0017] S5. According to the new phase distribution, update the rotation angle of the meta-atoms to generate a new metasurface coding matrix;

[0018] S6. Repeat steps S2 to S5 until the error between the electric field distribution in the target area and the preset target field reaches the preset threshold or the number of iterations reaches the upper limit;

[0019] S7. Output the final metasurface coding matrix for dynamically adjusting the rotation angle of the meta-atoms to achieve precise focusing of the electromagnetic beam.

[0020] Preferably, the steps of the improved variational mode decomposition (VMD) algorithm include:

[0021] A1. Input the received reflected echo signal and set the initial penalty factor α int and scaling factor ζ;

[0022] A2. Decompose the input signal into multiple Intrinsic Mode Functions (IMFs), each IMF representing a frequency component in the signal;

[0023] A3. For each IMF, calculate its central frequency ω i , and dynamically adjust the penalty factor α r according to the preset reference frequency ω i :

[0024]

[0025] A4. By dynamically adjusting the penalty factor, isolate the IMFs related to the breathing frequency while suppressing the high-frequency harmonic noise generated by limb movements;

[0026] A5. From the isolated IMFs, identify the IMFs related to the breathing frequency, and extract the peak points of the breathing waveform through the peak detection algorithm to calculate the breathing cycle and breathing frequency.

[0027] Preferably, the mechanically programmable metasurface module adopts Pancharatnam-Berry (PB) phase reflection units with a preset period, a circular dielectric substrate, a rectangular metal strip embedded in the center, and symmetrically distributed circular microstructures.

[0028] Preferably, the signal transceiver module is installed on a tripod at a preset height, and the data streams of the wearable sensor and the signal transceiver module are synchronized through a computer to achieve timestamp alignment.

[0029] Preferably, the mechanical control unit is only powered on during coding switching, and the motor is powered off during monitoring to save energy, and the steady-state power consumption of the system approaches zero within a preset monitoring period.

[0030] The present invention also provides a multi-person vital sign monitoring method based on the above system, including the following steps:

[0031] B1. Adopt a modular detachable metasurface architecture, realize 360° continuous phase regulation through a mechanically programmable metasurface, support rapid reconfiguration of wavefront coding, and dynamically generate a three-dimensional focused beam to lock on the chest regions of multiple target human bodies;

[0032] B2. Optimize the metasurface coding matrix through the Gerchberg-Saxton (GS) algorithm, dynamically adjust the rotation angle of the meta-atoms, and focus the electromagnetic energy on the target chest region;

[0033] B3. Collect the physiological echo signals of multiple target human bodies through the signal transceiver module, and synchronize the data streams of the wearable sensor and the signal transceiver module through a computer to achieve timestamp alignment;

[0034] B4. Combine with the improved variational mode decomposition algorithm to separate the limb movement noise and physiological echo signals of multiple target humans;

[0035] B5. Synchronously monitor the respiratory signals of multiple targets in a complex indoor environment and extract the respiratory frequency through a signal processing algorithm;

[0036] Preferably, in step B1, the electromagnetic wave is focused on the chest regions of different target humans in sequence through a time-division multiplexing strategy to achieve multi-target respiratory monitoring.

[0037] Preferably, the time-division multiplexing strategy supports a minimum angular resolution of 11°, and can distinguish multiple targets within 0.5 meters of each other.

[0038] Preferably, in step B4, the improved variational mode decomposition algorithm is dynamically adjusted through a frequency-domain penalty factor, effectively separating the limb movement harmonics and physiological signals of multiple target humans, and the accuracy of respiratory frequency extraction reaches 98.7%.

[0039] Beneficial effects

[0040] (1) Innovative motion noise suppression: Through the three-dimensional dynamic beam focusing technology of the mechanically programmable metasurface, physical layer separation of the chest echo signal and limb movement interference is achieved. The average error of respiratory frequency monitoring in the state of swinging arms / walking is reduced to 0.5 times per minute, the accuracy is increased by 10 times compared with the traditional non-focusing scheme, and the signal-to-noise ratio is increased by more than 8 dB.

[0041] (2) Zero-power sustainable monitoring: Based on the mechanical rotation phase regulation mechanism, the steady-state power consumption of an 800-unit system approaches zero within a 20-second monitoring period, saving 99% of energy compared with the electrically controlled metasurface scheme, supporting 72-hour continuous monitoring, and breaking through the energy bottleneck of long-term deployment in existing technologies.

[0042] (3) Modular multi-scenario adaptation: Adopting a detachable superatom architecture, it supports rapid attachment and deployment on the surface of furniture, has the characteristics of being detachable and reusable, and is environmentally friendly; at the same time, compared with the electrically controlled metasurface module, the manufacturing cost is reduced by 77.5%, significantly improving the system economy and environmental friendliness.

[0043] (4) Intelligent signal analysis enhancement: In a complex indoor environment, dual-target 11° angular resolution monitoring can be achieved. By integrating the improved variational mode decomposition (VMD) algorithm and the dynamic adjustment strategy of the frequency-domain penalty factor, limb movement harmonics and physiological signals are effectively separated. The accuracy of respiratory frequency extraction in the multi-person synchronous monitoring scenario reaches 98.7%, an increase of 42% compared with the traditional EMD method.

[0044] (5) Generate a time-varying coding matrix based on the Gerchberg-Saxton optimization algorithm to achieve precise millimeter-wave focusing on the thoracic region, with the signal-to-noise ratio increased by 15 dB compared to the metal reflector. Description of the Drawings

[0045] Figure 1 This is a schematic diagram of the BioMeta system architecture proposed by the present invention, showing the dynamic manipulation of electromagnetic waves in three-dimensional space by a mechanically programmable metasurface and the non-contact vital sign monitoring scenario.

[0046] Figure 2 This is the core design of the programmable BioMeta described in the present invention. (a) shows the electromagnetic wave focusing based on biometa, with a modular meta-atom mechanical structure including a rotating shaft, a dielectric substrate, and a metal resonant unit; (b) and (c) show the 19-step rotation angles (0° - 180°) of the meta-atom and the corresponding phase response curves, achieving 360° continuous regulation; (d) and (e) are examples of dynamic reconstruction of the coding matrix, showing the normalized electric field distribution at different focusing positions in the xoz and xoy planes (focusing depth 600 mm, lateral offset ±200 mm).

[0047] Figure 3 This is the simulation verification result based on BioMeta. (a) shows the beam focusing coding optimization process generated by the Gerchberg-Saxton algorithm; (b) shows the construction of a time-varying electromagnetic human model, including the parameter settings of breathing (0.2 Hz periodic conductivity change) and limb movement (1 Hz mechanical swing); (c) shows the comparison of the electromagnetic intensity at the chest with / without metasurface focusing (x-axis scanning range ±1 m, peak intensity increased by 8.7 dB); (d) shows the comparison of the time-domain waveforms of the extracted breathing signals, with the signal-to-noise ratio of the focusing scheme increased by 14.3 dB; (e) shows the comparison of the spectrograms of the human echo signals at a single moment with metasurface electromagnetic wave focusing and without metasurface beam focusing; (f) shows the comparison of the human breathing frequency estimation errors with metasurface electromagnetic wave focusing and without metasurface beam focusing over a period of time. Detailed Implementation Manner

[0048] The non-invasive human breathing monitoring system based on a modular and reconfigurable metasurface (BioMeta) proposed by the present invention suppresses limb movement noise through three-dimensional dynamic wavefront control technology to achieve high-precision and low-power multi-target vital sign perception. The meta-atoms can be connected to the surface of furniture (such as walls, desktops) through magnetic adsorption or snap-in interfaces, supporting arbitrary shape arrangement and rapid disassembly and recycling.

[0049] The present invention consists of a reconfigurable metasurface module, a signal transceiver module, a mechanical control unit, and a data processing module, and its core function is to dynamically control the electromagnetic wave beam and focus it on the target human chest area.

[0050] The BioMeta metasurface consists of 800 detachable mechanically rotatable meta-atoms. Each meta-atom is a Pancharatnam-Berry (PB) phase reflection unit with a period of 25 mm. The unit adopts a metallic square lattice structure, with a circular dielectric substrate of radius 12 mm, embedded with a rectangular metal strip (20 mm in length and 10 mm in width) and symmetrically distributed circular microstructures (radius 2.5 mm) at the center. Each meta-atom supports 19 rotation angles (0°–180°, step size 10°), and realizes 360° continuous phase regulation through a stepper motor drive. The reflection phase has a linear relationship with the rotation angle. Compared with traditional electronically controlled metasurfaces (PIN diodes or varactor diodes), BioMeta does not require continuous power supply and only consumes energy during coding switching (single energy consumption is less than 1 mW), significantly reducing the system power consumption.

[0051] The signal transceiver module uses a Universal Software Radio Peripheral (USRP-2974) as the signal source, paired with a pair of high-gain horn antennas, installed on a tripod at a height of 1.3 m, responsible for transmitting continuous wave signals (CW) and receiving reflected echoes respectively. The computer (running LabVIEW) synchronizes the wearable sensor (collecting abdominal pressure reference signals) with the USRP data stream to achieve timestamp alignment. Based on the Gerchberg-Saxton (GS) algorithm, the metasurface coding matrix is optimized, and the rotation angles of the meta-atoms are dynamically adjusted to focus the electromagnetic energy on the target chest area, filtering out environmental noises (such as WiFi and Bluetooth interferences) and high-frequency harmonics generated by limb movements.

[0052] The steps of the Gerchberg-Saxton (GS) algorithm include:

[0053] S1. Initialize the electric field distribution in the target focus area and initialize the metasurface coding matrix;

[0054] S2. According to the current metasurface coding matrix, calculate the phase distribution of the electromagnetic wave on the metasurface, and propagate the phase distribution to the target area through Fourier transform to obtain the electric field distribution in the target area;

[0055] S3. Compare the calculated electric field distribution in the target area with the preset target field, apply constraints in the target area, retain the amplitude information of the target field, but use the calculated phase information;

[0056] S4. Propagate the corrected electric field distribution in the target area back to the metasurface through inverse Fourier transform to obtain a new metasurface phase distribution;

[0057] S5. According to the new phase distribution, update the rotation angles of the meta-atoms to generate a new metasurface coding matrix;

[0058] S6. Repeat steps S2 to S5 until the error between the electric field distribution in the target area and the preset target field reaches the preset threshold or the number of iterations reaches the upper limit;

[0059] S7. Output the final metasurface coding matrix, which is used to dynamically adjust the rotation angle of the meta-atoms to achieve precise focusing of the electromagnetic beam.

[0060] The data processing module uses an improved variational mode decomposition (VMD) algorithm to process the human echo signal. In the improved VMD algorithm, a frequency-dependent penalty factor α is proposed to adaptively extract sub-signals with narrower frequency bands. Specifically, for the i-th sub-signal (i ≤ I), its penalty factor α i is denoted as

[0061]

[0062] where α int is the initial penalty factor, and ζ is the scaling factor. ω i is the frequency of the i-th IMF, and ω r is the reference frequency, which is preset according to the observation of the breathing frequency. For the adaptive α i , when the central frequency of the i-th IMF approaches its reference value ω r , a greater penalty will be given, resulting in a narrower frequency band of the signal. The entire human echo signal processing process is completed in MATLAB. In the present invention, first, the out-of-band noise in the human echo signal is filtered, and then the improved VMD algorithm is used to extract the breathing signal. Finally, peak detection is used to estimate the human breathing rate. Through three-dimensional electric field distribution simulation verification, the field strength at the chest position after focusing is increased by about 300%, and the interference in the limb area is reduced by 60%.

[0063] In the mechanical control unit, the phase-angle relationship of each meta-atom is calibrated through a preset test signal to establish a phase codebook; the required phase distribution is calculated according to the target position, and the corresponding meta-atom is driven to rotate to the specified angle; and it only needs to be powered when the coding is switched, and the motor is powered off during monitoring to save energy. The system can switch the coding matrix in a time-sharing manner, successively focusing the electromagnetic wave on the chests of different individuals, and realizing multi-target breathing monitoring through time division. It supports a minimum angular resolution of 11°, and can distinguish multiple targets within 0.5 meters adjacent to each other.

[0064] Based on the above system, the present invention also provides a multi-person vital sign monitoring method based on the system, including the following steps:

[0065] S1. Adopt a modular detachable metasurface architecture to achieve 360° continuous phase regulation through mechanically programmable metasurfaces, support rapid reconfiguration of wavefront coding, dynamically generate three-dimensional focused beams to lock on the thoracic regions of multiple target human bodies; sequentially focus electromagnetic waves on the chest regions of different target human bodies through a time-division multiplexing strategy to achieve multi-target breathing monitoring. The time-division multiplexing strategy supports a minimum angular resolution of 11°, and can distinguish multiple targets within 0.5 meters of each other.

[0066] Among them, the specific steps for the time-division multiplexing strategy to achieve multi-target breathing monitoring include:

[0067] 1. Target positioning and initialization

[0068] Step S1-1: Identify multiple human targets in the monitoring area through a laser and computer vision system, and obtain the three-dimensional coordinates (x, y, z) of their chest regions.

[0069] Step S1-2: According to the number of targets (N) and the total monitoring duration (T), divide the total time into N equal sub-time slices (Δt = T / N), and each sub-time slice corresponds to the monitoring task of one target.

[0070] 2. Dynamic switching of beam focusing

[0071] Step S2-1: In the first time slice Δt1, generate the coding matrix of the first target (Target1) through the Gerchberg-Saxton (GS) algorithm. The host computer switches the superatom codebook to the response state according to the coding matrix, so as to focus the electromagnetic beam on the chest region of Target1 (such as Figure 2 (as shown in the focusing effect in (d)).

[0072] Step S2-2: The signal transceiver module transmits a continuous wave (CW) and receives the physiological echo signal of Target1, and synchronously records the timestamp t1.

[0073] Step S2-3: Before switching to the next target, the mechanical control unit cuts off the power supply of the stepper motor, and the superatom maintains the current angle unchanged (zero-power state).

[0074] 3. Multi-target cyclic monitoring

[0075] Step S3-1: In the time slice Δt2, quickly switch to the second target (Target2), and repeat steps S2-1 to S2-3 to focus the beam on the chest region of Target2.

[0076] Step S3-2: Sequentially complete the monitoring of all targets (Target1 → Target2 →... → TargetN) in a cycle, and each target is scanned at least once within a single cycle (such as Figure 3(c) Schematic diagram of time division).

[0077] 4. Signal Synchronization and Processing

[0078] Step S4-1: Divide the echo signals of multiple people through the timestamps recorded in Step S2-2, so as to obtain the original echo signal data of each person;

[0079] Step S4-2: Process the signals of each target using the improved VMD algorithm, extract the respiratory echo signals of each person, and estimate the breathing frequency of this person according to the number of peaks in each segment of the signal (usually 1 signal peak corresponds to 1 breath of this person);

[0080] S2. Optimize the metasurface coding matrix through the Gerchberg-Saxton (GS) algorithm, dynamically adjust the rotation angle of the meta-atoms, and focus the electromagnetic energy on the target chest area;

[0081] S3. Collect the physiological echo signals of multiple target human bodies through the signal transceiver module, and divide the multi-person echo signals by timestamp alignment;

[0082] S4. Combine the improved variational mode decomposition algorithm to separate the limb movement noise and physiological echo signals of multiple target human bodies, and the accuracy rate of breathing frequency extraction reaches 98.7%.

[0083] S5. Synchronously monitor the breathing signals of multiple targets in a complex indoor environment, and extract the breathing frequency through a signal processing algorithm.

Claims

1. A non-contact vital signs monitoring system based on a mechanically programmable metasurface, characterized in that: Includes the following modules: Mechanically programmable metasurface module: It is composed of multiple detachable mechanically rotating meta-atoms. Each meta-atom supports multiple preset rotation angles and can achieve 360° continuous phase control through stepper motor drive. Signal transceiver module: It uses a general software radio peripheral as the signal source and is equipped with a pair of high-gain horn antennas to transmit continuous wave signals (CW) and receive reflected echoes. It uses the Gerchberg-Saxton (GS) algorithm to optimize the metasurface coding matrix, dynamically adjust the metaatom rotation angle, and focus the electromagnetic energy on the target chest area. Mechanical control unit: calibrates the phase-angle relationship of each superatom through preset test signals, establishes a phase codebook, calculates the required phase distribution according to the target position, and drives the corresponding superatom to rotate to the specified angle; Data processing module: Combined with the improved variational mode decomposition algorithm, it separates limb movement noise from physiological echo signals and extracts respiratory frequency.

2. The non-contact vital signs monitoring system according to claim 1, characterized in that: The steps of the Gerchberg-Saxton (GS) algorithm include: S1, initializing the electric field distribution of the target focus area and initializing the metasurface encoding matrix; S2. Calculate the phase distribution of the electromagnetic wave on the metasurface according to the current metasurface coding matrix, and propagate the phase distribution to the target area through Fourier transform to obtain the electric field distribution of the target area; S3, comparing the calculated electric field distribution of the target area with the preset target field, imposing constraints on the target area, retaining the amplitude information of the target field, but using the calculated phase information; S4, propagating the corrected electric field distribution of the target area back to the metasurface through inverse Fourier transform to obtain a new metasurface phase distribution; S5. According to the new phase distribution, the rotation angle of the meta-atom is updated to generate a new metasurface encoding matrix; S6, repeating steps S2 to S5 until the error between the electric field distribution in the target area and the preset target field reaches a preset threshold, or the number of iterations reaches an upper limit; S7. Output the final metasurface coding matrix, which is used to dynamically adjust the rotation angle of the metaatom and achieve precise focusing of the electromagnetic beam.

3. The non-contact vital signs monitoring system according to claim 1, characterized in that: The steps of the modified variational mode decomposition (VMD) algorithm include: A1. Input the received reflected echo signal and set the initial penalty factor α int and the scaling factor ζ; A2. Decompose the input signal into multiple intrinsic mode functions (IMFs), each IMF represents a frequency component in the signal; A3. For each IMF, calculate its center frequency ω i , and according to the preset reference frequency ω r , dynamically adjust the penalty factor α i : A4, by dynamically adjusting the penalty factor, the IMF related to the respiratory frequency is separated, while the high-frequency harmonic noise generated by limb movement is suppressed; A5. From the separated IMFs, identify the IMFs related to the respiratory frequency, extract the peak points of the respiratory waveform through the peak detection algorithm, and calculate the respiratory cycle and respiratory frequency.

4. The non-contact vital signs monitoring system according to claim 1, characterized in that: The mechanical programmable metasurface module adopts a Pancharatnam-Berry (PB) phase reflection unit with a preset period and a circular medium as the substrate, with a rectangular metal strip and symmetrically distributed circular microstructures embedded in the center.

5. The non-contact vital signs monitoring system according to claim 1, characterized in that: The signal transceiver module is installed on a tripod at a preset height, and the wearable sensor and the signal transceiver module data stream are synchronized by a computer to achieve timestamp alignment.

6. The non-contact vital signs monitoring system according to claim 1, characterized in that: The mechanical control unit is powered only when the encoding is switched, and the motor is powered off during the monitoring period to save energy. The steady-state power consumption of the system approaches zero within a preset monitoring period.

7. A method for monitoring vital signs of multiple persons based on the system according to any one of claims 1 to 6, characterized in that: The following steps are involved: B1. Adopting modular detachable metasurface architecture, 360° continuous phase control is achieved through mechanical programmable metasurface, supporting rapid reconstruction of wavefront coding, and dynamically generating three-dimensional focused beams to lock multiple target human chest areas; B2. Optimize the metasurface encoding matrix through the Gerchberg-Saxton (GS) algorithm, dynamically adjust the metaatom rotation angle, and focus the electromagnetic energy on the target chest area; B3. Collect physiological echo signals of multiple target human bodies through the signal transceiver module, and synchronize the data streams of the wearable sensor and the signal transceiver module through a computer to achieve time stamp alignment; B4. Combine the improved variational mode decomposition algorithm to separate the limb motion noise and physiological echo signals of multiple target human bodies; B5. Synchronously monitor multiple target breathing signals in a complex indoor environment and extract the breathing frequency through signal processing algorithms.

8. The method for monitoring multiple vital signs according to claim 7, characterized in that: In the step B1, electromagnetic waves are sequentially focused on the chest areas of different target human bodies through a time multiplexing strategy to achieve multi-target respiratory monitoring.

9. The method for monitoring multiple vital signs according to claim 8, characterized in that: The time multiplexing strategy supports a minimum angular resolution of 11° and can distinguish multiple targets within an adjacent 0.5 meter.

10. The method for monitoring multiple vital signs according to claim 7, characterized in that: In step B4, the improved variational mode decomposition algorithm dynamically adjusts the frequency domain penalty factor to effectively separate the limb motion harmonics and physiological signals of multiple target human bodies, and the accuracy of respiratory frequency extraction reaches 98.7%.

Images