Multi-person heartbeat and breathing information monitoring method and system

Abstract

The invention discloses a multi-person heartbeat and breathing information monitoring method and system. The method comprises the following steps: a radar antenna adopts a one-transmitting and N-receiving mode to transmit a wide wave beam; receiving echo signals by the N receiving antennas, performing multi-target identification by using a digital beam forming algorithm to obtain an initial azimuth angle of each target, and correcting the initial azimuth angle of each target by using a sum-difference angle measurement algorithm to obtain a final azimuth angle of each target; mixing the echo signal of each target azimuth angle with the local oscillator signal to obtain an intermediate frequency signal; taking the intermediate frequency signal as the input of discrete wavelet transform, obtaining the frequency spectrum of each target, and locking the distance unit information of each target; meanwhile, a discrete wavelet transform coefficient is dynamically adjusted in combination with a self-adaptive filtering algorithm based on a minimum mean square error so as to adapt to the change of the signal; and performing signal processing on the distance unit information of each target to obtain multi-target vital sign information. According to the invention, multi-person heartbeat and breathing information monitoring can be realized, and the real-time performance and accuracy of the monitoring system can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of millimeter wave radar signal processing, and in particular to a method and system for monitoring the heartbeat and breathing information of multiple people. Background Art

[0002] With the continuous progress of society and people's increasing attention to personal health, the monitoring of human vital signs has become an important issue. In particular, the monitoring of heartbeat and respiratory signals is directly related to multiple health indicators, so fast and accurate measurement technology has received widespread attention. At present, most common vital sign monitoring devices, such as blood pressure monitors, smart bracelets, finger-clip pulse oximeters and electrocardiograms, use contact measurement methods. The advantages of this type of method are high measurement accuracy and reliable data, so it is very suitable for use in environments such as homes and hospitals. However, contact measurement also has some limitations. For example, for patients with damaged skin such as burns, it is difficult for contact sensors to fit directly on the skin, making measurement difficult. In addition, for special groups that require long-term monitoring, such as the elderly and infants, long-term contact monitoring may interfere with their normal sleep and daily life. In this context, non-contact vital sign monitoring technology is particularly important.

[0003] This type of technology can monitor indicators such as heartbeat and breathing through telemetry, thus avoiding the need for direct contact with the skin, reducing discomfort to the patient, and also providing solutions for special situations where contact monitoring methods cannot be used. However, when performing multi-target physiological information monitoring, the identification of different targets is a difficult problem to solve. So far, most studies have focused on a single target, but in actual application scenarios, it is necessary to monitor the vital signs of two or more targets, such as real-time monitoring of physiological information of multiple beds.

[0004] The difficulty of multi-target respiration and heartbeat monitoring lies in the differentiation of multiple targets. The mainstream method of multi-target differentiation is to differentiate based on the different radial distances of the targets. However, its disadvantages are also very obvious. In order to obtain a higher distance resolution, this method requires a high bandwidth of the radar system, and when multiple targets are located in the same distance unit, they cannot be differentiated. At the same time, since the radar has different echo intensities for targets at different distances, this also indirectly leads to a reduction in measurement accuracy.

[0005] In the prior art, the invention patent with the patent publication number CN115399747A discloses a multi-person vital signs monitoring method based on millimeter wave radar. It effectively separates multiple targets from the distance-angle diagram and realizes target positioning based on the connected domain analysis method, and extracts the phase signal at each target position to estimate the breathing and heart rate. This patent can distinguish multiple targets at the same distance or the same angle, and realize vital signs estimation in the same distance scene. However, this patent first processes the intermediate frequency signal, and then uses the MVDR beamforming algorithm to estimate the angle, processes the rangefft points of each azimuth and distance, and finally determines the target. This processing method has several disadvantages: First, in step 1 of the patent, a set of echo data is first obtained to obtain the radar three-dimensional data matrix Y∈C M×N×L , where M is the number of sampling frames, N is the number of sampling points per frame, and L is the number of channels. The need to accumulate M frames of data and then perform subsequent processing will affect the real-time monitoring. Second, in step 2 of the patent, in the radar three-dimensional data matrix Y, each row of each channel data performs FFT along the fast time dimension, and processing all data of L channels requires a large amount of calculations. Third, in steps 3 and 4 of the patent, the distance-angle diagram is obtained according to the MVDR angle spectrum estimation, and then the distance-angle diagram is preprocessed to filter out image data greater than the preset detection threshold. The above algorithm is a global search algorithm, which occupies a large amount of memory and has higher requirements for hardware resources. Summary of the invention

[0006] The technical problem to be solved by the present invention is to realize the monitoring of heartbeat and breathing information of multiple persons while ensuring the real-time performance and accuracy of the monitoring system.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] A method for monitoring heartbeat and breathing information of multiple persons, comprising:

[0009] The radar antenna adopts 1 transmit and N receive, emitting a wide beam;

[0010] The N receiving antennas receive the echo signals, and use the digital beamforming algorithm to perform multi-target recognition, obtain the initial azimuth of each target, and then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target;

[0011] Mix the echo signal of each target azimuth with the local oscillator signal to obtain an intermediate frequency signal;

[0012] The intermediate frequency signal is used as the input of discrete wavelet transform to obtain the spectrum of each target and lock the distance unit information of each target; at the same time, the discrete wavelet transform coefficients are dynamically adjusted to adapt to the changes of the signal by combining the adaptive filtering algorithm based on minimum mean square error;

[0013] Signal processing is performed on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain the vital signs information of multiple targets.

[0014] In one embodiment of the present invention, obtaining the initial azimuth of each target includes:

[0015] Get the radar antenna transmission signal and the echo signal received after a period of time:

[0016] Acquire a receiving signal of each receiving antenna according to an array element of each receiving antenna, a spacing between array elements, and an incident angle of an incident signal;

[0017] Convert the received signals of N receiving antennas into vector form;

[0018] The digital beamforming algorithm is used to weight the received signal in vector form to compensate for the phase difference caused by the time delay between the array elements, and to achieve the same-direction superposition of the N receiving antennas in a certain direction during the output process, so that the output power at the incident angle is maximized. The spatial power spectrum is constructed to complete the angle estimation of the target azimuth and obtain the initial azimuth of each target.

[0019] In one embodiment of the present invention, the maximum output power of digital beamforming is obtained by the following formula:

[0020] P DBF (w) = a H (θ k )Ra(θ k );

[0021] Where P DBF (w) represents the maximum output power of digital beamforming, w represents the weight vector of N receiving antennas, and a(θ k ) is represented as the steering vector, e is a natural constant, j is an imaginary unit, d L is the spacing between array elements, λ is the signal wavelength, T is the transpose, θ k is represented as the incident angle of the kth incident signal, R is represented as the covariance matrix of the output signals of the N receiving antennas, and H is represented as the conjugate transpose;

[0022] The spatial power spectrum is obtained by changing the weight vector to perform spatial beam scanning, and a detection threshold is preset. The beam direction greater than the detection threshold is the initial azimuth of the target.

[0023] In one embodiment of the present invention, obtaining the final azimuth angles of each target includes:

[0024] According to the initial azimuth angle of the target, beam 1 is formed in this angle direction, and beam 2 is formed in the adjacent angle direction. The two beams have the same shape and partially overlap.

[0025] The line connecting the intersection of beam 1 and beam 2 and the origin is taken as the equal signal axis, and the direction of the equal signal axis is recorded as

[0026] Perform sum and difference processing on the signals simultaneously received by beam 1 and beam 2 to obtain a sum signal and a difference signal respectively;

[0027] In the direction of the equal signal axis Nearby, the sum signal and the difference signal are approximately represented by using the first-order Taylor expansion to obtain the difference and ratio;

[0028] Use the difference and ratio to judge the angle where the target deviates from the direction of the equal signal axis The size and direction of the angle, and then the final azimuth of each target is obtained;

[0029] Among them, the difference and ratio are calculated by the following formula:

[0030]

[0031] In the formula, u Δ (θ) is expressed as a difference signal, u ∑ (θ) is expressed as a sum signal, θ re It is expressed as the angle of the target deviating from the direction of the equal signal axis. angle, F(θ) is the antenna directivity function, and S is the intermediate variable;

[0032] The final azimuth of each target is angle θ re and sum.

[0033] In one embodiment of the present invention, the intermediate frequency signal S IF (x) is obtained by the following formula:

[0034]

[0035] In the formula, A IF It is represented by the amplitude of the intermediate frequency signal, B is represented by the bandwidth of the signal, and T c Expressed as the signal period, f c It is represented as the starting frequency of the signal, τ is represented as the delay between transmitting the signal and receiving the echo signal, It is represented as phase difference and t is represented as time.

[0036] In one embodiment of the present invention, a discrete wavelet transform is performed on the intermediate frequency signal, using the following formula:

[0037]

[0038] Where ψ is the mother wavelet function, J, L∈Z, a0, b0 are constants, and Z is an integer;

[0039] Discrete wavelet transform coefficient DWT(a0 J ,Lb0) is expressed as:

[0040]

[0041] Where y(t) represents the input signal of discrete wavelet transform.

[0042] In one embodiment of the present invention, the minimum mean square error algorithm minimizes the mean square value of the output error sequence e(n) by adjusting the filter coefficients, and modifies the filter coefficients according to this criterion, and approximates the desired signal by iteratively updating the filter coefficients;

[0043] Among them, the input signal is DWT(a0 J ,Lb0), the output signal y(n) of the filter at time n can be expressed as:

[0044]

[0045] Where M is the order of the filter, w i (n) is the filter coefficient at time n;

[0046] The output error sequence e(n) is obtained by:

[0047] e(n)=D(n)-y(n);

[0048] Where D(n) represents the expected output signal;

[0049] The desired signal is approximated by iteratively updating the filter coefficients, using the following formula:

[0050] w i (n+1)=w i (n)+2μ·e(n)·DWT(a0 J ,Lb0);

[0051] In the formula, μ represents the step size factor.

[0052] In one embodiment of the present invention, the signal processing includes: performing phase unwrapping,

[0053] Assume the current phase value is φ n , the phase value at the next moment is φ n+1 ;

[0054] When φ n+1 -φn ≥π, it indicates that the phase signal has jumped, and execute φ n+1 =φ n -2π;

[0055] When φ n+1 -φ n <π, execute φ n+1 =φ n +2π, to ensure that the phase signal jump is less than π;

[0056] The first-order phase difference is used to remove the constant phase offset. The adjacent phase values ​​are subtracted to remove the constant offset contained in the adjacent frames. Assume that the current frame phase is φ n , the phase of the previous frame is φ n-1 , the difference between the two φ n -φ n-1 As the current phase value φ n .

[0057] In one embodiment of the present invention, after phase information extraction, an IIR bandpass filter is used to separate the respiratory and heartbeat signals, and then a spectrum diagram of the signal is obtained by spectrum estimation and then a peak search is performed to obtain the respiratory frequency and heartbeat frequency.

[0058] The present invention also provides a multi-person heartbeat and respiration information monitoring system, which uses the multi-person heartbeat and respiration information monitoring method described in any one of claims 1 to 8, comprising:

[0059] Radar antenna module, used for radar antenna adopting 1 transmit and N receive, emitting wide beam;

[0060] The target azimuth module is used to receive the echo signals from N receiving antennas, and use the digital beamforming algorithm to perform multi-target recognition to obtain the initial azimuth of each target; then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target;

[0061] The intermediate frequency signal module is used to mix the echo signal of each target azimuth with the local oscillator signal to obtain the intermediate frequency signal;

[0062] The target distance module is used to use the intermediate frequency signal as the input of the discrete wavelet transform, obtain the spectrum of each target, and lock the distance unit information of each target; at the same time, the discrete wavelet transform coefficients are dynamically adjusted to adapt to the changes of the signal in combination with the adaptive filtering algorithm based on the minimum mean square error;

[0063] The multi-target vital signs module is used to perform signal processing on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain multi-target vital signs information.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] The present invention first uses a digital beamforming algorithm to obtain the initial angle direction of each target within the angle coverage range of the millimeter wave radar, and then uses the sum-difference angle measurement algorithm to obtain the initial angle deviation value, thereby obtaining a more accurate azimuth, thereby improving the accuracy of angle estimation, and then uses discrete wavelet transform to obtain the distance unit of each target with a determined angle direction. Compared with the algorithm that first determines the distance unit and then estimates the angle, the amount of algorithm calculation can be reduced, and the hardware memory usage can be reduced. Reducing the amount of algorithm calculation can reduce the computing resources required by the processor, thereby reducing the power consumption of the hardware device and reducing the generation of heat, which is beneficial for the use of portable devices or long-term monitoring systems. In addition, the reduction in the amount of calculation means an increase in data processing speed, ensuring the real-time nature of multi-target vital signs monitoring and achieving efficient multi-target monitoring.

[0066] The present invention adopts a 1-transmit-N-receive radar, firstly uses a digital beamforming algorithm and a sum-difference angle measurement algorithm to obtain the angle information of each target; then locks the distance unit of each target through discrete wavelet transform, and dynamically adjusts the discrete wavelet transform coefficient to adapt to the change of the signal in combination with an adaptive filtering algorithm based on minimum mean square error; then performs phase extraction and processing, separates the heartbeat and breathing signals through a bandpass filter, and obtains the vital sign data of each target through spectrum estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 The present invention is a flowchart of a method for monitoring heartbeat and respiration information of multiple people according to an embodiment of the present invention.

[0068] Figure 2 FIG. 4 is a schematic diagram of digital beamforming according to an embodiment of the present invention.

[0069] 3(a) and (b) are schematic diagrams of digital beamforming target positioning according to an embodiment of the present invention.

[0070] 4(a), (b), and (c) are schematic diagrams of sum and difference angle measurement according to an embodiment of the present invention.

[0071] Figure 5 Schematic diagram of target monitoring results according to an embodiment of the present invention.

[0072] Figure 6 The present invention is a block diagram of a multi-person heartbeat and respiratory information monitoring system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0073] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described in conjunction with the accompanying drawings of the specification.

[0074] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0075] See also Figure 1 This embodiment discloses a method for monitoring heartbeat and breathing information of multiple persons, including:

[0076] S10, the radar antenna adopts 1 transmit and N receive, emitting a wide beam.

[0077] In one embodiment of the present invention, a millimeter wave radar with 1 transmitting antenna and N receiving antennas transmits a wide beam, with the front of the radar as an angle of 0°, and covers an angle range of -90° to 90° in front of the radar, which is specifically determined by the transmitting antenna, and transmits a linear frequency modulated continuous wave signal s through the millimeter wave radar. T (t):

[0078]

[0079] In the formula, A T is the amplitude of the transmitted signal, f c is the starting frequency of the signal, B is the bandwidth of the signal, T c is the signal period, is the phase, c is the speed of light, and t is the time. In addition, in this embodiment, the angle range in front of the radar is [-60°, 60°].

[0080] S20, the N receiving antennas receive the echo signals, and use the digital beamforming algorithm to perform multi-target recognition, obtain the initial azimuth of each target, and then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target.

[0081] In one embodiment of the present invention, the echo signal received after a time τ is s R (t):

[0082]

[0083] In the formula, A R is the amplitude of the received signal, and τ is the delay from transmitting the signal to receiving the echo signal.

[0084] In this embodiment, when monitoring the vital signs of multiple targets, the radar can collect signals from different directions. The phase information directly extracted may be the aliasing of multiple target phases or the phase information of a target with stronger energy. The role of beamforming is to point the beam to the direction of interest, so that the information required for the direction can be obtained more accurately.

[0085] Digital beamforming is a signal processing technology that uses a sensor array to receive signals in a directional manner. The principle is to process the data of the antenna array, such as weighting, adding some weight constraints, etc., to form a certain beam shape, so that the desired angle signal interferes constructively and other angle signals interfere destructively. The role of the beam is to effectively enhance the target signal and suppress other non-target signals to a certain extent. This technology can not only effectively improve the signal-to-noise ratio, but also has a certain noise reduction effect.

[0086] N receiving antennas form a receiving antenna array. Correspondingly, the receiving antenna array has N array elements, using a uniform linear array, and the spacing between the array elements is d L , the incident angle of the incident signal is θ k (k=1,2,...,K), take the first array element as the reference array element, assume that the signal is uncorrelated with the noise, then the received signal x of the nth array element n (t) is expressed as:

[0087]

[0088] Where s(t) is the incident angle θ k The array receives the signal, λ is the signal wavelength, e is a natural constant, j is an imaginary unit, θ k is the incident angle of the kth incident signal.

[0089] Write the received signal of the receiving antenna array in the form of a vector:

[0090] X(t) = A(θ)S(t);

[0091] where X(t)=[x1(t),x2(t),...,x N (t)] T is the received signal vector of N array elements, A(θ)=[a(θ1),a(θ2),...,a(θ K )] is a N×K-dimensional steering vector, where T is the matrix transpose, S(t)=[s1(t),s2(t),...,s K (t)] T is the incident signal vector.

[0092] The weighted vectors of antenna array elements are used to compensate for the phase difference caused by time delay between array elements, and to achieve unidirectional superposition in a certain direction during the array output process, so as to maximize the output power in this direction, and to complete the angle estimation of the target direction by constructing the spatial power spectrum.

[0093] After weighting the signals received by N array elements, the output signal Y(t) is:

[0094] Y(t)=w H X(t);

[0095] Among them, the weight vector of N array elements is expressed as w = [w1, w2, ..., w N ] T , w H represents the conjugate transpose of w.

[0096] The output power p(w) of the receiving antenna array is expressed as:

[0097] p(w)=E[|Y(t))| 2 ]=w H E[X(t)X H (t)]w=w H Rw;

[0098] Where R = E[X(t)X H (t)] is the covariance matrix of the array output signal, and E[·] represents the expectation.

[0099] The digital beamforming algorithm optimizes the weight vector to maximize the output power in a specific direction. k The beam forming of k ), then the maximum output power of digital beamforming is P DBF (w) is:

[0100] P DBF (w) = a H (θ k )Ra(θ k )

[0101] The spatial power spectrum is obtained by changing the weight vector to perform spatial beam scanning, and a detection threshold is preset. The beam direction greater than the detection threshold is the initial azimuth of the target.

[0102] See also Figure 2 As shown, the number of array elements in the antenna array is N = 4, a uniform linear array is used, and the spacing between the array elements is d L=λ / 2, λ is the signal wavelength. The weight vector is changed within [-60°, 60°] to perform spatial beam scanning to obtain the spatial power spectrum. The preset detection threshold is 2. The beam direction with an amplitude greater than the detection threshold in the spatial power spectrum is the initial azimuth of the target.

[0103] Please refer to FIG. 3 , where (a) and (b) correspond to the target positioning schematic diagrams of human target 1 and human target 2 respectively, and the initial azimuth angles of human targets 1 and 2 are -20° and 35° respectively.

[0104] Due to the limited number of radar array elements, there will be a large error in determining the azimuth of the target using digital beamforming. The sum-difference angle measurement algorithm can more accurately measure the angle of the target. In the sum-difference angle measurement algorithm, the sum beam has the largest gain in the pointing direction, and the difference beam forms a zero point in the pointing direction. The sum-difference ratio is used to obtain the deviation angle of the target relative to the axial direction.

[0105] After the approximate azimuth of the target is determined by the digital beamforming algorithm, beam 1 is formed in this angle direction, and beam 2 is formed in the adjacent angle direction. The two beams have the same shape and partially overlap. The signals received by the two beams are processed by sum and difference to obtain the sum signal and difference signal and the corresponding angle detection curve. The angle detection curve is the curve of the amplitude ratio of the sum and difference signals changing with the error angle. The intersection of the two beams and the line connecting the origin is the equal signal axis, and the pointing angle is If the target echo signal is incident from the direction of the equal signal axis, the target angle error is 0, and the difference signal of the two beams is 0. If the target deviates from the equal signal axis, the difference signal output amplitude is proportional to the error angle, and its sign is determined by the direction of deviation. More accurate target angle information can be obtained based on the angle at which the target echo signal deviates from the equal signal axis.

[0106] The directivity functions F1(θ) and F2(θ) of beam 1 and beam 2 are:

[0107]

[0108] Among them, θ D is the deviation angle between the equal signal axis and the direction of the beam maximum value, and F(θ) is the antenna directivity function.

[0109] The signals received by beam 1 and beam 2 are:

[0110] u1(θ)=KF1(θ)=KF(θ D -θ re )

[0111] u2(θ)=KF2(θ)=KF(θ D +θ re )

[0112] Among them, θ reThe angle at which the target deviates from the direction of the equal signal axis The angle, K is the signal amplitude coefficient, K is a constant. For the sum-difference method of angle measurement, the sum signal u ∑ (θ) and the difference signal u Δ (θ) is calculated as follows:

[0113] u ∑ (θ)=u1(θ)+u2(θ)

[0114] u Δ (θ)=u1(θ)-u2(θ)

[0115] In the direction of the equal signal axis The first-order Taylor expansion is used to approximate the above formula:

[0116]

[0117] The difference and ratio are:

[0118]

[0119] in,

[0120] It can be seen that the difference and ratio Proportional to the target deflection angle θ re , so the difference and ratio can be used to judge the angle θ re The size and direction of each target. Specifically, the final azimuth of each target is angle θ re and sum.

[0121] Please refer to Figure 4(a), the sum and difference angle measurement is performed on target 1, the azimuth of beam 1 is -20°, and the azimuth of beam 2 is -15°. Figure 4(b) is the sum signal and difference signal of beam 1 and beam 2, and Figure 4(c) is the angle detection curve of the difference and sum ratio. It can be seen from the figure that the ratio of the difference signal to the sum signal is approximately a slanted line, and the difference and sum ratio corresponds to the angle value one by one. The sum and difference angle measurement process of other targets is the same as that of target 1, and the more accurate azimuths of target 1 and target 2 are -19.1° and 34.2°.

[0122] S30, mixing the echo signal of each target azimuth with the local oscillator signal to obtain an intermediate frequency signal.

[0123] In one embodiment of the present invention, after obtaining the final azimuth angles of each target, the echo beams of the final azimuth angles of each target are mixed with the local oscillator signal to obtain an intermediate frequency signal beam. When the radar receives echo beams of multiple targets, after mixing by the mixer, intermediate frequency signals with different frequencies are obtained. The intermediate frequency beam is S IF (x):

[0124]

[0125] In the formula, A IF Expressed as the amplitude of the intermediate frequency signal, Expressed as phase difference.

[0126] S40, taking the intermediate frequency signal as the input of discrete wavelet transform, obtaining the spectrum of each target, locking the distance unit information of each target; at the same time, combining the adaptive filtering algorithm based on minimum mean square error to dynamically adjust the discrete wavelet transform coefficients to adapt to the change of the signal.

[0127] In one embodiment of the present invention, discrete wavelet transform is used to lock the distance unit of each target, and the discrete wavelet transform coefficients are dynamically adjusted to adapt to the change of the signal in combination with the adaptive filtering algorithm based on the minimum mean square error (LMS). Discrete wavelet transform is a time-frequency analysis method that can decompose the signal into frequency components of different scales. The input signal y(t) is subjected to discrete wavelet transform, and the expression is:

[0128]

[0129] In the formula, ψ is the mother wavelet function, a is expressed as the scale factor, and the power series is discretized as a=a0J; b is expressed as the displacement factor, which is uniformly valued, b=La0Jb0; where J, L∈Z, a0, b0 are constants, and Z is an integer. The discretization here is for the scale factor a and the displacement factor b, rather than the discretization of the time variable t. The selection of b0 depends on the mother wavelet function ψ.

[0130] Discrete wavelet transform coefficient DWT(a0 J ,Lb0) is expressed as:

[0131]

[0132] Where y(t) represents the input signal of discrete wavelet transform.

[0133] The signal is filtered using the minimum mean square error algorithm. The goal of the minimum mean square error algorithm is to minimize the mean square value of the output error sequence e(n) by adjusting the filter coefficients, and to modify the filter coefficients according to this criterion. The filter coefficients are updated iteratively to approximate the desired signal. The input signal is DWT(a0 J ,Lb0), the output signal y(n) of the filter at time n can be expressed as:

[0134]

[0135] Where M is the order of the filter, w i (n) is the filter coefficient at time n.

[0136] The output error sequence e(n) is the difference between the desired output signal D(n) and the actual output signal y(n):

[0137] e(n)=D(n)-y(n);

[0138] The filter coefficients are updated at each iteration step according to the error signal to minimize the mean square error between the output signal and the desired signal:

[0139] w i (n+1)=w i (n)+2μ·e(n)·DWT(a0 J ,Lb0);

[0140] Where μ is the step size factor, which is used to control the update amplitude. The condition for the minimum mean square error algorithm to converge is: 0<μ<1 / λ max ,λ max is the largest eigenvalue of the autocorrelation matrix of the input signal.

[0141] In this embodiment, specifically, the mother wavelet function selects the Haar wavelet function, which is the only real-valued wavelet that has compact support, symmetry, and orthogonality. It is defined as:

[0142]

[0143] The multi-layer decomposition of wavelet transform can separate the signal into wavelet coefficients of different scales, which can reveal the mutation information hidden in the normal signal.

[0144] The minimum mean square error algorithm is an easy-to-implement, robust, and widely used algorithm. Selecting a suitable expected signal D(n), the filter coefficient adjustment algorithm tries to make y(n) close to D(n), the expected value of the difference e(n) between the expected signal D(n) and the filter output y(n) is minimized, and the filter coefficient w is modified according to this criterion. i (n).

[0145] S50, performing signal processing on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain multi-target vital sign information.

[0146] In one embodiment of the present invention, the target distance unit locked after the final target azimuth angles and discrete wavelet transform obtained by sum and difference angle measurement algorithm are corrected cannot determine whether it is a human target, and it is necessary to perform operations such as unwrapping and first-order difference on its phase signal to determine which is the human target to be measured. Because the phase difference generated by the human respiratory and heartbeat signals is time-varying, while the phase signals of other static objects remain basically unchanged. The phase information of the human target is extracted, and the phase extraction utilizes the inverse tangent function principle. The phase value amplitude range obtained by the inverse tangent function is [-π,π]. Therefore, the real phase value generated by the respiratory and heartbeat signals is folded in the range of [-π,π]. To obtain the real phase value, it is necessary to perform phase unwrapping on the extracted phase value to unwrap the phase data from the range of [-π,π] to make it continuous phase data.

[0147] In this embodiment, the specific principle is: suppose the current phase value is φ n , the phase value at the next moment is φ n+1 When φ n+1 -φ n ≥π, it means that the phase signal has jumped, execute φ n+1 =φ n -2π; when φ n+1 -φ n <π, execute φ n+1 =φ n +2π, to ensure that the phase signal jump is less than π. There is a constant phase offset in the unwrapped phase signal. Since the constant phase offset of each frame of data is the same, it is necessary to use the first-order phase difference to remove the constant phase offset. The adjacent phase values ​​are subtracted, that is, the phase value of the next frame minus the phase value of the previous frame. The constant offset contained in the adjacent frames is removed by subtracting. Assume that the current frame phase is φ n , the phase of the previous frame is φ n-1 , the difference between the two φ n -φ n-1 As the current phase value φ n .

[0148] In this embodiment, after the phase information is extracted, the breathing and heartbeat signals are separated using an IIR bandpass filter, and then the spectrum diagram of the signal is obtained by spectrum estimation and then a peak search is performed to obtain the breathing frequency and the heartbeat frequency.

[0149] In one embodiment of the present invention, the phase information of two target distance units in all frame data is extracted and processed, mainly including phase unwrapping and phase first-order difference operations. The essential difference between a stationary object and a human target is that the phase difference of a human body changes over time due to the effects of breathing and heartbeat, while the phase difference of a stationary object is close to zero due to the lack of relative motion. Using this difference, the method of calculating the variance can be used to determine whether the target is a human target. The mean square error of the first-order difference of the two target phases is compared to determine whether a human target exists. If so, the human target is screened out. The signal is then divided into two paths for filtering. The frequency range of a normal human breathing signal is 0.1-0.5Hz, and the frequency range of a heartbeat signal is 0.8-2Hz. This embodiment uses an IIR bandpass filter for filtering to separate the breathing signal and the heartbeat signal. The Fourier transform-based spectrum estimation is performed on the breathing signal and the heartbeat signal, respectively, to obtain a spectrum diagram for peak search. The frequency corresponding to the maximum peak is the breathing frequency and the heartbeat frequency. Figure 5 As shown in the figure, the target detection results are that the azimuths of target one and target two are -19.1° and 34.2° respectively. The distances of the targets locked by discrete wavelet transform and minimum mean square error algorithm are 0.6m and 0.8m respectively. Finally, the breathing and heart rates of the two targets are obtained by signal processing. The results show that the breathing rate of target one is 16 times per minute and the heart rate is 76 times per minute. The breathing rate of target two is 14 times per minute and the heart rate is 82 times per minute.

[0150] See also Figure 6 As shown, the present invention also provides a multi-person heartbeat and breathing information monitoring system, which uses the multi-person heartbeat and breathing information monitoring method described above, including:

[0151] Radar antenna module, used for radar antenna adopts 1 transmit and N receive, emitting wide beam.

[0152] The target azimuth module is used to receive the echo signals from N receiving antennas, and use the digital beamforming algorithm to perform multi-target recognition, obtain the initial azimuth of each target, and then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target.

[0153] The intermediate frequency signal module is used to mix the echo signal of each target azimuth with the local oscillator signal to obtain the intermediate frequency signal.

[0154] The target distance module is used to use the intermediate frequency signal as the input of the discrete wavelet transform, obtain the spectrum of each target, and lock the distance unit information of each target; at the same time, the discrete wavelet transform coefficients are dynamically adjusted in combination with the adaptive filtering algorithm based on the minimum mean square error to adapt to the changes in the signal.

[0155] The multi-target vital signs module is used to perform signal processing on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain multi-target vital signs information.

[0156] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0157] The above-described embodiments merely represent implementation methods of the invention. The protection scope of the present invention is not limited to the above-described embodiments. For those skilled in the art, several modifications and improvements may be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for monitoring heartbeat and breathing information of multiple people, characterized in that: include: The radar antenna adopts 1 transmit and N receive, emitting a wide beam; The N receiving antennas receive the echo signals, and use the digital beamforming algorithm to perform multi-target recognition, obtain the initial azimuth of each target, and then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target; Mix the echo signal of each target azimuth with the local oscillator signal to obtain an intermediate frequency signal; The intermediate frequency signal is used as the input of discrete wavelet transform to obtain the spectrum of each target and lock the distance unit information of each target; at the same time, the discrete wavelet transform coefficients are dynamically adjusted to adapt to the changes of the signal by combining the adaptive filtering algorithm based on minimum mean square error; Signal processing is performed on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain the vital signs information of multiple targets.

2. The method for monitoring heartbeat and breathing information of multiple persons according to claim 1, characterized in that: Get the initial azimuth of each target, including: Get the radar antenna transmission signal and the echo signal received after a period of time: Acquire a receiving signal of each receiving antenna according to an array element of each receiving antenna, a spacing between array elements, and an incident angle of an incident signal; Convert the received signals of N receiving antennas into vector form; The digital beamforming algorithm is used to weight the received signal in vector form to compensate for the phase difference caused by the time delay between the array elements, and to achieve the same-direction superposition of the N receiving antennas in a certain direction during the output process, so that the output power at the incident angle is maximized. The spatial power spectrum is constructed to complete the angle estimation of the target azimuth and obtain the initial azimuth of each target.

3. The method for monitoring heartbeat and breathing information of multiple persons according to claim 2, characterized in that: The maximum output power of digital beamforming is obtained by the following formula: P DBF (w)=a H (θ k )Ra(θ k ); Where P DBF (w) represents the maximum output power of digital beamforming, w represents the weight vector of N receiving antennas, and a(θ k ) is represented as the steering vector, e is a natural constant, j is an imaginary unit, d L is the spacing between array elements, λ is the signal wavelength, T is the transpose, θ k is represented as the incident angle of the kth incident signal, R is represented as the covariance matrix of the output signals of the N receiving antennas, and H is represented as the conjugate transpose; The spatial power spectrum is obtained by changing the weight vector to perform spatial beam scanning, and a detection threshold is preset. The beam direction greater than the detection threshold is the initial azimuth of the target.

4. The method for monitoring heartbeat and respiration information of multiple persons according to claim 1, characterized in that: Get the final azimuth of each target, including: According to the initial azimuth angle of the target, beam 1 is formed in this angle direction, and beam 2 is formed in the adjacent angle direction. The two beams have the same shape and partially overlap. The line connecting the intersection of beam 1 and beam 2 and the origin is taken as the equal signal axis, and the direction of the equal signal axis is recorded as Perform sum and difference processing on the signals simultaneously received by beam 1 and beam 2 to obtain a sum signal and a difference signal respectively; In the direction of the equal signal axis Nearby, the sum signal and the difference signal are approximately represented by using the first-order Taylor expansion to obtain the difference and ratio; Use the difference and ratio to judge the angle where the target deviates from the direction of the equal signal axis The size and direction of the angle, and then the final azimuth of each target is obtained; Among them, the difference and ratio are calculated by the following formula: In the formula, u Δ (θ) is expressed as a difference signal, u ∑ (θ) is expressed as a sum signal, θ re It is expressed as the angle of the target deviating from the direction of the equal signal axis. angle, F(θ) is the antenna directivity function, and S is the intermediate variable; The final azimuth of each target is angle θ re and sum.

5. The method for monitoring heartbeat and respiration information of multiple persons according to claim 1, characterized in that: Intermediate frequency signal S IF (x) is obtained by the following formula: In the formula, A IF It is represented by the amplitude of the intermediate frequency signal, B is represented by the bandwidth of the signal, and T c Expressed as the signal period, T c It is represented as the starting frequency of the signal, τ is represented as the delay between transmitting the signal and receiving the echo signal, It is represented as phase difference and t is represented as time.

6. The method for monitoring heartbeat and breathing information of multiple persons according to claim 1, characterized in that: Perform discrete wavelet transform on the intermediate frequency signal through the following formula: Where, ψ is the mother wavelet function, J, L∈Z, a0, b0 are constants, and Z is an integer; Discrete wavelet transform coefficient DWT(a0 J ,Lb0) is expressed as: Where y(t) represents the input signal of discrete wavelet transform.

7. The method for monitoring heartbeat and respiration information of multiple persons according to claim 6, characterized in that: The minimum mean square error algorithm minimizes the mean square value of the output error sequence e(n) by adjusting the filter coefficients, and modifies the filter coefficients according to this criterion, and approximates the desired signal by iteratively updating the filter coefficients; Among them, the input signal is DWT(a0 J ,Lb0), the output signal y(n) of the filter at time n can be expressed as: Where M is the order of the filter, w i (n) is the filter coefficient at time n; The output error sequence e(n) is obtained by: e(n)=D(n)-y(n); Where D(n) represents the expected output signal; The desired signal is approximated by iteratively updating the filter coefficients, using the following formula: In i (n+1)=in i (n)+2μ e(n) DWT(a0 J ,Lb0); In the formula, μ represents the step size factor.

8. The method for monitoring heartbeat and breathing information of multiple persons according to claim 1, characterized in that: Signal processing includes: phase unwrapping, Assume the current phase value is φ n , the phase value at the next moment is φ n+1 ; When φ n+1 -φ n ≥π, it indicates that the phase signal has jumped, and execute φ n+1 =φ n -2π; When φ n+1 -φ n <π, execute φ n+1 =φ n +2π, to ensure that the phase signal jump is less than π; The first-order phase difference is used to remove the constant phase offset. The adjacent phase values ​​are subtracted to remove the constant offset contained in the adjacent frames. Assume that the current frame phase is φ n , the phase of the previous frame is φ n-1 , the difference between the two φ n -φ n-1 As the current phase value φ n .

9. The method for monitoring heartbeat and respiration information of multiple persons according to claim 8, characterized in that: After phase information extraction, the breathing and heartbeat signals are separated using an IIR bandpass filter. The spectrum diagram of the signal is obtained through spectrum estimation and peak search is performed to obtain the breathing frequency and heartbeat frequency.

10. A multi-person heartbeat and respiratory information monitoring system, characterized in that: The method for monitoring heartbeat and respiratory information of multiple persons according to any one of claims 1 to 9 comprises: Radar antenna module, used for radar antenna adopting 1 transmit and N receive, emitting wide beam; The target azimuth module is used to receive the echo signals from N receiving antennas, and use the digital beamforming algorithm to perform multi-target recognition to obtain the initial azimuth of each target; then use the sum and difference angle measurement algorithm to correct the initial azimuth of each target to obtain the final azimuth of each target; The intermediate frequency signal module is used to mix the echo signal of each target azimuth with the local oscillator signal to obtain the intermediate frequency signal; The target distance module is used to use the intermediate frequency signal as the input of the discrete wavelet transform, obtain the spectrum of each target, and lock the distance unit information of each target; at the same time, the discrete wavelet transform coefficients are dynamically adjusted to adapt to the changes of the signal in combination with the adaptive filtering algorithm based on the minimum mean square error; The multi-target vital signs module is used to perform signal processing on the distance unit information of each target, including phase processing, bandpass filtering and spectrum estimation, to obtain multi-target vital signs information.

Images