Pelvic floor muscle micro-motion signal detection and respiratory interference removal method
The variational mode separation algorithm removes respiratory interference in millimeter wave radar detection, solving the interference problem in pelvic floor muscle motion signal detection, improving detection accuracy and robustness, and ensuring monitoring accuracy and reliability.
Patent Information
- Application Number
- CN202411581163.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-07
- Publication Date
- 2025-02-25
AI Technical Summary
When using millimeter-wave radar to detect pelvic floor muscle movement signals, interference with physiological signals such as breathing and heartbeat is difficult to effectively remove, resulting in the accuracy and reliability of pelvic floor muscle movement signals.
By establishing equation constraints, the energy overlap between the pelvic floor muscle movement signal and the respiratory interference signal is minimized, and the breathing and its harmonic interference is adaptively separated and suppressed, and the pure pelvic floor muscle movement signal is extracted.
It improves the detection accuracy of pelvic floor muscle movement signals, enhances the robustness of the detection, adapts to physiological differences in different individuals, and ensures the accuracy and reliability of pelvic floor muscle movement monitoring.
Smart Images
Figure CN119498812A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of millimeter wave radar and pelvic floor muscle movement detection, and in particular relates to a method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference. Background Art
[0002] The pelvic floor muscles are an important muscle group for the human body to control urination, defecation and maintain the normal function of the pelvic organs. In recent years, with the change of lifestyle and the increase of postpartum women and the elderly, pelvic floor muscle dysfunction has become increasingly common. Pelvic floor muscle dysfunction not only leads to health problems such as urinary incontinence and pelvic organ prolapse, but also seriously affects the quality of life of patients. In order to effectively carry out rehabilitation training, functional assessment and clinical treatment of the pelvic floor muscles, the demand for accurate and convenient pelvic floor muscle movement monitoring technology is growing.
[0003] Traditional pelvic floor muscle assessment methods mostly use bioelectrodes or pressure sensors. Although these methods can provide direct muscle activity information, they have certain limitations. First, such devices are generally contact or invasive, complicated to operate, and easily cause discomfort to patients. In addition, frequent operation will lead to reduced data consistency. Secondly, these methods are limited by the detection environment and it is difficult to provide continuous and dynamic monitoring results, especially in daily life. The application is greatly restricted.
[0004] In order to solve these problems, non-contact millimeter-wave radar technology has become a new means of detecting pelvic floor muscle movement. Millimeter-wave radar is a detection technology based on high-frequency electromagnetic waves, which has the characteristics of strong penetration, good anti-interference ability and high precision. In medical testing, millimeter-wave radar can be used to detect tiny physiological signals, such as breathing, heartbeat, and muscle movement. Through the micro-motion detection technology of millimeter-wave radar, it is possible to capture the subtle movements of the pelvic floor muscles in real time without contacting the skin, which is extremely convenient and comfortable for patients.
[0005] However, in actual applications, the signals detected by millimeter-wave radar not only include the movement of the target muscles, but also capture the interference of physiological signals such as breathing and heartbeat. The interference of respiratory signals and their harmonics is particularly prominent, making it difficult to accurately extract the pelvic floor muscle movement signals. The presence of respiratory interference will directly affect the judgment of the movement state of the pelvic floor muscles and reduce the accuracy and reliability of monitoring. In this context, how to effectively remove the interference of breathing and its harmonics in the mixed signals collected by millimeter-wave radar has become a key technical difficulty. Summary of the invention
[0006] Technical problem solved: The present invention discloses a method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference, which uses millimeter-wave radar for non-contact detection of human pelvic floor muscle groups, and is more convenient and comfortable. The variational mode separation algorithm adopted at the same time has significant advantages and can effectively solve the problem of respiratory signal interference. In the present invention, by establishing an equality constraint, the energy overlap between the pelvic floor muscle movement signal and the respiratory interference signal is minimized, and the respiratory and its harmonic interference are adaptively separated and suppressed, ensuring that the pelvic floor muscle movement signal finally extracted is purer. This algorithm not only improves the detection accuracy, but also can adapt to the physiological differences of different individuals, and enhances its robustness in practical applications.
[0007] Technical solution:
[0008] A method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference, the method comprising the following steps:
[0009] Step A, millimeter wave radar echo data collection. The millimeter wave radar is aimed at the pelvic floor muscle part, and the transmitting antenna transmits an electromagnetic wave signal, which is reflected when it touches the surface of the pelvic floor muscle part and is received by the receiving antenna. The pelvic floor muscle movement signal is modulated in the echo signal, and then a series of front-end processing is performed to obtain a beat frequency signal and is sampled by the ADC as a digital signal.
[0010] Step B, motion signal waveform extraction. After obtaining the digital signal sampled in the above step, it is necessary to extract the motion signal waveform modulated in the collected data. First, the distance gate where the pelvic floor muscle is located is determined by distance-dimensional Fourier transform, and then the signal of the distance gate where the pelvic floor muscle is located is phase extracted to obtain a time domain signal containing pelvic floor muscle movement and respiratory interference movement.
[0011] Step C, performing variational mode separation on the time domain signal containing pelvic floor muscle movement and respiratory interference obtained in the above step. Based on variational mode decomposition, variational mode separation minimizes the energy overlap of pelvic floor muscle movement signals and respiratory and harmonic interference signals through equality constraints, and adaptively suppresses respiratory and harmonic signals, thereby obtaining a pure pelvic floor muscle movement signal.
[0012] Step A further comprises:
[0013] FMCW radar is used to transmit electromagnetic waves to sense pelvic floor muscle movement, where the radar frequency band includes 24 GHz, 60 GHz or 77 GHz; the FMCW radar transmission signal can be expressed as:
[0014] s t (t fast )=exp[j(2πf c t fast +πktfast 2 +θ)],t fast ∈(-T / 2,T / 2)
[0015] where f c is the center frequency, T is the interval of the chirp signal, t fast is the fast time, k is the frequency modulation slope, k = B / T, B is the bandwidth, θ is the initial phase, and the echo signal s received by the radar r (t) is the delay of the transmitted signal after a certain period of time, and the signal received by the receiving antenna is
[0016] s r (t fast )=s t (t fast -t delay )
[0017] t delay It represents the delay of the echo signal. After the echo signal passes through the low noise amplifier, mixing, and filtering, the beat frequency signal is obtained.
[0018]
[0019] Where R represents the absolute distance from the radar antenna to the pelvic floor muscle, λ c represents the wavelength. The radar signal is transmitted and received every t time, which can be written as
[0020]
[0021] R(t) can be regarded as the distance between the pelvic floor muscle and the radar antenna varying with time, that is, R(t) includes the pelvic floor muscle movement signal.
[0022] Finally, the digital signal is obtained through ADC sampling, which can be used for subsequent signal processing.
[0023] Step B further comprises:
[0024] The fast time Fourier transform of the beat frequency signal yields:
[0025]
[0026] in is the beat frequency, which is proportional to the distance of the pelvic floor muscle. Select a suitable f′ so that f′ is closest to f0, and then you get the distance dimension signal:
[0027]
[0028] Where η = σTsinc (T (f′-f0)) is a fixed constant, and the phase information of s (t) contains the pelvic floor muscle movement signal and the respiratory interference signal. In actual use, the reflected signal of other stationary targets will produce DC interference, that is:
[0029]
[0030] Where d is a complex number, indicating DC interference, the above formula is written as an I / Q signal:
[0031]
[0032] It can be seen that the phase information of s(t) contains the pelvic floor muscle movement signal and respiratory interference. Since the presence of DC will lead to the failure of phase recovery, it is necessary to remove the DC first, and the phase signal obtained by inverse tangent demodulation is:
[0033]
[0034] Where R(t) is the time domain waveform containing the pelvic floor muscle movement signal and the respiratory interference signal.
[0035] Step C further comprises:
[0036] As mentioned above, the acquired time domain signal contains the respiratory motion mode r r (t) and pelvic floor muscle movement mode r p (t), that is
[0037] R(t)=r p (t)+r r (t)
[0038] The respiratory motion modes are concentrated around their center frequency and are therefore solved by minimizing the bandwidth as a variational problem:
[0039]
[0040] Where J1 represents the bandwidth of the breathing signal. Through Hilbert transform, the relevant analytical signal is obtained, and then it is moved to the baseband through spectrum shifting. Finally, the bandwidth is estimated by the L2 norm of the gradient. represents the derivative, ω r Represents the fundamental frequency of the breathing signal.
[0041] The overlapping spectrum of the pelvic floor muscle movement mode and the respiratory movement mode should be minimized, that is, the energy of the pelvic floor muscle movement mode should be suppressed in the frequency band where the respiratory movement mode is located. In order to meet these restrictions, the r p The energy of (t) is considered as r p (t) and r r(t). Therefore, a series of frequency responses are used:
[0042]
[0043] Among them, the filter At frequency ω = kω r , k = 1, 2, ... K has infinite gain, which leads to infinite sensitivity to the frequency of the fundamental breathing frequency and its harmonics, γ k It is a parameter to adjust the notch depth of the filter.
[0044] In order to minimize r p (t) and r r (t) The spectrum overlap energy between the two, increasing the penalty function
[0045]
[0046] Among them, β k (t) is the filter The inverse Fourier transform of .
[0047] To this end, the problem of finding the desired pattern can be modeled as a constrained minimization problem, where the combination of J1 and J2 is minimized under the linear equality constraint:
[0048]
[0049] r p (t)+r r (t) = R(t)
[0050] Where α represents the parameter for adjusting J1 and J2.
[0051] Step C further comprises:
[0052] By combining the quadratic penalty term and the Lagrange multiplier, the augmented Lagrangian function can be expressed as:
[0053]
[0054] Where λ represents the Lagrange multiplier.
[0055] The problem is solved as a series of iterative sub-optimizations using the alternating direction method of the multiplier algorithm. Each time a sub-optimization problem is solved, the result is iteratively updated in the original problem, that is, one of the variables (r p , r p ,ω r ) while keeping the other variables at their most recent values, and then repeat the process for the other two variables.
[0056] Beneficial effects:
[0057] First, the method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference of the present invention uses millimeter wave radar for non-contact detection of human pelvic floor muscle groups, which is more convenient and comfortable than traditional contact equipment. The method is simple and effective, has reliable performance, and is easy to implement.
[0058] Second, the pelvic floor muscle micro-motion signal detection and respiratory interference removal method of the present invention adopts a variational mode separation algorithm based on a variational mode decomposition algorithm to minimize the energy overlap between the pelvic floor muscle movement signal and the respiratory and harmonic interference signals, thereby achieving adaptive suppression of respiratory and harmonic interference signals and obtaining a pure pelvic floor muscle movement signal. This algorithm not only improves the accuracy of detection, but also can adapt to the physiological differences of different individuals, enhancing its robustness in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 This is a flow chart of the method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference of the present invention. DETAILED DESCRIPTION
[0060] The following examples will enable those skilled in the art to more fully understand the present invention, but are not intended to limit the present invention in any way.
[0061] See also Figure 1 The present invention discloses a method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference, and the method comprises the following steps:
[0062] Step A, millimeter wave radar echo data collection. The millimeter wave radar is aimed at the pelvic floor muscle part, and the transmitting antenna transmits an electromagnetic wave signal, which is reflected when it touches the surface of the pelvic floor muscle part and is received by the receiving antenna. The pelvic floor muscle movement signal is modulated in the echo signal, and then a series of front-end processing is performed to obtain a beat frequency signal and is sampled by the ADC as a digital signal.
[0063] Step B, motion signal waveform extraction. After obtaining the digital signal sampled in the above step, it is necessary to extract the motion signal waveform modulated in the collected data. First, the distance gate where the pelvic floor muscle is located is determined by distance-dimensional Fourier transform, and then the signal of the distance gate where the pelvic floor muscle is located is phase extracted to obtain a time domain signal containing pelvic floor muscle movement and respiratory interference movement.
[0064] Step C, performing variational mode separation on the time domain signal containing pelvic floor muscle movement and respiratory interference obtained in the above step. Based on variational mode decomposition, variational mode separation minimizes the energy overlap of pelvic floor muscle movement signals and respiratory and harmonic interference signals through equality constraints, and adaptively suppresses respiratory and harmonic signals, thereby obtaining a pure pelvic floor muscle movement signal.
[0065] The pelvic floor muscle micro-motion signal detection and respiratory interference removal method of the present invention corresponds to a pelvic floor muscle micro-motion signal detection and respiratory interference removal method system, which includes a signal acquisition and acquisition module, a signal preprocessing module, and a respiratory interference suppression and pelvic floor muscle movement extraction module. The signal acquisition and acquisition module is used to detect pelvic floor muscle movement using a millimeter wave radar system, and capture the echo signal containing the pelvic floor muscle movement; the signal preprocessing module obtains an orthogonal I / Q signal after Fourier transformation and pelvic floor muscle position corresponding range gate selection, and then performs DC estimation and inverse tangent phase demodulation on the I / Q signal to obtain a mixed signal containing pelvic floor muscle movement and respiratory interference; the respiratory interference suppression and pelvic floor muscle movement extraction module, based on variational mode decomposition, minimizes the energy overlap of pelvic floor muscle movement signals and respiratory and its harmonic interference signals through equation constraints, and adaptively suppresses respiratory and its harmonic signals, thereby obtaining a pure pelvic floor muscle movement signal.
[0066] The signal collection and acquisition process includes:
[0067] FMCW radar is used to transmit electromagnetic waves to sense pelvic floor muscle movement, where the radar frequency band includes 24 GHz, 60 GHz or 77 GHz; the FMCW radar transmission signal can be expressed as:
[0068] s t (t fast )=exp[j(2πf c t fast +πkt fast 2 +θ)],t fast ∈(-T / 2,T / 2)
[0069] where f c is the center frequency, T is the interval of the chirp signal, t fast is the fast time, k is the frequency modulation slope, k = B / T, B is the bandwidth, θ is the initial phase, and the echo signal s received by the radar r (t) is the delay of the transmitted signal after a certain period of time, and the signal received by the receiving antenna is
[0070] s r (t fast )=s t (t fast -t delay )
[0071] t delay It represents the delay of the echo signal. After the echo signal passes through the low noise amplifier, mixing, and filtering, the beat frequency signal is obtained.
[0072]
[0073] Where R represents the absolute distance from the radar antenna to the pelvic floor muscle, λ c represents the wavelength. The radar signal is transmitted and received every t time, which can be written as
[0074]
[0075] R(t) can be regarded as the distance between the pelvic floor muscle and the radar antenna varying with time, that is, R(t) includes the pelvic floor muscle movement signal.
[0076] Finally, the digital signal is obtained through ADC sampling, which can be used for subsequent signal processing.
[0077] The signal preprocessing process includes:
[0078] The fast time Fourier transform of the beat frequency signal yields:
[0079]
[0080] in is the beat frequency, which is proportional to the distance of the pelvic floor muscle. Select a suitable f′ so that f′ is closest to f0, and then you get the distance dimension signal:
[0081]
[0082] Where η = σTsinc (T (f′-f0)) is a fixed constant, and the phase information of s (t) contains the pelvic floor muscle movement signal and the respiratory interference signal. In actual use, the reflected signal of other stationary targets will produce DC interference, that is:
[0083]
[0084] Where d is a complex number, indicating DC interference, the above formula is written as an I / Q signal:
[0085]
[0086] It can be seen that the phase information of s(t) contains the pelvic floor muscle movement signal and respiratory interference. Since the presence of DC will lead to the failure of phase recovery, it is necessary to remove the DC first, and the phase signal obtained by inverse tangent demodulation is:
[0087]
[0088] Where R(t) is the time domain waveform containing the pelvic floor muscle movement signal and the respiratory interference signal.
[0089] The respiratory interference suppression and pelvic floor muscle movement extraction module includes:
[0090] As mentioned above, the acquired time domain signal contains the respiratory motion mode r r (t) and pelvic floor muscle movement mode r p (t), that is
[0091] R(t)=r p (t)+r r (t)
[0092] The respiratory motion modes are concentrated around their center frequency and are therefore solved by minimizing the bandwidth as a variational problem:
[0093]
[0094] Where J1 represents the bandwidth of the breathing signal. Through Hilbert transform, the relevant analytical signal is obtained, and then it is moved to the baseband through spectrum shifting. Finally, the bandwidth is estimated by the L2 norm of the gradient. represents the derivative, ω r Represents the fundamental frequency of the breathing signal.
[0095] The overlapping spectrum of the pelvic floor muscle movement mode and the respiratory movement mode should be minimized, that is, the energy of the pelvic floor muscle movement mode should be suppressed in the frequency band where the respiratory movement mode is located. In order to meet these restrictions, the r p The energy of (t) is considered as r p (t) and r r (t). Therefore, a series of frequency responses are used:
[0096]
[0097] Among them, the filter At frequency ω = kω r , k = 1, 2, ... K has infinite gain, which leads to infinite sensitivity to the fundamental frequency of breathing and its harmonics, γ k It is a parameter to adjust the notch depth of the filter.
[0098] In order to minimize r p (t) and r r (t) The spectrum overlap energy between the two, increasing the penalty function
[0099]
[0100] Among them, β k (t) is the filter The inverse Fourier transform of .
[0101] To this end, the problem of finding the desired pattern can be modeled as a constrained minimization problem, where the combination of J1 and J2 is minimized under the linear equality constraint:
[0102]
[0103] Where α represents the parameter for adjusting J1 and J2.
[0104] By combining the quadratic penalty term and the Lagrange multiplier, the augmented Lagrangian function can be expressed as:
[0105]
[0106] Where λ represents the Lagrange multiplier.
[0107] The problem is solved as a series of iterative sub-optimizations using the alternating direction method of the multiplier algorithm. Each time a sub-optimization problem is solved, the result is iteratively updated in the original problem, that is, one of the variables (r p , r p ,ω r ) while keeping the other variables at their most recent values, and then repeat the process for the other two variables.
[0108] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.
Claims
1. A method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference, characterized in that: The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference comprises the following steps: Step A, collecting millimeter wave radar echo data; specifically including: The millimeter wave radar aims at the pelvic floor muscle part and transmits electromagnetic wave signals through the transmitting antenna. When the electromagnetic wave signals transmitted by the antenna touch the surface of the pelvic floor muscle part, they are reflected and received by the receiving antenna. The received pelvic floor muscle movement signals are modulated into the echo signals. After the front-end processing, the beat frequency signals are obtained and sampled into digital signals by the ADC. Step B, extracting the motion signal waveform; specifically comprising: Extract the motion signal waveform modulated in the collected data from the sampled digital signal; determine the range gate where the pelvic floor muscle is located through distance-dimensional Fourier transform, and then perform phase extraction on the signal of the range gate where the pelvic floor muscle is located to obtain a time domain signal containing pelvic floor muscle movement and respiratory interference movement; Step C, performing variational mode separation on the obtained time domain signal containing pelvic floor muscle movement and respiratory interference, minimizing the energy overlap of pelvic floor muscle movement signals and respiratory and its harmonic interference signals through equality constraints, adaptively suppressing respiratory and its harmonic signals, and obtaining a pure pelvic floor muscle movement signal.
2. The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference according to claim 1, characterized in that: Step A further comprises: FMCW radar is used to transmit electromagnetic waves to sense pelvic floor muscle movement, wherein the radar frequency band includes 24GHz, 60GHz or 77GHz; the FMCW radar transmission signal is expressed as: s t (t fast )=exp[j(2πf c t fast +πkt fast 2 +θ)],t fast ∈(-T / 2,T / 2) where f c is the center frequency, T is the interval of the chirp signal, t fast is the fast time, k is the frequency modulation slope, k = B / T, B is the bandwidth, θ is the initial phase, and the echo signal s received by the radar r (t) is the delay of the transmitted signal after a certain period of time. The signal received by the receiving antenna is: s r (t fast )=s t (t fast -t delay ) where t delay It represents the delay of the echo signal. After the echo signal passes through the low noise amplifier, mixing and filtering, the beat frequency signal is obtained as follows: Where R represents the absolute distance from the radar antenna to the pelvic floor muscle, λ c represents the wavelength; the radar signal is transmitted and received once every t time, and the radar signal is: Where R(t) is the distance between the pelvic floor muscle site and the radar antenna varying with time; The digital signal is obtained through ADC sampling.
3. The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference according to claim 1, characterized in that: Step B further comprises: The fast time Fourier transform of the beat frequency signal yields: in is the beat frequency, which is proportional to the distance of the pelvic floor muscle. Select a suitable f′ so that f′ is closest to f0, and get the distance dimension signal: Where η = σTsinc (T (f′-f0)) is a fixed constant, the phase information of s (t) contains the pelvic floor muscle movement signal and the respiratory interference signal; the reflected signal of other stationary targets produces DC interference, that is: Where d is a complex number, indicating DC interference, the above formula is written as an I / Q signal: The phase information of s(t) includes pelvic floor muscle movement signals and respiratory interference; After removing DC, the phase signal obtained by inverse tangent demodulation is: Where R(t) is the time domain waveform containing the pelvic floor muscle movement signal and the respiratory interference signal.
4. The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference according to claim 1, characterized in that: Step C further comprises: The time domain signal R(t) contains the respiratory motion mode r r (t) and pelvic floor muscle movement mode r p (t): R(t)=r p (t)+r r (t) The respiratory motion modes are concentrated around their center frequency and are solved by minimizing the bandwidth as a variational problem: Where J1 represents the bandwidth of the breathing signal. Through Hilbert transform, the relevant analytical signal is obtained, and then moved to the baseband through spectrum shifting. Finally, the bandwidth is estimated by the L2 norm of the gradient. represents the derivative, ω r Indicates the fundamental frequency of the breathing signal. ; The filtered r within the respiratory spectrum p The energy of (t) is considered as r p (t) and r r An indicator of the spectral overlap between (t) using a series of frequency responses is: Among them, the filter At frequency ω = kω r ,k=1,2,…K has infinite gain, γ k It is a parameter to adjust the notch depth of the filter; Add a penalty function to minimize r p (t) and r r The spectrum overlap energy between (t): Among them, β k (t) is the filter The inverse Fourier transform of The problem of finding the desired pattern is modeled as a constrained minimization problem, where the combination of J1 and J2 is minimized under the linear equality constraint: r p (t)+r r (t)=R(t) Where α represents the parameter for adjusting J1 and J2.
5. The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference according to claim 4, characterized in that: The process of solving the equality constrained minimization problem includes: By combining the quadratic penalty term and the Lagrange multiplier, the augmented Lagrangian function is expressed as: Where λ represents the Lagrange multiplier; The equality constraint minimization problem is solved as a series of iterative sub-optimizations by the alternating direction method of the multiplier algorithm; each time a sub-optimization problem is solved, and the result is iteratively updated in the original problem. At each iteration, one of the variables (r p , r p ,ω r ) while keeping the latest values of other variables unchanged; The solved r p It is a pure pelvic floor muscle movement signal with respiratory interference signals removed.
6. The method for detecting pelvic floor muscle micro-motion signals and removing respiratory interference according to claim 1, characterized in that: The variational mode separation algorithm in step C is based on the variational mode decomposition algorithm, and achieves adaptive suppression of respiratory and harmonic interference signals by minimizing the energy overlap between pelvic floor muscle movement signals and respiratory and harmonic interference signals.