Search

CN-122000071-A - Cerebral apoplexy risk diagnosis and early warning system and method based on millimeter wave radar fusion

CN122000071ACN 122000071 ACN122000071 ACN 122000071ACN-122000071-A

Abstract

The invention relates to the technical field of intelligent medical monitoring and discloses a cerebral apoplexy risk diagnosis and early warning system and method based on millimeter wave radar fusion. The system comprises a sensor module, a micro-characterization decoupling module, a space isomerism module, an early warning judging module and a diagnosis evaluation plane mapping risk prediction point, wherein the sensor module is used for collecting data of a target object, the micro-characterization decoupling module is used for acquiring radio frequency echo signals and executing complex characteristic reconstruction, determining a space reflection anchor point and stripping physiological characteristic waveforms, the space isomerism module is used for constructing a virtual anatomical projection plane and generating a kinetic energy distribution thermodynamic diagram to calculate an asymmetry index, and the early warning judging module is used for executing cross-modal time sequence fusion to obtain an isomerism energy unbalance vector, mapping a risk prediction point on the diagnosis evaluation plane and executing grading risk early warning according to an evolution slope. The invention realizes the space visual mapping of the microscopic kinetic energy distribution of the thoracic cavity, solves the problem of high report missing rate of single-mode monitoring, and improves the early warning precision of pathological evolution in the early stage of sudden stroke.

Inventors

  • LI JINZHENG
  • CAI ZHENG
  • ZHANG YANG
  • ZHAO HONGCHAO

Assignees

  • 云南方圆计量校准检测服务有限公司

Dates

Publication Date
20260508
Application Date
20260407

Claims (10)

  1. 1. Cerebral apoplexy risk diagnosis and early warning system based on millimeter wave radar fusion, which is characterized in that the system comprises: the micro-feature decoupling module is used for acquiring a radio frequency echo signal representing micro-motion characteristics of a target object, performing complex feature reconstruction on the radio frequency echo signal to obtain an angle energy spectrum representing space reflection intensity, performing peak detection and coordinate mapping on the angle energy spectrum, determining a space reflection anchor point, performing phase unwrapping and physical displacement mapping based on the space reflection anchor point to obtain a physiological displacement sequence, performing multi-scale digital filtering on the physiological displacement sequence, and stripping physiological characteristic waveforms; The space isomerism module is used for constructing a virtual anatomic projection plane based on a space reflection anchor point, calling physiological characteristic waveforms in the virtual anatomic projection plane to execute physiological kinetic energy focusing mapping of grid dimensions to obtain a kinetic energy distribution thermodynamic diagram, dividing a double-side monitoring area in the kinetic energy distribution thermodynamic diagram, and executing two-dimensional space discrete summation calculation based on the double-side monitoring area to obtain an asymmetry index; The early warning judging module is used for performing cross-modal time sequence fusion based on the physiological characteristic waveform to obtain heterogeneous energy unbalance vectors, constructing a diagnosis and evaluation plane according to the asymmetry index to map the heterogeneous energy unbalance vectors to obtain risk prediction points, calculating an evolution slope through the risk prediction points, and performing grading risk early warning based on the evolution slope to obtain a risk early warning set; A bed body; And the sensor module at least comprises a millimeter wave radar sensor and a weighing sensor array.
  2. 2. The warning system for diagnosing risk in cerebral stroke based on millimeter wave radar fusion according to claim 1, wherein the complex feature reconstruction comprises: Continuously radiating a plurality of frequency modulation continuous waves to a monitored area containing a target object according to a preset pulse period through a millimeter wave radar sensor, capturing a reflected signal modulated by a thoracic cavity physiological micro-fluctuation phase, performing analog multiplication operation and low-pass filtering on the reflected signal to extract an intermediate frequency signal, and quantitatively sampling to obtain a radio frequency echo signal; Performing fast Fourier transform on the radio frequency echo signals to generate a distance spectrum, and defining a distance measurement unit with the largest energy amplitude in the distance spectrum as a target distance measurement unit; Extracting complex voltage values containing real part information and imaginary part information according to the position of the radio frequency echo signal in the target ranging unit, and combining the complex voltage values into a complex sampling sequence according to a time sequence; Calculating the instantaneous power of each complex voltage value and accumulating to obtain a single-frame total energy value, dividing the single-frame total energy value by the number of the frequency modulation continuous waves, and converting the single-frame total energy value into average energy power; and obtaining a background energy mean value of the monitored area in an empty state, taking the reciprocal of the background energy mean value as an environment correction coefficient, and multiplying the average energy power by the environment correction coefficient to obtain effective signal energy.
  3. 3. The warning system for cerebral apoplexy risk diagnosis based on millimeter wave radar fusion according to claim 2, wherein the installation modes of the millimeter wave radar sensor comprise two types, the first installation mode is to fixedly install the millimeter wave radar sensor above a bed head, and the second installation mode is to arrange the millimeter wave radar sensor in a mattress of a bed body.
  4. 4. The warning system for diagnosing risk in cerebral stroke based on millimeter wave radar fusion according to claim 3, wherein the method for acquiring the spatial reflection anchor point comprises the following steps: Acquiring complex voltage values of each independent channel of the millimeter wave radar sensor at the position of the target ranging unit, and sequencing the complex voltage values to construct a space measurement vector; presetting a plurality of candidate angles, calculating theoretical phase offset of each independent channel under each candidate angle, constructing angle guide vectors, and forming a reference vector set by each angle guide vector; Respectively performing inner product operation on the space measurement vector and each angle guide vector to obtain space reflection intensity, and combining corresponding candidate angles to obtain an angle energy spectrum; Searching a local extreme point with the largest value in the angle energy spectrum as an energy peak amplitude, and mapping the energy peak amplitude to obtain a target azimuth angle and a target pitch angle; Determining a radial distance value based on the target ranging unit, and calculating a transverse component, a longitudinal component and a height component by combining a target azimuth angle and a target pitch angle under a constructed radar local coordinate system; And performing rotation operation on the transverse component, the longitudinal component and the height component by utilizing a millimeter wave radar sensor, generating a rotated transverse component, a rotated longitudinal component and an absolute vertical height, and defining a three-dimensional physical coordinate point formed by the three components as a space reflection anchor point.
  5. 5. The mmwave radar fusion-based stroke risk diagnosis and early warning system of claim 4, wherein the physiological signature comprises: Calculating the original phase value of each complex voltage value in the complex sampling sequence by using an arctangent function, and arranging according to the time sequence of the pulse period to obtain an original phase sequence; calculating a phase difference value between adjacent pulse periods in an original phase sequence, performing numerical compensation operation according to the relation between the phase difference value and a preset jump threshold value, and eliminating phase winding to construct a continuous phase sequence; converting the phase values in the continuous phase sequence combined with the inherent wavelength of the radio frequency echo signal into an instantaneous displacement characteristic value, and combining according to a time sequence to obtain a physiological displacement sequence; Performing frequency selective filtering on the physiological displacement sequence in parallel in a preset respiratory frequency interval and a heartbeat frequency interval respectively by using a zero-phase translation digital band-pass filter, and extracting a respiratory characteristic waveform and a heartbeat characteristic waveform; and combining the respiration characteristic waveform and the heartbeat characteristic waveform to obtain the physiological characteristic waveform.
  6. 6. The warning system for stroke risk diagnosis based on millimeter wave radar fusion of claim 5, wherein the kinetic energy distribution thermodynamic diagram comprises: Constructing a two-dimensional physical section by taking a space reflection anchor point as a geometric reference center, intercepting a rectangular area, dividing the rectangular area into a plurality of two-dimensional discrete grids with three-dimensional geometric coordinates, and forming a virtual anatomic projection surface by all the two-dimensional discrete grids; Calculating the physical space linear distance from the two-dimensional discrete grid to each independent channel, and determining the theoretical phase delay amount by combining the inherent wavelength of the radio frequency echo signal; performing reverse phase rotation on the complex voltage value by utilizing the theoretical phase delay amount to perform phase compensation and coherent accumulation to obtain a local time domain reflection signal; In a preset respiratory frequency interval and a heartbeat frequency interval, performing frequency selective filtering on the local time domain reflection signals to obtain local physiological micro-motion components; And calculating kinetic energy density points by utilizing local physiological micro-motion components, and combining the kinetic energy density points corresponding to all the two-dimensional discrete grids according to space topology positions to generate a kinetic energy distribution thermodynamic diagram.
  7. 7. The warning system for stroke risk diagnosis based on millimeter wave radar fusion of claim 6, wherein the asymmetry index comprises: Taking a geometric reference center as an origin of coordinates, taking an anatomical sagittal plane projection line along a longitudinal direction parallel to a radar local coordinate system, and dividing a virtual anatomical projection plane into a left monitoring area and a right monitoring area; Extracting kinetic energy density points corresponding to each two-dimensional discrete grid in a left monitoring area to perform two-dimensional space discrete summation operation to obtain a left inching energy total value, extracting kinetic energy density points corresponding to each two-dimensional discrete grid in a right monitoring area to perform two-dimensional space discrete summation operation to obtain a right inching energy total value; Acquiring multi-frame radio frequency echo signals of a target object in a historical healthy baseline period, calculating the ratio of a left micro-energy total value to a right micro-energy total value corresponding to each frame of radio frequency echo signals, and taking the arithmetic average of all the ratios as a calibration factor; And calculating the calibration factor by combining the left micro-energy total value and the right micro-energy total value to obtain an asymmetry index.
  8. 8. The warning system for diagnosing risk in cerebral apoplexy based on millimeter wave radar fusion according to claim 7, wherein the method for acquiring the risk prediction point comprises the following steps: Obtaining effective peaks of the heartbeat characteristic waveform, calculating time difference values between adjacent effective peaks, and arranging according to a time sequence to obtain a heartbeat interval sequence; Resampling and time-frequency converting the heart beat interval sequence to obtain a power spectrum density function, respectively executing fixed integral operation on the power spectrum density function in a preset low-frequency band and high-frequency band to extract low-frequency power and high-frequency power, dividing the low-frequency power by the high-frequency power to obtain an energy ratio, and combining the energy ratio with a short-time variation to obtain a heart rate variability index; performing statistical calculation on the heart beat interval sequence to obtain atrial fibrillation load characteristics; carrying out statistical calculation on the respiratory characteristic waveform to obtain the physiological characteristics of the heart and lung; Setting a characteristic integration sliding window, and pressing the asymmetric index and the heart rate variability index into the characteristic integration sliding window to respectively obtain an asymmetric index sequence and a heart rate variability index sequence; and respectively calculating arithmetic average values of the asymmetric index sequence and the heart rate variability index sequence in the feature integration sliding window, and mapping the arithmetic average values of the asymmetric index sequence and the heart rate variability index sequence in a diagnosis evaluation plane to obtain a risk prediction point.
  9. 9. The warning system for risk diagnosis in cerebral stroke based on millimeter wave radar fusion according to claim 8, wherein the risk warning set comprises: Acquiring and calculating standard deviation and arithmetic mean value of an asymmetric index sequence and a heart rate variability index sequence in a historical health baseline period, and establishing the arithmetic mean value of the asymmetric index sequence and the heart rate variability index sequence as a personalized health anchor point in a diagnosis evaluation plane; Combining the standard deviation of the asymmetric index sequence, the standard deviation of the heart rate variability index sequence and a preset risk tolerance coefficient to construct a dynamic risk target area taking the personalized healthy anchor point as a geometric center; when the risk prediction point is separated from the dynamic risk target area, calculating the physical distance variation between the risk prediction point and the personalized health anchor point, and defining the physical distance variation as Euclidean distance increment; dividing the Euclidean distance increment by a preset total observation time length to obtain an evolution slope; Comparing the evolution slope with a preset first slope threshold value and a preset second slope threshold value, judging the risk level, triggering a corresponding early warning prompt, and combining the risk levels to obtain a risk early warning set.
  10. 10. The cerebral apoplexy risk diagnosis and early warning method based on millimeter wave radar fusion is applied to the cerebral apoplexy risk diagnosis and early warning system based on millimeter wave radar fusion as claimed in any one of claims 1 to 9, and is characterized in that the method comprises the following steps: acquiring a radio frequency echo signal representing microscopic motion characteristics of a target object, performing complex characteristic reconstruction on the radio frequency echo signal to obtain an angle energy spectrum representing spatial reflection intensity, performing peak detection and coordinate mapping on the angle energy spectrum, determining a spatial reflection anchor point, performing phase unwrapping and physical displacement mapping based on the spatial reflection anchor point to obtain a physiological displacement sequence, performing multi-scale digital filtering on the physiological displacement sequence, and stripping physiological characteristic waveforms; Constructing a virtual anatomic projection plane based on a space reflection anchor point, calling physiological characteristic waveforms in the virtual anatomic projection plane to execute physiological kinetic energy focusing mapping of grid dimensions to obtain a kinetic energy distribution thermodynamic diagram, dividing a double-side monitoring area in the kinetic energy distribution thermodynamic diagram, and executing two-dimensional space discrete summation calculation based on the double-side monitoring area to obtain an asymmetry index; Performing cross-modal time sequence fusion based on physiological characteristic waveforms to obtain heterogeneous energy unbalance vectors, constructing a diagnosis evaluation plane according to asymmetry indexes to map the heterogeneous energy unbalance vectors to obtain risk prediction points, calculating an evolution slope through the risk prediction points, and performing hierarchical risk early warning based on the evolution slope to obtain a risk early warning set.

Description

Cerebral apoplexy risk diagnosis and early warning system and method based on millimeter wave radar fusion Technical Field The invention relates to the technical field of intelligent medical monitoring, in particular to a cerebral apoplexy risk diagnosis and early warning system and method based on millimeter wave radar fusion. Background With the rapid development of intelligent monitoring and ambulatory medical technology, non-contact vital sign monitoring has shown great clinical value in early warning and risk assessment of cerebral stroke. Prognosis of cerebral stroke is highly dependent on the timeliness of treatment, but most patients lack significant clinical symptoms before onset, resulting in missing the optimal treatment window. Currently, millimeter wave radar technology can realize noninductive monitoring on the details of a heartbeat cycle and a breathing mode by virtue of the advantages of high precision, non-contact, privacy protection and the like, however, the traditional monitoring method generally faces the bottlenecks of 'single dimension' and 'static threshold'. The prior art focuses on macroscopic monitoring of a single physiological parameter or relies on isolated vital sign fluctuations for risk assessment. This ignores the complex multimodal physiological co-evolution effects of stroke precursors, namely the deep coupling relationship between autonomic imbalance of heart, breathing pattern disorder and neuromotor hypofunction. In an actual scene, only the radial distance information is mastered, the target thoracic cavity and surrounding static strong reflectors cannot be effectively distinguished, so that the spatial directivity of a physiological signal source is fuzzy, and weak physiological micro-vibration is often covered due to the imperfect symmetry of human anatomy structures and the interference of environmental background noise, so that the symmetry break of unilateral neuromuscular movement caused by early stroke damage is difficult to accurately quantify. Although the traditional prediction model can identify partial abnormality, the pre-warning sensitivity and the false alarm rate are difficult to balance due to incapability of effectively decoupling heterogeneous interferences such as individual body type, detection distance, environmental attenuation and the like, so that how to change macroscopic vital sign records into cross-system collaborative analysis aiming at cerebral apoplexy pathology mechanism from static monitoring of single parameters to dynamic deep fusion of multi-mode physiological characteristics is difficult, the limitation of individual difference and environmental noise is broken through, and ultra-early accurate pre-warning of cerebral apoplexy risk is realized, and the method is a technical problem to be solved in the field. In the prior art, china patent with the authority of publication number CN118648887B discloses a non-contact real-time physiological sign monitoring system based on millimeter wave radar, the system comprises units such as millimeter wave radar, clutter suppression module, signal separation module and the like, intermediate frequency signals are obtained through transmitting electromagnetic waves, a four-dimensional data matrix is constructed, a constant false alarm rate detection, self-adaptive distance unit selection and optimization variation modal decomposition algorithm are utilized, non-contact extraction of physiological signs such as respiration, heartbeat and the like is realized, the problem of environmental clutter suppression and physiological signal separation is solved at the core, and the accuracy of sign monitoring is improved. The Chinese patent with the grant bulletin number of CN116269249B discloses a cerebral apoplexy risk prediction method and a cerebral apoplexy risk prediction system, wherein the risk prediction system is constructed by collecting multiple physiological indexes of a human body and combining an algorithm model, so that quantitative evaluation of cerebral apoplexy occurrence risk is realized, and data support is provided for home and clinical cerebral apoplexy risk screening. However, although the two prior arts have a certain application value in non-contact physiological monitoring and cerebral apoplexy risk prediction, the core pain points of multi-mode feature fusion, spatial signal positioning and individual dynamic evaluation in early cerebral apoplexy early warning cannot be solved. The CN118648887B focuses on the extraction and separation of the single-dimensional cardiopulmonary signs, does not involve bilateral motion symmetry analysis of the chest of the human body, cannot capture the energy imbalance characteristics caused by the damage of the single-sided neuromuscular in the early stage of cerebral apoplexy, and lacks the deep conversion logic from physiological signals to pathological early warning. CN116269249B relies on static threshold analysis of conventional physiologi