Search

CN-121987179-A - Respiration monitoring system based on sensor

CN121987179ACN 121987179 ACN121987179 ACN 121987179ACN-121987179-A

Abstract

The invention relates to the technical field of respiration monitoring, and provides a respiration monitoring system based on a sensor, which comprises the following components: the system comprises a non-contact radar sensor, a signal preprocessing unit, an activity state monitoring module, a dynamic threshold adjusting unit, a respiratory signal reconstruction module and a respiratory event judging module, and is further additionally provided with a data fusion calibration unit. The sensor collects original respiratory vibration data of the thoracic cavity, an initial baseband signal is extracted through preprocessing, the monitoring module evaluates the activity state, the adjusting unit is adaptive to configuration parameters, the reconstruction module generates a high-fidelity respiratory effort signal through filtering and phase compensation, the judging module extracts feature recognition and classifies respiratory events, and the calibrating unit fuses the multi-source data to verify and correct results. The system can be deployed in multiple devices, realizes interference-free long-term monitoring, can accurately identify respiratory events and subdivide types, and greatly improves monitoring precision and suitability.

Inventors

  • LI YANG

Assignees

  • 西安交通大学医学院第一附属医院

Dates

Publication Date
20260508
Application Date
20260310

Claims (9)

  1. 1. A sensor-based respiration monitoring system comprising: at least one non-contact radar sensor configured to transmit electromagnetic waves to a thoracic region of a subject and receive echo signals thereof to obtain raw respiratory vibration data; A signal preprocessing unit, connected to the non-contact radar sensor, configured to perform clutter suppression and phase demodulation on the raw respiratory vibration data to extract an initial baseband signal containing respiratory displacement information; An activity state monitoring module configured to evaluate the physical activity intensity of the subject in real time based on the raw respiratory vibration data or the initial baseband signal, and generate an activity state indication signal; The dynamic threshold value adjusting unit is connected with the activity state monitoring module and is configured to adaptively adjust the signal-to-noise ratio calculation window length and the filtering cut-off frequency according to the activity state indication signal to generate a dynamic signal-to-noise ratio threshold value matched with the current activity state; The respiratory signal reconstruction module is respectively connected to the signal preprocessing unit and the dynamic threshold adjustment unit, and is configured to adaptively filter the initial baseband signal based on the dynamic signal-to-noise ratio threshold and reconstruct a respiratory waveform by utilizing a nonlinear phase compensation algorithm so as to generate a high-fidelity respiratory effort signal; and the respiratory event judging module is connected with the respiratory signal reconstructing module and is configured to perform time-frequency domain feature extraction based on the high-fidelity respiratory effort signal so as to identify an apnea, hypopnea or respiratory effort related arousal event and output a corresponding respiratory event evaluation result.
  2. 2. The sensor-based respiration monitoring system according to claim 1, characterized in that the non-contact radar sensor is a frequency modulated continuous wave radar or an ultra wideband pulse radar, the working frequency band of which is located in the 60GHz to 80GHz millimeter wave band.
  3. 3. The sensor-based respiration monitoring system of claim 1, wherein the activity state monitoring module comprises: the energy calculation operator module is used for calculating the signal energy integral of the original respiratory vibration data in the sliding time window; And the motion artifact identification sub-module is connected to the energy calculation sub-module and is used for comparing the signal energy integral with a preset static energy baseline, judging the motion state when the signal energy integral exceeds a first multiple of the static energy baseline, and judging the bed-leaving state when the signal energy integral exceeds a second multiple, wherein the first multiple is smaller than the second multiple.
  4. 4. The sensor-based respiration monitoring system of claim 1, wherein the dynamic threshold adjustment unit comprises: the table look-up sub-module pre-stores mapping relations between different activity intensity levels, signal-to-noise ratio threshold values and filter coefficients; And the parameter mapping sub-module queries the table look-up sub-module according to the activity state indication signal so as to output a corresponding dynamic signal-to-noise ratio threshold value and a real-time filter configuration parameter to the respiratory signal reconstruction module.
  5. 5. The sensor-based respiration monitoring system of claim 1, wherein the respiration signal reconstruction module further comprises: The lower limit and the upper limit of the passband frequency of the self-adaptive bandpass filter are adjusted in real time by the dynamic threshold adjusting unit according to the current respiratory frequency estimated value and the activity state; And the phase compensator is used for carrying out phase correction on the signal after passing through the adaptive band-pass filter by adopting a Hilbert transform or extended Kalman filtering algorithm so as to eliminate respiratory waveform distortion caused by nonlinear phase shift.
  6. 6. The sensor-based respiratory monitoring system of claim 1, wherein the respiratory event determination module comprises: The characteristic extraction submodule is used for extracting the amplitude variation coefficient, the peak interval and the power spectral density of the high-fidelity respiratory effort signal; and the obstructive and central type apnea classification sub-module is connected with the characteristic extraction sub-module and is configured to judge whether an apnea event is obstructive, central type or mixed type according to the presence and the phase characteristics of the respiratory effort and the waveform form of the high-fidelity respiratory effort signal.
  7. 7. The sensor-based respiratory monitoring system of claim 1, further comprising a data fusion calibration unit, coupled to the respiratory signal reconstruction module, configured to receive and fuse blood oxygen saturation signals or body position signals from auxiliary sensors, to assist in verifying respiratory event assessment results output by the respiratory event determination module.
  8. 8. The sensor-based respiration monitoring system according to claim 1, characterized in that the system is deployed in a bedside device, smart wearable device or wheelchair for undisturbed long-term respiration monitoring of a user.
  9. 9. A method of monitoring a sensor-based respiration monitoring system according to any of claims 1-8 comprising the steps of: S1, acquiring original respiratory vibration data of a tested person through a non-contact radar sensor; S2, preprocessing the original respiratory vibration data, and extracting an initial baseband signal containing respiratory displacement information; s3, evaluating the physical activity intensity of the tested person in real time and generating an activity state indicating signal; S4, adaptively adjusting a signal-to-noise ratio threshold according to the activity state indication signal to generate a dynamic signal-to-noise ratio threshold; s5, carrying out self-adaptive filtering and nonlinear phase compensation on the initial baseband signal based on the dynamic signal-to-noise ratio threshold value, and reconstructing to generate a high-fidelity respiration effort signal; S6, carrying out feature extraction and pattern recognition on the high-fidelity respiratory effort signal, and outputting a respiratory event evaluation result.

Description

Respiration monitoring system based on sensor Technical Field The invention relates to the technical field of respiration monitoring, in particular to a respiration monitoring system based on a sensor. Background The respiratory monitoring is a core technical means in the fields of sleep medicine, home health care, chronic disease monitoring and the like, accurate respiratory data acquisition and respiratory event identification can provide key basis for diagnosis and intervention of sleep apnea syndrome, chronic obstructive pulmonary disease and other symptoms, and meanwhile, in the fields of a care institution and home care, the respiratory monitoring for a user for a long time can realize timely early warning of abnormal conditions and reduce health risks, so that an efficient, accurate and interference-free respiratory monitoring system becomes an important research and development direction in the related fields. The existing respiration monitoring technology has many objective defects that firstly, a contact sensor is adopted, wearing constraint is strong, discomfort is easy to generate, long-term non-interference monitoring cannot be realized, single sensor data is easy to be misjudged due to environmental interference, two signal processing parameters are fixed values, the two signal processing parameters cannot be dynamically adjusted according to the moving states of a tested person, the person leaves a bed and other moving states, movement artifacts are easy to cause signal distortion and monitoring accuracy is greatly reduced, three signal filtering is easy to generate nonlinear phase shift to cause waveform distortion, only basic respiration events can be simply identified, type subdivision cannot be carried out on the apnea, and the requirement of clinical accurate analysis is difficult to meet. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a respiration monitoring system based on a sensor, which solves the problems of constraint in contact monitoring, easy receptor movement interference due to fixed signal processing parameters, easy distortion of respiration signals and low recognition precision of respiration events, which are not subdivided in the traditional respiration monitoring system. The respiration monitoring system based on the sensor comprises at least one non-contact radar sensor, is configured to emit electromagnetic waves to the chest region of a tested person and receive echo signals thereof so as to acquire original respiration vibration data; A signal preprocessing unit, connected to the non-contact radar sensor, configured to perform clutter suppression and phase demodulation on the raw respiratory vibration data to extract an initial baseband signal containing respiratory displacement information; An activity state monitoring module configured to evaluate the physical activity intensity of the subject in real time based on the raw respiratory vibration data or the initial baseband signal, and generate an activity state indication signal; The dynamic threshold value adjusting unit is connected with the activity state monitoring module and is configured to adaptively adjust the signal-to-noise ratio calculation window length and the filtering cut-off frequency according to the activity state indication signal to generate a dynamic signal-to-noise ratio threshold value matched with the current activity state; The respiratory signal reconstruction module is respectively connected to the signal preprocessing unit and the dynamic threshold adjustment unit, and is configured to adaptively filter the initial baseband signal based on the dynamic signal-to-noise ratio threshold and reconstruct a respiratory waveform by utilizing a nonlinear phase compensation algorithm so as to generate a high-fidelity respiratory effort signal; and the respiratory event judging module is connected with the respiratory signal reconstructing module and is configured to perform time-frequency domain feature extraction based on the high-fidelity respiratory effort signal so as to identify an apnea, hypopnea or respiratory effort related arousal event and output a corresponding respiratory event evaluation result. Preferably, the non-contact radar sensor is a frequency modulation continuous wave radar or an ultra-wideband pulse radar, and the working frequency band of the non-contact radar sensor is located in a millimeter wave band from 60GHz to 80 GHz. Preferably, the activity state monitoring module includes: the energy calculation operator module is used for calculating the signal energy integral of the original respiratory vibration data in the sliding time window; And the motion artifact identification sub-module is connected to the energy calculation sub-module and is used for comparing the signal energy integral with a preset static energy baseline, judging the motion state when the signal energy integral exceeds a first multiple of the static energy base