Search

CN-121987905-A - Breathing machine false triggering prevention method, equipment and medium based on self-adaptive Kalman filtering

CN121987905ACN 121987905 ACN121987905 ACN 121987905ACN-121987905-A

Abstract

The invention discloses a breathing machine false triggering prevention method, equipment and medium based on self-adaptive Kalman filtering, and relates to the technical field of medical equipment. The method comprises the steps of obtaining respiratory flow signals and respiratory pressure signals of a patient in real time, constructing a state space model, carrying out recursive optimal state estimation on the state space model by using the respiratory flow signals and the respiratory pressure signals as observation input and adopting an adaptive Kalman filter, carrying out real-time separation to obtain optimal estimated values of all components, extracting multidimensional characteristic parameters based on the optimal estimated values and calculating comprehensive decision indexes, when the respiratory flow signals exceed a trigger threshold, identifying whether a current event is cardiogenic oscillation according to comparison results of the comprehensive decision indexes and the decision threshold, and controlling a trigger response of a breathing machine according to the identification results. The invention fundamentally solves the problem of false triggering of the breathing machine caused by cardiogenic oscillation, and remarkably improves the man-machine synchronism and the treatment safety.

Inventors

  • LIU YU
  • YANG QUANGANG
  • LV YIZHI
  • FENG KAI

Assignees

  • 可孚医疗科技股份有限公司

Dates

Publication Date
20260508
Application Date
20260121

Claims (10)

  1. 1. An anti-false triggering method of a respirator based on adaptive Kalman filtering, which is characterized by comprising the following steps: Acquiring a respiratory flow signal and a respiratory pressure signal of a patient in real time; constructing a state space model comprising a respiratory flow component, a respiratory pressure component, a flow heartbeat component and a pressure heartbeat component; the respiratory flow signal and the respiratory pressure signal are used as observation input, and an adaptive Kalman filter is adopted to carry out recursive optimal state estimation on the state space model, so as to obtain optimal estimated values of respiratory flow components, respiratory pressure components, flow heart beat components and pressure heart beat components; Calculating a multidimensional characteristic parameter based on the optimal estimated values of the flow heart beat component, the pressure heart beat component, the respiratory flow component and the respiratory pressure component; Calculating a comprehensive decision index based on the multidimensional characteristic parameters, and identifying whether the current event is cardiogenic oscillation or not according to a comparison result of the comprehensive decision index and a decision threshold when the respiratory flow signal exceeds a trigger threshold; And controlling the event response of the breathing machine to the respiratory flow signal exceeding a trigger threshold according to the identification result, wherein the trigger is restrained if the cardiac oscillation is identified, and the trigger is allowed to be executed if the normal respiratory effort is identified.
  2. 2. The method for preventing false triggering of a ventilator based on adaptive kalman filtering according to claim 1, wherein the state space model includes a state transition equation and an observation equation, and the specific expressions are respectively: , ; , ; , ; Wherein, the 、 State vectors at the k-1 th time and the k-1 th time respectively; a state transition matrix representing a kth time; Process noise representing the kth time; A respiratory flow component representing a kth time; Indicating the respiratory flow rate change at the kth time; a respiratory pressure component representing a kth time; a respiratory pressure change rate at the kth time; representing a flow heart beat component at a kth time; Representing a pressure heart beat component at a kth time; an observation vector indicating a kth time; Representing an observation matrix; An observation noise at the kth time; a respiratory flow observation value at the kth time; a respiratory pressure observation value at a kth time; Representing a signal sampling time interval; Representing the flow decay factor; representing the pressure decay coefficient.
  3. 3. The respiratory machine false triggering prevention method based on the adaptive Kalman filtering according to claim 1, wherein the recursive optimal state estimation of the state space model by adopting the adaptive Kalman filter comprises the following iterative steps: Based on the state optimal posterior estimation value at the last iteration moment and the state transition equation of the state space model, predicting a state prior estimation value at the current moment and a corresponding prior estimation error covariance matrix; calculating the difference between the actual observation signal vector at the current moment and the observation predicted value obtained according to the state priori estimated value to obtain an innovation vector; estimating and updating a process noise covariance matrix and an observation noise covariance matrix of the state space model in real time based on the statistical characteristics of the innovation vector in a preset sliding time window; calculating a Kalman gain matrix at the current moment according to the prior estimation error covariance matrix and the updated observation noise covariance matrix; weighting the innovation vector by using the Kalman gain matrix, and correcting the state prior estimated value to obtain a state optimal posterior estimated value at the current moment; Updating the posterior estimation error covariance matrix of the state optimal posterior estimation value, and using the posterior estimation error covariance matrix for the state prediction step of the next iteration moment.
  4. 4. The adaptive kalman filter based ventilator false triggering prevention method of claim 1, wherein the multi-dimensional feature parameters include flow heart beat energy ratio, pressure heart beat energy ratio, flow change pattern score, flow-pressure heart beat component coherence, heart beat flow component relative amplitude and heart beat pressure component relative amplitude.
  5. 5. The adaptive kalman filter based ventilator false triggering prevention method according to claim 4, wherein the flow heart beat energy ratio, the pressure heart beat energy ratio, the flow change mode score, the flow-pressure heart beat component coherence, the heart beat flow component relative amplitude and the heart beat pressure component relative amplitude are calculated by the following formulas: ; ; ; ; , ; Wherein, the Representing flow heart beat energy ratio; the optimal estimated value of the heart beat component of the flow at the ith moment in the sliding time window is represented, M represents the length of the sliding time window, k represents the index of the current moment; Representing an optimal estimated value of the respiratory flow component at the ith moment in the sliding time window; Representing a minimum value that prevents denominator from being zero; Representing the pressure heart beat energy ratio; representing an optimal estimate of the pressure heart beat component at the i-th moment in the sliding time window; representing an optimal estimate of the respiratory pressure component at the i-th instant in the sliding time window; a score representing a flow rate change pattern; 、 、...、 The respiratory flow rate change rate optimal estimated values at the current time, the previous time, the first L times are respectively represented, wherein L represents the number of respiratory flow rate change rates participating in coherence calculation; representing flow-pressure heart beat component coherence; representing an average value of the optimal estimated value of the flow heart beat component in the sliding time window; Representing an average value of the optimal estimate of the pressure heart beat component within the sliding time window; representing the relative amplitude of the heart beat flow component; Representing the relative amplitude of the heart beat pressure component; Representing the optimal estimated value of the flow heart beat component at the current moment k; Representing an optimal estimated value of the pressure heart beat component at the current moment k; Representing an optimal estimated value of the respiratory flow component at the current moment k; Representing the optimal estimate of the respiratory pressure component at the current instant k.
  6. 6. The method for preventing false triggering of a ventilator based on adaptive kalman filtering according to any one of claims 1-5, wherein the method further comprises a physiological monitoring step, specifically comprising: and calculating the real-time heartbeat frequency and/or the heartbeat intensity of the patient based on the respiratory flow component, the respiratory pressure component, the flow heartbeat component and the optimal estimated value of the pressure heartbeat component.
  7. 7. The adaptive kalman filter based ventilator false triggering prevention method of claim 6, wherein calculating the heartbeat frequency comprises: Band-pass filtering the optimal estimated value of the flow heart beat component; The peaks of the filtered signal are detected and the heart beat frequency is calculated from the time intervals between successive peaks.
  8. 8. The adaptive kalman filter based ventilator false triggering prevention method of claim 6, wherein calculating the heart beat intensity comprises: calculating a heart beat energy envelope according to the optimal estimated values of the flow heart beat component and the pressure heart beat component; Calculating breathing background energy according to the breathing flow component, the breathing pressure component, the flow heart beat component and the optimal estimated value of the pressure heart beat component; the heart beat intensity is calculated according to the heart beat energy envelope and respiratory background energy.
  9. 9. A ventilator device comprising a memory, a processor and a computer program or instructions stored on the memory, wherein the processor executes the computer program or instructions to implement the adaptive kalman filter based ventilator false triggering prevention method of any of claims 1-8.
  10. 10. A computer readable storage medium having stored thereon a computer program or instructions which, when executed by a processor, implement the method for preventing false triggering of a ventilator based on adaptive kalman filtering as defined in any one of claims 1 to 8.

Description

Breathing machine false triggering prevention method, equipment and medium based on self-adaptive Kalman filtering Technical Field The invention belongs to the technical field of medical equipment, and particularly relates to a breathing machine false triggering prevention method, equipment and medium based on self-adaptive Kalman filtering. Background Ventilators typically rely on monitoring the patient's airflow signal in real-time to identify their spontaneous respiratory effort in a pressure-supported ventilation mode. When the airflow value at the end of expiration reaches a preset trigger threshold, the breathing machine automatically switches from the expiration phase to the inspiration phase so as to realize man-machine synchronization. However, in clinical practice, it has been found that the heart beat of a patient periodically compresses the chest and pulmonary vessels, producing a small periodic airflow fluctuation that is independent of spontaneous breathing, known as cardiogenic oscillations. This oscillation is particularly pronounced at the end of expiration, as its amplitude may reach or even exceed the trigger threshold of the ventilator, resulting in false triggering of the ventilator. Frequent false triggering not only can interfere with the normal breathing rhythm of a patient and increase breathing work, but also can cause man-machine countermeasure and abnormal fluctuation of pressure in the lung, thereby increasing the risk of lung injury and reducing the ventilation treatment effect. In order to reduce the occurrence of false triggering, various triggering judgment optimization schemes based on airflow signal analysis have been proposed in the prior art. For example, patent document CN115970109a discloses a respiratory ventilation prediction preprocessing method, which predicts the respiratory state (inspiration or expiration) of the next step by jointly judging the continuously sampled flow slope (K) and the real-time calculated tidal volume (e.g. TV1, TV 2) in the end expiration or end inspiration "hold phase" (i.e. the mean value of the air flow is in a stable interval close to zero), and adjusts the respiratory control parameters in advance according to the result. Specifically, at the end of expiration, if a flow slope K greater than a threshold K3 and an expiration tidal volume TV2 greater than a certain proportion of the predicted tidal volume is detected, the patient is predicted to begin inhaling and the control parameters are adjusted accordingly. However, this solution still has significant limitations in practical applications. The trigger judgment logic is essentially dependent on the combination condition of the flow change trend (slope) and the tidal volume accumulation value. In a clinical scenario, particularly at the end of expiration, the patient may have had a tidal volume accumulation that meets the trigger conditions due to spontaneous respiratory effort, and the airflow fluctuations resulting from the heart beat may likewise form a signal pattern that meets the slope threshold requirement. Since cardiogenic oscillations and real respiratory effort may have similar short-term varying characteristics on flow waveforms, and existing solutions do not physiologically distinguish the sources of the signals, it is difficult to effectively identify cardiogenic components in the airflow signal based only on a combined criterion of flow slope and tidal volume. Therefore, in the case of significant cardiac oscillations, the scheme still has high false triggering risk, and the false triggering problem caused by cardiac airflow fluctuation cannot be fundamentally solved. In summary, the prior art lacks an effective method for separating and identifying cardiac oscillation components from respiratory airflow signals accurately in real time, which results in a ventilator that is susceptible to cardiac pulsation interference in a pressure support ventilation mode and has insufficient triggering accuracy. Therefore, a technical solution that can decompose and identify the signal source is urgently needed to improve the anti-interference capability and clinical reliability of the ventilator trigger mechanism. Disclosure of Invention Aiming at the defects in the prior art, the invention aims to provide a breathing machine false triggering prevention method, equipment and medium based on self-adaptive Kalman filtering, so as to solve the technical problems that the traditional triggering judgment method cannot be effectively distinguished and false triggering frequently occurs and further the breathing rhythm of a patient is disturbed and the breathing work and the lung injury risk are increased because cardiac airflow fluctuation (cardiac oscillation) and spontaneous breathing effort of the patient are similar to each other on flow signals under the pressure support ventilation mode of the traditional breathing machine. The invention solves the technical problems by adopting the