An oxygen generator control system and method for the department of respiratory medicine

Abstract

The present invention relates to the technical field of oxygen generator control, and particularly relates to an oxygen generator regulation system and method for the department of respiratory medicine, including: obtaining the degree of abnormal patient status of each unstable sequence segment in the respiratory frequency time series sequence of a patient, combining the oxygen saturation time series sequence and the respiratory depth time series sequence, and then obtaining the possibility that each unstable sequence segment is noise, decomposing each unstable sequence segment, so as to obtain the final abnormal degree of the patient status of each unstable sequence segment, thereby obtaining prediction data, inputting the prediction data into a PID controller, and outputting an oxygen generator regulation instruction at the current moment. By assigning different prediction weights to the noise data, stable data, and unstable data in the time series sequence, the present invention reduces the influence of the noise data, improves the importance of the unstable data when the patient has an abnormality, thereby ensuring the accuracy of the prediction data and improving the accuracy of the dynamic regulation of the oxygen generator.

Description

Technical Field

[0001] The present invention relates to the technical field of oxygen generator control, and particularly relates to an oxygen generator regulation system and method for the department of respiratory medicine. Background Art

[0002] An oxygen generator is a medical device. Its application in the department of respiratory medicine is mainly to provide continuous oxygen supply to help patients with respiratory dysfunction or hypoxia symptoms maintain normal blood oxygen levels. It has a wide range of applications in medical treatment, first aid, and home care. The dynamic regulation of the oxygen flow of the oxygen generator is to adjust the oxygen output according to the patient's real-time oxygen demand to ensure that the patient obtains sufficient oxygen without wasting resources, effectively improving the effect of oxygen therapy and the patient's comfort.

[0003] Predicting the patient's oxygen demand in the next period of time is the basis of dynamic regulation. Generally, predictive regulation is achieved by real-time monitoring of physiological parameters such as the patient's respiratory rate, depth, and oxygen saturation. Existing problems: The monitoring data of the sensors worn on the patient's body will inevitably be affected by the patient's movement, and there may be certain measurement errors in the sensors themselves and electromagnetic interference and mechanical vibration that may exist in the working environment of the oxygen generator, resulting in abnormal noise data fluctuations in the monitoring data. When the patient is in an abnormal state, it will also cause fluctuations in the monitoring data, such as emotional agitation, anxiety, or sudden pain stimuli. There is a situation where the data fluctuations caused by abnormal states are misjudged as noise data fluctuations, which will affect the authenticity of the monitoring data collection, and then affect the accuracy of the data prediction achieved through the monitoring data. The inaccurate monitoring data prediction results may lead to low accuracy of the oxygen generator control, that is, inaccurate dynamic regulation of the oxygen generator flow. Summary of the Invention

[0004] The present invention provides an oxygen generator regulation system and method for the department of respiratory medicine to solve the existing problems.

[0005] The oxygen generator regulation system and method of the present invention adopt the following technical solutions:

[0006] An embodiment of the present invention provides an oxygen generator regulation device for the department of respiratory medicine. The device includes a memory and a processor. The processor is used to process the instructions stored in the memory to implement the following process:

[0007] Obtain the oxygen saturation time series, respiratory rate time series, and respiratory depth time series of the patient;

[0008] Divide the respiratory rate time series into several stable sequence segments and unstable sequence segments, and obtain the initial abnormal degree of each unstable sequence segment according to the difference in respiratory rate between the stable sequence segments and the unstable sequence segments;

[0009] Based on the initial degree of abnormality of each unstable sequence segment in the respiratory rate time series, as well as the oxygen saturation time series and the respiratory depth time series, obtain the possibility that each unstable sequence segment in the respiratory rate time series is noise;

[0010] Decompose each unstable sequence segment in the respiratory rate time series to obtain a trend term; based on the peak value in the trend term, the difference value of the data in the trend term, the possibility that each unstable sequence segment in the respiratory rate time series is noise, and the initial degree of abnormality, obtain the true degree of abnormality of each unstable sequence segment in the respiratory rate time series;

[0011] Based on the true degree of abnormality of each unstable sequence segment in the respiratory rate time series, obtain prediction data; input the prediction data into a PID controller to output an oxygen generator flow regulation instruction at the current moment.

[0012] Furthermore, the division of the respiratory rate time series into several stable sequence segments and unstable sequence segments includes the following specific steps:

[0013] In the respiratory rate time series, take the normalized value of the absolute value of the difference between the th and the th respiratory rates, and denote it as the instability of the th respiratory rate;

[0014] Denote the respiratory rates with instability less than a preset first threshold as stable respiratory rates, and denote the respiratory rates with instability greater than or equal to the preset first threshold as unstable respiratory rates;

[0015] Form stable sequence segments by adjacent stable respiratory rates, and form unstable sequence segments by adjacent unstable respiratory rates.

[0016] Furthermore, the obtaining of the initial degree of abnormality of each unstable sequence segment according to the differences in respiratory rates in the stable sequence segments and unstable sequence segments includes the following specific steps:

[0017] In the respiratory rate time series, take the normalized value of the difference between the mean value of all respiratory rates within the th unstable sequence segment and the mean value of all respiratory rates within all stable sequence segments adjacent to the th unstable sequence segment, and denote it as the relative difference of the th unstable sequence segment;

[0018] Based on the relative difference, duration, and differences in respiratory rates of the th unstable sequence segment, obtain the initial degree of abnormality of the th unstable sequence segment.

[0019] Further, based on the relative difference, duration, and the difference in respiratory rate between the th unstable sequence segment, the initial degree of abnormality of the th unstable sequence segment is obtained, and the specific steps are as follows:

[0020] Subtract the minimum value from the maximum value among all respiratory rates within the th unstable sequence segment, and multiply the result by the variance of all respiratory rates within the th unstable sequence segment. Denote the product as the instability degree of the th unstable sequence segment;

[0021] Multiply the instability degree of the th unstable sequence segment, the relative difference of the th unstable sequence segment, and the duration of the th unstable sequence segment. Denote the product as the initial degree of abnormality of the th unstable sequence segment.

[0022] Further, based on the initial degree of abnormality of each unstable sequence segment in the respiratory rate time series, and the oxygen saturation time series and the respiratory depth time series, the probability that each unstable sequence segment in the respiratory rate time series is noise is obtained, and the specific steps are as follows:

[0023] Use the start and end times of each unstable sequence segment in the respiratory rate time series to divide the oxygen saturation time series and the respiratory depth time series into several unstable sequence segments respectively;

[0024] According to the method for obtaining the relative difference of each unstable sequence segment in the respiratory rate time series, obtain the relative difference of each unstable sequence segment in the oxygen saturation time series and the respiratory depth time series respectively;

[0025] Based on the difference between the relative differences of each unstable sequence segment in the respiratory rate time series, the oxygen saturation time series, and the respiratory depth time series, obtain the probability that each unstable sequence segment in the respiratory rate time series is noise.

[0026] Further, based on the difference between the relative differences of each unstable sequence segment in the respiratory rate time series, the oxygen saturation time series, and the respiratory depth time series, the probability that each unstable sequence segment in the respiratory rate time series is noise is obtained, and the specific steps are as follows:

[0027] Subtract the The absolute value of the difference between the relative differences of the unstable sequence segments, plus the sum of the absolute values of the differences between the relative differences of the respiratory rate time series and the oxygen saturation time series at the th unstable sequence segment, is denoted as the first sum value;

[0028] The DTW distance between the respiratory rate time series and the respiratory depth time series at the th unstable sequence segment, plus the sum of the DTW distances between the respiratory rate time series and the oxygen saturation time series at the th unstable sequence segment, is denoted as the second sum value;

[0029] The product of the first sum value and the second sum value is denoted as the probability that the th unstable sequence segment in the respiratory rate time series is noise.

[0030] Furthermore, obtaining the true anomaly degree of each unstable sequence segment in the respiratory rate time series according to the peak value in the trend item, the difference value of the data in the trend item, the probability that each unstable sequence segment in the respiratory rate time series is noise, and the initial anomaly degree includes the following specific steps:

[0031] In the trend item of the th unstable sequence segment in the respiratory rate time series, the sequence composed of the maximum value in the trend item and all the data before the maximum value is denoted as the target sequence, and the sequence composed of the maximum value in the trend item and all the data after the maximum value is denoted as the reference sequence;

[0032] The ratio of the number of negative numbers in the first-order difference sequence of the target sequence to the number of data in the target sequence is denoted as the first ratio;

[0033] The ratio of the number of positive numbers in the first-order difference sequence of the reference sequence to the number of data in the reference sequence is denoted as the second ratio;

[0034] The product of the mean of the first ratio and the second ratio and the number of peak values in the trend item is denoted as the th unstable sequence segment's noise influence credibility;

[0035] According to the th unstable sequence segment's noise influence credibility, the probability of being noise, and the initial anomaly degree, obtain the true anomaly degree of each unstable sequence segment in the respiratory rate time series.

[0036] Furthermore, obtaining the true anomaly degree of each unstable sequence segment in the respiratory rate time series according to the th unstable sequence segment's noise influence credibility, the probability of being noise, and the initial anomaly degree includes the following specific steps:

[0037] Calculate the inverse proportional normalization value of the product of the credibility of the noise influence of the th unstable sequence segment and the probability that the th unstable sequence segment is noise. Denote the normalization value of the product of the inverse proportional normalization value and the initial abnormal degree of the th unstable sequence segment as the true abnormal degree of the th unstable sequence segment in the respiratory rate time series.

[0038] Furthermore, obtaining prediction data according to the true abnormal degree of each unstable sequence segment in the respiratory rate time series includes the following specific steps:

[0039] Denote the sum value of the true abnormal degree of the th unstable sequence segment in the respiratory rate time series and a preset second threshold as the weight of each respiratory rate within the th unstable sequence segment in the respiratory rate time series;

[0040] Set the weight of each respiratory rate within each stable sequence segment in the respiratory rate time series to a preset third threshold;

[0041] Perform data prediction on the respiratory rate time series using a weighted ARIMA model according to the weight of each respiratory rate in the respiratory rate time series to obtain prediction data;

[0042] Obtain the prediction data of the respiratory depth time series and the oxygen saturation time series respectively according to the acquisition method of the prediction data of the respiratory rate time series.

[0043] The beneficial effects of the technical solution of the present invention are:

[0044] In the embodiment of the present invention, the initial abnormal degree of each unstable sequence segment in the respiratory rate time series of a patient is obtained, thereby determining the initial abnormal degree of each unstable sequence segment, providing data support for the possibility that the unstable sequence segment is noise and the true abnormal degree in the subsequent stage. By combining the oxygen saturation time series and the respiratory depth time series, the possibility that each unstable sequence segment is noise is obtained, thereby analyzing the influence of noise, further ensuring the accuracy of detecting the abnormal state of the patient, and thus improving the accuracy of oxygen supply regulation. Each unstable sequence segment is decomposed to obtain the true abnormal degree of each unstable sequence segment, thereby obtaining prediction data, and thus obtaining the true abnormal degree without noise interference, ensuring the credibility of the prediction data, and thus improving the credibility of the oxygen generator flow regulation instruction. The prediction data is input into a PID controller, and an oxygen generator flow regulation instruction at the current moment is output. Thus, by assigning different prediction weights to the noise data, stable data, and unstable data in the time series, the present invention reduces the influence of noise data, that is, it can eliminate the influence of noise on the accuracy of unstable data caused by abnormal states, improves the importance of unstable data when the patient has an abnormality, thereby ensuring the accuracy of prediction data and improving the accuracy of dynamic regulation of the oxygen generator flow for respiratory medicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0046] Figure 1 It is a flowchart of the steps of a method for regulating an oxygen generator for respiratory medicine according to the present invention;

[0047] Figure 2 It is a schematic diagram of the global data trend change of the unstable sequence segment in the respiratory rate time series without noise influence. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific embodiments, structures, features, and effects of an oxygen generator regulation system and method for respiratory medicine according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs.

[0050] The following specifically describes the specific solutions of an oxygen generator regulation system and method for the department of respiratory medicine provided by the present invention in conjunction with the accompanying drawings.

[0051] An oxygen generator regulation device for the department of respiratory medicine provided by an embodiment of the present invention, the device includes a memory and a processor, and the processor is used to process instructions stored in the memory to implement the steps of an oxygen generator regulation method for the department of respiratory medicine.

[0052] Please refer to Figure 1 , which shows the step flow chart of an oxygen generator regulation method provided by an embodiment of the present invention. The method includes the following steps:

[0053] Step S001: Obtain the oxygen saturation time series, respiratory frequency time series, and respiratory depth time series of the patient.

[0054] The SpO2 sensor worn on the patient and the sensors for respiratory frequency and depth are respectively used to collect the oxygen saturation, respiratory frequency, and respiratory depth in real time, and the oxygen saturation time series, respiratory frequency time series, and respiratory depth time series are respectively obtained.

[0055] It should be noted that: The main respiratory parameters that need to be monitored for the dynamic regulation of the oxygen generator flow are the oxygen saturation, respiratory frequency, and respiratory depth. This is because when the amount of oxygen inhaled by the patient during breathing decreases, the oxygen saturation will decrease, and the oxygen generator needs to increase the output flow. Conversely, when the amount of oxygen inhaled by the patient during breathing increases, the oxygen saturation will increase, and the oxygen generator needs to reduce the output flow. Among them, the acquisition frequency of the three types of data is the same, once every 10 seconds, and the acquisition duration is from the start of wearing the oxygen generator to the current moment. This is used as an example for description. Since the patient needs to wear the oxygen generator, the respiratory changes caused by strenuous exercise are not considered in this embodiment.

[0056] Step S002: Divide the respiratory frequency time series into several stable sequence segments and unstable sequence segments, and obtain the initial abnormal degree of each unstable sequence segment according to the difference in respiratory frequency between the stable sequence segments and the unstable sequence segments.

[0057] It should be noted that: under normal circumstances, the oxygen saturation, respiratory rate, and respiratory depth of patients change little over time. This is mainly because the human body has a series of self-regulatory mechanisms to maintain physiological homeostasis and ensure the normal operation of the body's functions. However, when the respiratory rate time series is affected by noise or abnormal conditions, the respiratory rate of the patient will fluctuate, and the breathing often becomes irregular, manifested as a disorder of the respiratory rhythm or irregular changes in depth, resulting in unstable fluctuations in the data. To facilitate subsequent analysis of the possibility that each unstable sequence segment is noise and the true degree of abnormality, the data characteristics of the unstable sequence segments are analyzed by combining the differences in respiratory rates between the stable sequence segments and the unstable sequence segments to quantify the initial degree of abnormality of each unstable sequence segment.

[0058] Preferably, in an embodiment of the present invention, the method for obtaining the initial degree of abnormality of each unstable sequence segment in the respiratory rate time series includes:

[0059] In the respiratory rate time series, the normalization value of the absolute value of the difference between the th and the th respiratory rates is denoted as the instability of the th respiratory rate.

[0060] In the above manner, the instability of each respiratory rate is obtained.

[0061] It should be noted that: in this embodiment, the instability of the last respiratory rate is set to the instability of the penultimate respiratory rate, and this is used as an example for description. Among them, the minimum-maximum normalization method is used to normalize the absolute value of the difference between adjacent respiratory rates to between 0 and 1, which is a well-known technique.

[0062] A first threshold is preset to be 0.3, and this is used as an example for description.

[0063] In the respiratory rate time series, the smaller the instability, the greater the possibility that the corresponding respiratory rate is a stable respiratory rate, and the greater the instability, the greater the possibility that the corresponding respiratory rate is an unstable respiratory rate. Therefore, the respiratory rates with instability less than the preset first threshold are denoted as stable respiratory rates, and the respiratory rates with instability greater than or equal to the preset first threshold are denoted as unstable respiratory rates.

[0064] In the respiratory rate time series, the adjacent stable respiratory rates form stable sequence segments, and the adjacent unstable respiratory rates form unstable sequence segments. Thus, the respiratory rate time series is divided into several stable sequence segments and unstable sequence segments.

[0065] It should be noted that: if a certain stable breathing frequency has no adjacent stable breathing frequency, then this stable breathing frequency is defined as an unstable breathing frequency; if an unstable breathing frequency has no adjacent unstable breathing frequency, then this unstable breathing frequency is defined as a stable breathing frequency, so as to ensure that the length of the sequence segment is greater than 1. Among them, the stable sequence segment reflects the data when the patient's condition is stable, while the unstable sequence segment reflects the data when the patient's condition is unstable. The greater the difference between the unstable sequence segment and its adjacent stable sequence segment, the greater the change in the patient's breathing condition.

[0066] In the time series of breathing frequencies, subtract the mean value of all breathing frequencies within the th unstable sequence segment from the mean value of all breathing frequencies within all stable sequence segments adjacent to the th unstable sequence segment, and then take the normalized value of the difference, which is denoted as the relative difference of the th unstable sequence segment in the time series of breathing frequencies.

[0067] It should be noted that: the formula for calculating the relative difference of the th unstable sequence segment in the time series of breathing frequencies is:

[0068]

[0069] In the formula, is the relative difference of the th unstable sequence segment in the time series of breathing frequencies, is the mean value of all breathing frequencies within the th unstable sequence segment in the time series of breathing frequencies, is the mean value of all breathing frequencies within all stable sequence segments adjacent to the th unstable sequence segment in the time series of breathing frequencies. is a linear normalization function, which is used to normalize the data value to between 0 and 1. Normalization is to eliminate the dimension and facilitate the comparative analysis of data in different dimensions. When the breathing frequencies within the unstable sequence segment change more greatly and violently, and the duration of the unstable sequence segment is longer, it indicates that the patient needs to make a greater change in the breathing pattern and a longer time to restore normal breathing, which is more in line with the data fluctuation characteristics reflected by the real abnormal condition.

[0070] In the time series of breathing frequencies, multiply the difference between the maximum value and the minimum value of all breathing frequencies within the th unstable sequence segment by the variance of all breathing frequencies within the th unstable sequence segment, and the result is denoted as the instability degree of the th unstable sequence segment in the time series of breathing frequencies. For the The instability degree of the nth unstable sequence segment, the relative difference of the th unstable sequence segment, and the product of the duration of the th unstable sequence segment are denoted as the initial abnormality degree of the th unstable sequence segment in the respiratory rate time series. The relative difference of the th unstable sequence segment, and the product of the duration of the th unstable sequence segment are denoted as the initial abnormality degree of the th unstable sequence segment in the respiratory rate time series. Initial abnormality degree of the nth unstable sequence segment in the respiratory rate time series.

[0071] It should be noted that: The calculation formula for the initial abnormality degree of the th unstable sequence segment in the respiratory rate time series is: The calculation formula for the initial abnormality degree of the nth unstable sequence segment in the respiratory rate time series is:

[0072]

[0073] In the formula, is the initial abnormality degree of the th unstable sequence segment in the respiratory rate time series, is the initial abnormality degree of the nth unstable sequence segment in the respiratory rate time series, is the relative difference of the th unstable sequence segment in the respiratory rate time series, is the relative difference of the nth unstable sequence segment in the respiratory rate time series, is the duration of the th unstable sequence segment in the respiratory rate time series, is the duration of the nth unstable sequence segment in the respiratory rate time series, is the variance of all respiratory rates in the th unstable sequence segment in the respiratory rate time series, is the variance of all respiratory rates in the nth unstable sequence segment in the respiratory rate time series, and are the maximum and minimum values of all respiratory rates in the th unstable sequence segment in the respiratory rate time series, respectively. are the maximum and minimum values of all respiratory rates in the nth unstable sequence segment in the respiratory rate time series, respectively. is the instability degree of the th unstable sequence segment in the respiratory rate time series. The larger its value, the more irregular the breathing and the larger the range of respiratory rate changes. is the instability degree of the nth unstable sequence segment in the respiratory rate time series. The larger its value, the more irregular the breathing and the larger the range of respiratory rate changes.

[0074] Step S003: According to the initial abnormality degree of each unstable sequence segment in the respiratory rate time series, and the oxygen saturation time series and the respiratory depth time series, obtain the possibility that each unstable sequence segment in the respiratory rate time series is noise.

[0075] It should be noted that: Due to the monitoring data of the sensor worn on the patient's body, it will inevitably be affected by the patient's movement, and there may be certain measurement errors in the sensor itself and electromagnetic interference and mechanical vibration that may exist in the working environment of the oxygen generator, resulting in abnormal noise data fluctuations in the monitoring data, which will lead to a low credibility of the initially calculated abnormality degree. Therefore, it is necessary to distinguish the noise data fluctuations and reduce the influence of noise on the respiratory rate data.

[0076] It should be further noted that: in the time series of data collected by different types of sensors, the probability that noise data appears simultaneously in different types of sensors is relatively low, and the probability that the magnitudes of noise data are similar is also relatively low. This is because the occurrence time of noise data is random, and the magnitude is also random. When a patient experiences emotional agitation, anxiety, or sudden pain stimulation, it usually causes an increase in breathing frequency and depth to increase oxygen intake and improve blood oxygen saturation to relieve the discomfort felt by the body. When the patient's discomfort gradually subsides, the breathing frequency and depth will decrease, causing the blood oxygen saturation to gradually return to a relatively normal level. That is, when the patient's state changes, these three types of data change simultaneously, and the change trends are similar.

[0077] Preferably, in an embodiment of the present invention, the method for obtaining the probability that each unstable sequence segment in the breathing frequency time series is noise includes:

[0078] Using the start and end times of each stable sequence segment and each unstable sequence segment in the breathing frequency time series, the oxygen saturation time series and the breathing depth time series are respectively divided into a plurality of stable sequence segments and unstable sequence segments.

[0079] According to the method for obtaining the relative difference of each unstable sequence segment in the breathing frequency time series, the relative differences of each unstable sequence segment in the oxygen saturation time series and the breathing depth time series are respectively obtained.

[0080] The sum of the absolute values of the differences in the relative differences of the th unstable sequence segment in the breathing frequency time series and the breathing depth time series plus the sum of the absolute values of the differences in the relative differences of the th unstable sequence segment in the breathing frequency time series and the oxygen saturation time series is denoted as the first sum value. The sum of the DTW distances of the th unstable sequence segment in the breathing frequency time series and the breathing depth time series plus the sum of the DTW distances of the th unstable sequence segment in the breathing frequency time series and the oxygen saturation time series is denoted as the second sum value. The product of the first sum value and the second sum value is denoted as the probability that the th unstable sequence segment in the breathing frequency time series is noise.

[0081] It should be noted that: the calculation formula for the probability that the th unstable sequence segment in the breathing frequency time series is noise is:

[0082]

[0083] In the formula, is the The possibility that an unstable sequence segment is noise, is the absolute value of the difference between the relative differences of the th unstable sequence segment in the respiratory frequency time series and the respiratory depth time series, is the absolute value of the difference between the relative differences of the th unstable sequence segment in the respiratory frequency time series and the oxygen saturation time series, is the DTW distance of the th unstable sequence segment in the respiratory frequency time series and the respiratory depth time series, is the DTW distance of the th unstable sequence segment in the respiratory frequency time series and the oxygen saturation time series, is the first sum value, is the second sum value. Among them, the DTW distance is obtained by data matching of two unstable sequence segments according to the DTW algorithm. The DTW algorithm (Dynamic Time Warping) is a well-known technology. The greater the DTW distance, the less similar the data change trends of the two unstable sequence segments are, that is, the more likely they are to be noise data. And and the greater they are, the more it indicates that there is no simultaneous data fluctuation in the three types of data, that is, the more likely they are to be noise data.

[0084] Step S004: Decompose each unstable sequence segment in the respiratory frequency time series to obtain a trend term; according to the peak value in the trend term, the difference value of the data in the trend term, the possibility that each unstable sequence segment in the respiratory frequency time series is noise, and the initial abnormal degree, obtain the true abnormal degree of each unstable sequence segment in the respiratory frequency time series.

[0085] It should be noted that: The unstable data generated under the influence of noise is not the patient's true respiratory rate. Using the respiratory rate at this time for data prediction to control the oxygen generator will result in low accuracy of oxygen generator control; while the fluctuations in monitoring data caused by abnormal states can show the patient's true respiratory state, which is the basis for dynamically adjusting the oxygen flow of the oxygen generator. According to the possibility of noise, the noise data when the patient is in a stable state can be accurately distinguished. However, when the noise data is in the period of the patient's unstable state, it will cause the absolute value of the relative difference of the unstable sequence segments of different-dimensional data to be relatively large when the patient has emotional agitation, anxiety, or sudden pain stimulation, and the DTW distance of the unstable sequence segments of different-dimensional data is relatively large. Therefore, it is necessary to further analyze the influence of noise in the period of the patient's unstable state. During the process from the patient's emotional agitation, anxiety, or sudden pain stimulation to gradual relief, the respiratory rate, depth, and oxygen saturation will first increase suddenly and then gradually decrease until they return to normal. A schematic diagram of the global data trend change of the unstable sequence segment in the respiratory rate time series without noise influence is shown in Figure 2 as shown in Figure 2 where the horizontal axis is time and the vertical axis is the respiratory rate.

[0086] Here, the greater the true abnormal degree, the smaller the influence of the corresponding unstable sequence segment by noise fluctuations, the greater the possibility that the respiratory fluctuation is the patient's true respiratory state, and the higher the importance of each respiratory rate in the corresponding unstable sequence segment for subsequent data prediction; the smaller the true abnormal degree, the greater the influence of the corresponding unstable sequence segment by noise fluctuations, the smaller the possibility that the respiratory fluctuation is the patient's true respiratory state, and the lower the importance of each respiratory rate in the corresponding unstable sequence segment for subsequent data prediction.

[0087] Preferably, in an embodiment of the present invention, the method for obtaining the true abnormal degree of each unstable sequence segment in the respiratory rate time series includes:

[0088] Perform STL decomposition on the th unstable sequence segment in the respiratory rate time series to obtain the trend term.

[0089] Use the peak detection algorithm to obtain several peaks in the trend term.

[0090] It should be noted that: Both the peak detection algorithm and STL decomposition are well-known technologies, and the specific methods are not introduced here. Among them, the full Chinese name of STL decomposition is Seasonal-Trend decomposition procedure, and the full English name is Seasonal-Trenddecomposition procedure based on Loess, which can effectively extract the long-term trend from time series data.

[0091] In the trend term of the th unstable sequence segment in the respiratory rate time series, the sequence formed by the maximum value in the trend term and all the data before the maximum value is denoted as the target sequence, and the sequence formed by the maximum value in the trend term and all the data after the maximum value is denoted as the reference sequence.

[0092] It should be noted that: if there are multiple maximum values in the trend term, the first-occurring maximum value is taken.

[0093] In the trend term of the th unstable sequence segment in the respiratory rate time series, the ratio of the number of negative numbers in the first-order difference sequence of the target sequence to the number of data in the target sequence is denoted as the first ratio, and the ratio of the number of positive numbers in the first-order difference sequence of the reference sequence to the number of data in the reference sequence is denoted as the second ratio. The product of the mean of the first ratio and the second ratio and the number of peak values in the trend term is denoted as the noise influence credibility of the th unstable sequence segment in the respiratory rate time series.

[0094] In the respiratory rate time series, calculate the inverse proportional normalization value of the product of the noise influence credibility of the th unstable sequence segment and the probability that the th unstable sequence segment is noise. The normalization value of the product of this inverse proportional normalization value and the initial anomaly degree of the th unstable sequence segment is denoted as the true anomaly degree of the th unstable sequence segment in the respiratory rate time series.

[0095] It should be noted that: the calculation formula for the true anomaly degree of the th unstable sequence segment in the respiratory rate time series is:

[0096]

[0097] In the formula, is the true anomaly degree of the th unstable sequence segment in the respiratory rate time series, is the number of peak values in the trend term of the th unstable sequence segment in the respiratory rate time series, is the number of data in the target sequence in the trend term of the th unstable sequence segment in the respiratory rate time series, is the number of negative numbers in the first-order difference sequence of the target sequence in the trend term of the th unstable sequence segment in the respiratory rate time series, is the number of the The number of data in the reference sequence in the trend term of an unstable sequence segment is the number of positive numbers in the first-order difference sequence of the reference sequence in the trend term of the th unstable sequence segment in the respiratory rate time series, is the probability that the th unstable sequence segment in the respiratory rate time series is noise, is the initial abnormal degree of the th unstable sequence segment in the respiratory rate time series. is the first ratio, is the second ratio. represents the credibility of the noise influence of the th unstable sequence segment in the respiratory rate time series. is the linear normalization function used to normalize the data value between 0 and 1. is the exponential function with the natural constant as the base. In this embodiment,

[0098] is used to present the inverse proportional relationship and normalization process. The implementer can set the inverse proportional function and normalization function according to the actual situation. and The larger they are, the more seriously the unstable sequence segment is affected by noise. Therefore, is used to correct . And the greater the noise influence, the less accurate is. Therefore, is used to correct to obtain the true abnormal degree.

[0099] Step S005: Obtain prediction data according to the true abnormal degree of each unstable sequence segment in the respiratory rate time series; input the prediction data into the PID controller to output an oxygen generator flow regulation instruction at the current moment.

[0100] Preset the second threshold as 0.5 and the third threshold as 1, and describe it by taking this as an example.

[0101] The sum of the true abnormal degree of the th unstable sequence segment in the respiratory rate time series and the preset second threshold is denoted as the weight of each respiratory rate in the th unstable sequence segment in the respiratory rate time series.

[0102] In the above manner, the weight of each respiratory rate in each unstable sequence segment of the respiratory rate time series is obtained.

[0103] Set the weight of each respiratory rate in each stable sequence segment of the respiratory rate time series to a preset third threshold.

[0104] Thus, the weight of each respiratory rate in the respiratory rate time series is obtained.

[0105] According to the weight of each respiratory rate in the respiratory rate time series, use the weighted ARIMA model to predict the data of the respiratory rate time series and obtain the predicted data.

[0106] It should be noted that: The weighted ARIMA model (Weighted Autoregressive Integrated Moving Average) is a well-known technology, and the specific method will not be introduced here. In this embodiment, a smaller weight is given to the data affected by noise, and a larger weight is given to the monitoring data when the patient's state is abnormal. Thus, during the process of obtaining the predicted data, the influence of noise is reduced, and more consideration is given to the important monitoring data when the patient's state is abnormal, ensuring the accuracy of the predicted data.

[0107] According to the acquisition method of the predicted data of the respiratory rate time series, the predicted data of the respiratory depth time series and the oxygen saturation time series are obtained respectively.

[0108] Input the respiratory rate time series, the respiratory depth time series, the oxygen saturation time series, and the predicted data of the respiratory rate time series, the respiratory depth time series, and the oxygen saturation time series into the PID controller, and output a control instruction for the oxygen generator flow rate at the current moment.

[0109] Use the control instruction for the oxygen generator flow rate at the current moment to regulate the oxygen generator flow rate at the next moment of the current moment.

[0110] It should be noted that: Using a PID (Proportional-Integral-Derivative) controller for the dynamic regulation of the oxygen generator flow rate is a common and effective method. The PID controller is a well-known control technology. Thus, in the above manner, the control instruction for the oxygen generator flow rate at each subsequent moment is obtained, such as increasing the oxygen supply, decreasing the oxygen supply, and keeping the oxygen supply unchanged, thereby completing the dynamic regulation of the oxygen generator flow rate.

[0111] In summary, in the embodiment of the present invention, the oxygen saturation time series, the respiratory rate time series, and the respiratory depth time series of the patient are obtained, so as to obtain the initial abnormal degree of each unstable sequence segment in the respiratory rate time series. By combining the oxygen saturation time series and the respiratory depth time series, the possibility that each unstable sequence segment in the respiratory rate time series is noise is obtained. Each unstable sequence segment in the respiratory rate time series is decomposed, so as to obtain the true abnormal degree of each unstable sequence segment in the respiratory rate time series. According to the true abnormal degree of each unstable sequence segment in the respiratory rate time series, prediction data is obtained, and the prediction data is input into the PID controller to output an oxygen generator flow regulation instruction at the current moment. By assigning different prediction weights to the noise data, stable data, and unstable data in the time series, the present invention reduces the influence of the noise data, improves the importance of the unstable data when the patient has an abnormality, thereby ensuring the accuracy of the prediction data and improving the accuracy of the dynamic regulation of the oxygen generator flow.

[0112] The present invention also provides an oxygen generator regulation system for the department of respiratory medicine, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the foregoing oxygen generator regulation method for the department of respiratory medicine.

[0113] The foregoing is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An oxygen generator control device for the department of respiratory medicine, characterized in that, The device includes a memory and a processor, and the processor is configured to process instructions stored in the memory to implement the following process: Obtain the time series sequence of the patient's oxygen saturation, the time series sequence of the respiratory frequency, and the time series sequence of the respiratory depth; Divide the time series sequence of the respiratory frequency into a number of stable sequence segments and unstable sequence segments, and obtain the initial degree of abnormality of each unstable sequence segment according to the difference in the respiratory frequency between the stable sequence segments and the unstable sequence segments; According to the initial degree of abnormality of each unstable sequence segment in the time series sequence of the respiratory frequency, as well as the time series sequence of the oxygen saturation and the time series sequence of the respiratory depth, obtain the possibility that each unstable sequence segment in the time series sequence of the respiratory frequency is noise; Decompose each unstable sequence segment in the time series sequence of the respiratory frequency to obtain a trend term; According to the peak value in the trend term, the difference value of the data in the trend term, the possibility that each unstable sequence segment in the time series sequence of the respiratory frequency is noise, and the initial degree of abnormality, obtain the true degree of abnormality of each unstable sequence segment in the time series sequence of the respiratory frequency; According to the true degree of abnormality of each unstable sequence segment in the time series sequence of the respiratory frequency, obtain prediction data; input the prediction data into a PID controller, and output an oxygen generator flow regulation instruction at the current moment; The step of dividing the time series sequence of the respiratory frequency into a number of stable sequence segments and unstable sequence segments includes the following specific steps: In the respiratory rate time series, the normalized value of the absolute value of the difference between the -th and the -th respiratory rates is denoted as the instability of the -th respiratory rate; Denote the respiratory frequency with instability less than a preset first threshold as a stable respiratory frequency, and denote the respiratory frequency with instability greater than or equal to the preset first threshold as an unstable respiratory frequency; Form stable sequence segments with adjacent stable respiratory frequencies, and form unstable sequence segments with adjacent unstable respiratory frequencies; According to the peak value in the trend term, the difference value of the data in the trend term, the possibility that each unstable sequence segment in the time series sequence of the respiratory frequency is noise, and the initial degree of abnormality, obtain the true degree of abnormality of each unstable sequence segment in the time series sequence of the respiratory frequency, and the specific steps include: In the trend term of the th unstable sequence segment in the respiratory rate time series, the sequence formed by the maximum value in the trend term and all the data before the maximum value is denoted as the target sequence, and the sequence formed by the maximum value in the trend term and all the data after the maximum value is denoted as the reference sequence; Denote the ratio of the number of negative numbers in the first-order difference sequence of the target sequence to the number of data in the target sequence as a first ratio; Denote the ratio of the number of positive numbers in the first-order difference sequence of the reference sequence to the number of data in the reference sequence as a second ratio; Denote the product of the mean of the first ratio and the second ratio and the number of peaks in the trend term as the noise influence credibility of the th unstable sequence segment; According to the noise influence credibility, the possibility of noise, and the initial abnormal degree of the th unstable sequence segment, the true abnormal degree of each unstable sequence segment in the respiratory rate time series is obtained; The step of obtaining prediction data according to the true degree of abnormality of each unstable sequence segment in the time series sequence of the respiratory frequency includes the following specific steps: The sum of the true anomaly degree of the th unstable sequence segment in the respiratory rate time series and the preset second threshold is denoted as the weight of each respiratory rate within the th unstable sequence segment in the respiratory rate time series; Set the weight of each respiratory frequency in each stable sequence segment in the time series sequence of the respiratory frequency to a preset third threshold; According to the weights of each respiratory frequency in the time series sequence of the respiratory frequency, use a weighted ARIMA model to perform data prediction on the time series sequence of the respiratory frequency to obtain prediction data; According to the acquisition method of the prediction data of the time series sequence of the respiratory frequency, obtain the prediction data of the time series sequence of the respiratory depth and the time series sequence of the oxygen saturation respectively.

2. The oxygen generator control device for the department of respiratory medicine according to claim 1, wherein, The step of obtaining the initial degree of abnormality of each unstable sequence segment according to the difference in the respiratory frequency between the stable sequence segments and the unstable sequence segments includes the following specific steps: In the respiratory rate time series, the normalized value of the difference obtained by subtracting the mean value of all respiratory rates within the th unstable sequence segment from the mean value of all respiratory rates within all adjacent stable sequence segments of the th unstable sequence segment is denoted as the relative difference of the th unstable sequence segment; According to the relative difference, duration, and the difference in respiratory rate of the th unstable sequence segment, the initial abnormal degree of the th unstable sequence segment is obtained.

3. The oxygen generator control device for the department of respiratory medicine according to claim 2, characterized in that, The relative difference, duration, and difference in respiratory rate of the th unstable sequence segment are used to obtain the initial abnormality degree of the th unstable sequence segment. The specific steps are as follows: Subtract the difference between the maximum and minimum values of all respiratory frequencies within the th unstable sequence segment from the variance of all respiratory frequencies within the th unstable sequence segment, and denote the product as the instability degree of the th unstable sequence segment; Multiply the instability degree of the th unstable sequence segment, the relative difference of the th unstable sequence segment, and the duration of the th unstable sequence segment, and denote the result as the initial anomaly degree of the th unstable sequence segment.

4. The oxygen generator control device for the department of respiratory medicine according to claim 2, characterized in that, Obtaining the possibility that each unstable sequence segment in the respiratory rate time series is noise based on the initial abnormal degree of each unstable sequence segment in the respiratory rate time series and the oxygen saturation time series and the respiratory depth time series, including the following specific steps: Using the start and end times of each unstable sequence segment in the respiratory rate time series, respectively divide several unstable sequence segments in the oxygen saturation time series and the respiratory depth time series; According to the obtaining method of the relative difference of each unstable sequence segment in the respiratory rate time series, respectively obtain the relative difference of each unstable sequence segment in the oxygen saturation time series and the respiratory depth time series; According to the difference between the relative differences of each unstable sequence segment in the respiratory rate time series, the oxygen saturation time series and the respiratory depth time series, obtain the possibility that each unstable sequence segment in the respiratory rate time series is noise.

5. The oxygen generator control device for the department of respiratory medicine according to claim 4, wherein, The obtaining the possibility that each unstable sequence segment in the respiratory rate time series is noise according to the difference between the relative differences of each unstable sequence segment in the respiratory rate time series, the oxygen saturation time series and the respiratory depth time series, including the following specific steps: Add the absolute value of the difference in the relative difference of the th unstable sequence segment in the respiratory rate time series and the respiratory depth time series to the absolute value of the difference in the relative difference of the th unstable sequence segment in the respiratory rate time series and the oxygen saturation time series, and denote the sum value as the first sum value; Add the DTW distance of the th unstable sequence segment in the respiratory frequency time series and the respiratory depth time series to the DTW distance of the th unstable sequence segment in the respiratory frequency time series and the oxygen saturation time series, and denote the sum value as the second sum value; Denote the product of the first sum value and the second sum value as the possibility that the th unstable sequence segment in the respiratory rate time series is noise.

6. The oxygen generator control device for the department of respiratory medicine according to claim 1, characterized in that, The credibility of the noise influence, the probability of noise, and the initial abnormal degree according to the th unstable sequence segment are used to obtain the true abnormal degree of each unstable sequence segment in the respiratory rate time series. The specific steps are as follows: Calculate the inverse proportional normalization value of the product of the noise influence credibility of the th unstable sequence segment and the probability that the th unstable sequence segment is noise, and denote the normalization value of the product of the inverse proportional normalization value and the initial anomaly degree of the th unstable sequence segment as the true anomaly degree of the th unstable sequence segment in the respiratory rate time series.

Images