Portable oxygen supply device for the elderly and flow intelligent adjustment system

Abstract

The present application relates to oxygen supply technical field, specifically to a portable oxygen supply device for the elderly and a flow intelligent adjustment system, the flow intelligent adjustment system realizes the following flow intelligent adjustment method when executing: determining the unstable moment of breathing frequency, obtaining the filtering characteristics of each moment according to the breathing frequency change in the local time range of the unstable moment, filtering the breathing frequency to obtain the updated breathing frequency; according to the size relationship between the updated breathing frequency and the breathing depth of each moment and the corresponding preset baseline, determine the activity moment in the reference period of the current moment; according to the correlation between the activity period and the adjacent non-activity period after it in time sequence, determine the rest effect of the user, combine the activity intensity of the user in the activity period, obtain the accurate oxygen supply demand of the elderly, improve the accuracy of oxygen supply demand acquisition, and then the oxygen supply flow of the elderly can be accurately and reliably adjusted according to the oxygen supply demand.

Description

Technical Field

[0001] This invention relates to the field of oxygen supply technology, specifically to a portable oxygen supply device for the elderly and an intelligent flow regulation system. Background Technology

[0002] Portable oxygen supply devices for the elderly are miniaturized, portable, and independently powered oxygen supply equipment. Unlike traditional fixed home oxygen concentrators, they are characterized by their small size, portability, and independent power supply. They can meet the continuous oxygen needs of the elderly when they go out, travel, or move around at home. They are often equipped with intelligent control algorithms to achieve adaptive adjustment of oxygen supply flow, which can not only meet the oxygen needs of the elderly at home, but also provide a continuous oxygen supply when they are outdoors.

[0003] The existing intelligent flow regulation system for oxygen supply devices operates on the principle of determining the current oxygen demand of the elderly person and then adjusting the oxygen output flow rate accordingly. Currently, the current oxygen demand is determined solely based on the elderly person's respiratory rate. However, the respiratory rate of the elderly is inherently inconsistent with normal rates; for example, breathing is often shallow and rapid, rhythmic, and easily affected by activity. Directly relying on the collected respiratory rate to determine the oxygen demand will affect the accuracy of the demand assessment, thus impacting the accuracy of oxygen flow rate regulation. Summary of the Invention

[0004] To address the technical problem of low accuracy in obtaining existing oxygen demand, which affects the accuracy of oxygen flow rate regulation, the present invention aims to provide a portable oxygen supply device for the elderly and an intelligent flow rate regulation system. The specific technical solution adopted is as follows: In a first aspect of the present invention, a flow intelligent adjustment system for a portable oxygen supply device for the elderly is provided, comprising a memory and a processor; the memory is connected to the processor; the memory is used to store program instructions; the processor is used to implement the following flow intelligent adjustment method when the program instructions are executed: Determine the moment of instability based on the instability of the user's respiratory rate; Based on the changes in respiratory rate within a local time range during unstable moments, and the instability of respiratory rate during unstable moments, the filtering characteristics of each moment are obtained to filter the respiratory rate time series data and obtain updated respiratory rate time series data. Based on the relationship between the updated respiratory rate and respiratory depth at each moment and the corresponding preset baseline, the activity time within the reference time period at the current moment is determined; The rest effect of a user after an active period is determined based on the correlation between the active period and the adjacent inactive period that follows; the active period consists of at least two consecutive active moments in time. The activity intensity during the activity period is determined, and the rest effect is combined to obtain the user's oxygen supply demand at the current moment; the oxygen supply demand is used to indicate the adjustment of the user's oxygen supply flow.

[0005] In an exemplary embodiment, the process of obtaining the unstable moment includes: Determine the degree of fluctuation in respiratory rate at each moment to obtain the time of respiratory rate fluctuation; The respiratory instability at each moment is obtained by considering the number of fluctuating moments within the reference period and the maximum degree of fluctuation within the reference period. The respiratory instability is positively correlated with both the number of fluctuating moments and the maximum degree of fluctuation. The unstable moments are derived from the instability of the respiratory state at each moment.

[0006] In an exemplary embodiment, the process of obtaining the filtering features includes: The noise interference intensity at unstable moments is obtained based on the changes in respiratory rate within a local time range at unstable moments and the instability of respiratory state at unstable moments. The noise moment is obtained from the unstable moment based on the noise interference intensity. The filtering weight for the noise time is determined based on the noise interference intensity at the noise time, and the filtering weight is negatively correlated with the noise interference intensity; a preset constant is used as the filtering weight for other times besides the noise time; the filtering weight characterizes the filtering feature.

[0007] In an exemplary embodiment, the process of obtaining the noise interference intensity includes: The continuity of respiratory instability at a candidate unstable moment is determined based on the number and length of unstable periods within a reference period of the candidate unstable moment; the candidate unstable moment is any unstable moment; the unstable period consists of at least two unstable moments that are sequentially consecutive; the continuity is negatively correlated with the number of unstable periods and positively correlated with the length of the unstable period. The trend of respiratory instability at each moment within a local time range of the candidate unstable moment is determined to obtain the respiratory instability trend segment in which the candidate unstable moment is located; the respiratory instability trend segment consists of at least two moments that are sequentially consecutive and have the same trend. Based on the length of the unstable respiratory state trend segment and the overall trend of respiratory state instability, the trend of respiratory state instability at the candidate unstable moment is obtained. The noise interference intensity at the candidate unstable moment is obtained based on the respiratory state instability, continuity, and trend. The noise interference intensity is positively correlated with the respiratory state instability at the candidate unstable moment and negatively correlated with both continuity and trend.

[0008] In one exemplary embodiment, the process of obtaining the trend of change includes: Curve fitting is performed on the respiratory state instability within the local time range of the candidate unstable moment to obtain the fitted curve; Determine the slope of the tangent line for respiratory instability at each moment in the fitted curve; The trend of change is characterized by the sign of the slope of the tangent; The overall trend of respiratory instability in the unstable respiratory state trend segment is specifically the average value of the absolute value of the tangent slope at each moment in the unstable respiratory state trend segment. The process of acquiring the trend includes: The trend of respiratory instability at candidate unstable moments is obtained by averaging the length of the unstable trend segment and the absolute value of the tangent slope. The trend is positively correlated with both the length of the unstable trend segment and the average value of the tangent slope.

[0009] In an exemplary embodiment, the process of obtaining the activity moment includes: The moment within the reference time period of the current moment that meets the following conditions is determined as the active moment within the reference time period of the current moment: the updated respiratory rate is greater than the preset baseline of respiratory rate, and the respiratory depth is less than the preset baseline of respiratory depth.

[0010] In an exemplary embodiment, the process of obtaining the rest effect includes: Determine the duration characteristics of the inactive period relative to the duration of the active period; Determine the extent to which the overall respiratory rate during the active period exceeds the overall respiratory rate during the inactive period; Based on the duration characteristics and the degree of excess of the overall respiratory rate, the rest effect of the user after the activity period is obtained; the rest effect is positively correlated with both the duration characteristics and the degree of excess of the overall respiratory rate.

[0011] In an exemplary embodiment, the process of obtaining the activity intensity includes: Determine the respiratory intensity characteristics of the active period, wherein the respiratory intensity characteristics characterize the extent to which the maximum respiratory rate of the active period exceeds the maximum respiratory depth of the inactive period; The activity intensity of the activity period is obtained based on the duration of the activity period and the respiratory intensity characteristics; the activity intensity is positively correlated with both the duration of the activity period and the respiratory intensity characteristics.

[0012] In an exemplary embodiment, the process of obtaining the oxygen supply demand includes: Based on the activity intensity and rest effect during the activity period, an oxygen demand index for the activity period is obtained; the oxygen demand index is positively correlated with the activity intensity and negatively correlated with the rest effect. By integrating the oxygen demand indicators of each activity period within the reference time period at the current moment, the user's oxygen demand at the current moment can be obtained.

[0013] In a second aspect of the present invention, a portable oxygen supply device for the elderly is provided, comprising: a portable oxygen supply device body for the elderly, and the aforementioned intelligent flow regulation system.

[0014] This invention offers the following advantages: Due to the inherent variability in the physiological rhythms of users (i.e., the elderly), it first identifies unstable moments in the respiratory rate over time, avoiding including moments of normal respiratory rate in the data analysis, thus improving the relevance of the data analysis and reducing its workload. Addressing the susceptibility of elderly individuals' breathing to activity segments, it obtains filtering characteristics for each moment to filter the initially collected respiratory rate time-series data, resulting in more accurate updated respiratory rate time-series data. This effectively eliminates local random fluctuations and preserves respiratory rate changes related to actual activity or recovery processes. Furthermore, because the respiratory rate and depth of the elderly deviate significantly from the preset baseline during activity, this invention provides additional advantages. Small correlations, through double verification, determine the activity time, thus laying the foundation for accurate segmentation of activity and inactivity periods. By correlating the elderly's activity periods with adjacent inactivity periods, the rest effect after the activity period can be obtained. The rest effect represents the elderly's recovery ability after activity and can characterize their oxygen demand status. At the same time, the activity intensity during the activity period represents the physiological load caused by the elderly's activity and can also characterize their oxygen demand status. Combining these two characteristics, the accurate oxygen supply requirements of the elderly can be obtained, improving the accuracy and reliability of oxygen supply requirement acquisition. Consequently, the oxygen supply flow rate of the elderly can be accurately and reliably adjusted according to their oxygen supply requirements. Attached Figure Description

[0015] Figure 1 This is a flowchart of the intelligent flow regulation method corresponding to the intelligent flow regulation system in an embodiment of a portable oxygen supply device for the elderly provided by the present invention; Figure 2 This is a flowchart of the filtering feature acquisition process provided by the present invention; Figure 3 This is a flowchart of the process for obtaining the rest effect provided by the present invention. Detailed Implementation

[0016] An embodiment of a portable oxygen supply device for the elderly: This embodiment provides a portable oxygen supply device for the elderly, including: a portable oxygen supply device body and a flow intelligent adjustment system. The portable oxygen supply device body can be a common type of portable oxygen supply device. In one exemplary embodiment, its hardware structure includes: a high-pressure oxygen cylinder, a pressure reducing device, and oxygen delivery accessories. The high-pressure oxygen cylinder safely stores high-purity liquid oxygen, which is converted into a breathable gas through the vaporization of special materials. The pressure reducing device includes a main valve, a pressure reducing valve, a flow regulating valve, and a pressure gauge, etc., and its function is to reduce the pressure of the high-pressure oxygen to a safe pressure. The oxygen delivery accessories include an oxygen supply line, a nasal cannula, and a mask, etc., and their function is to deliver the depressurized oxygen to the user (specifically, the elderly) for inhalation.

[0017] The intelligent flow regulation system is used to intelligently regulate the oxygen flow rate in the portable oxygen supply device for the elderly. The system can be configured to include a memory and a processor; the memory is connected to the processor; the memory stores program instructions; the processor implements an intelligent flow regulation method when the program instructions are executed. Therefore, in specific implementations, the intelligent flow regulation system can be configured as a software system, achieving intelligent regulation of the oxygen flow rate through a software program corresponding to the internally executed intelligent flow regulation method; or it can be configured as a hardware system, including a data acquisition section and a data processing section. The data acquisition section collects the data required for intelligent flow regulation, and the data processing section receives the data collected by the data acquisition section, processes the data, and outputs the user's oxygen supply needs in a time sequence, thereby achieving accurate regulation of the oxygen flow rate. This embodiment does not limit the specific configuration of the intelligent flow regulation system.

[0018] To accurately determine a user's oxygen supply needs over time, this embodiment requires acquiring time-series data on the user's respiratory rate and respiratory depth during use of the portable oxygen supply device for the elderly. In an exemplary embodiment, a first gas flow sensor is installed in the oxygen supply pipeline. This first gas flow sensor can be specifically installed on the side near the high-pressure oxygen cylinder and can be a high-sensitivity gas flow sensor. When the user breathes, the oxygen flow rate fluctuates. The first gas flow sensor captures the periodic peaks of the airflow fluctuations to determine a complete breathing process, thereby obtaining the user's respiratory rate time-series data, i.e., the time sequence. A second gas flow sensor is installed at the inlet of the oxygen into the user's mask or nasal cannula (i.e., the user's inhalation inlet). This second gas flow sensor can also be a high-sensitivity gas flow sensor. The relative respiratory negative pressure amplitude is collected by a micro-pressure sensor in the second gas flow sensor as the respiratory depth (i.e., respiratory intensity), thereby obtaining the user's respiratory depth time-series data.

[0019] In this embodiment, the two gas flow sensors have the same sampling frequency and sample synchronously, ensuring that the respiratory rate and respiratory depth can be obtained at any sampling time (hereinafter referred to as time). The sampling frequency is set according to actual needs; this embodiment uses 2Hz as an example.

[0020] In other embodiments, respiratory rate and respiratory depth can also be acquired using other existing acquisition methods, such as embedding an inductor coil in an elastic band around the human chest and abdomen. The elastic band stretches and contracts with the human body's breathing, and the respiratory rate and respiratory depth are obtained through the stretching and contraction process.

[0021] To facilitate subsequent data processing, this embodiment can perform normalization processing on the time-series data of respiratory rate and respiratory depth separately to eliminate the influence of dimensions. In an exemplary embodiment, the normalization algorithm here includes two steps: a maximum and minimum value normalization algorithm and a sigmoid algorithm. Specifically, for respiratory rate, the maximum and minimum respiratory rates are determined from the user's historical respiratory rates. For example, within a historical period, the respiratory rates of the user in different states (including resting and several types of exercise) are obtained, and the maximum and minimum respiratory rates are determined from them. This not only ensures that the maximum and minimum respiratory rates conform to the user's actual respiratory rate range, but also ensures that there is a certain difference between the maximum and minimum respiratory rates. The respiratory rate is initially normalized using a maximum-minimum normalization algorithm. It should be understood that if, in extreme cases, the respiratory rate exceeds the maximum respiratory rate, the initial normalization result is greater than 1, and this result is retained. Conversely, if, in extreme cases, the respiratory rate falls below the minimum respiratory rate, the initial normalization result is less than 0, and this result is also retained. Then, the Sigmoid algorithm is used to further normalize the initial result, ensuring that while preserving the fluctuations in respiratory rate at various times, the final normalized result is limited to the range of 0 to 1. The respiratory rates mentioned below represent the normalized results.

[0022] Similarly, for breathing depth, the maximum and minimum breathing depths are determined from the user's historical breathing depths. For example, over a historical period, the user's breathing depths in different states (including resting and several types of exercise) are obtained, and the maximum and minimum breathing depths are determined from these. This ensures that the maximum and minimum breathing depths not only match the user's actual breathing depth range but also that there is a certain difference between the maximum and minimum breathing depths. A maximum-minimum value normalization algorithm is used for initial normalization of the breathing depth. It should be understood that if, in extreme cases, the breathing depth is higher than the maximum breathing depth, the initial normalization result is greater than 1, and this initial normalization result is retained; if, in extreme cases, the breathing depth is lower than the minimum breathing depth, the initial normalization result is less than 0, and this initial normalization result is retained. Then, the Sigmoid algorithm is used to further normalize the initial normalization result, so that while preserving the fluctuation of breathing depth at various times, the final normalization result is limited to the range of 0 to 1. The breathing depths mentioned below are the normalized results.

[0023] The intelligent flow regulation method implemented by the processor in the intelligent flow regulation system is as follows: Figure 1 As shown: Step S1: Determine the moment of instability based on the instability of the user's respiratory rate; Step S2: Based on the changes in respiratory rate within the local time range of unstable moments and the instability of respiratory rate at unstable moments, obtain the filtering characteristics for each moment to filter the respiratory rate time series data and obtain updated respiratory rate time series data. Step S3: Determine the activity time within the reference time period for the current moment based on the relationship between the updated respiratory rate and respiratory depth at each moment and the corresponding preset baseline; Step S4: Determine the user's rest effect after the active period based on the correlation between the active period and the adjacent inactive period; Step S5: Determine the activity intensity during the activity period and combine it with the rest effect to obtain the user's oxygen supply needs at the current moment.

[0024] The following is a detailed explanation of each step.

[0025] Step S1: Determine the moment of instability based on the instability of the user's breathing rate.

[0026] Elderly people may have irregular breathing rates, which can vary significantly depending on their activity level, such as resting or walking. Insufficient oxygen supply during activity can lead to hypoxia, chest tightness, and shortness of breath. Excessive oxygen supply during rest can result in oxygen waste, nasal dryness, and even the risk of oxygen toxicity from prolonged high-flow rates. Therefore, it is necessary to determine the user's actual oxygen needs and adjust the flow rate automatically.

[0027] Due to the unique respiratory characteristics of the elderly, their respiratory rate differs from that of healthy individuals. Therefore, directly relying on raw respiratory rate data to determine a user's current oxygen demand can compromise accuracy. Consequently, the initially acquired respiratory rate time-series data needs to be filtered to improve its reliability.

[0028] First, based on the time-series data of respiratory rate, unstable moments are determined according to the instability of the user's respiratory rate at various times. Unstable moments represent the moments when the respiratory rate fluctuates.

[0029] In this embodiment, a candidate time is defined as any time in the time sequence. A reference time period is determined for the candidate time, which is a time period associated with the candidate time. In this embodiment, the end time (i.e., the last time) of the reference time period is the candidate time. The duration of the reference time period is determined, specifically referring to the number of times it includes. The duration of the reference time period is set by the implementer according to adjustment needs or based on experience. In an exemplary embodiment, the reference time period includes 60 times. Therefore, the reference time period for the candidate time includes the candidate time and the 59 times preceding it.

[0030] The degree of fluctuation in respiratory rate at candidate time points is determined, where the degree of fluctuation represents the difference in respiratory rate between the candidate time point and its adjacent time points. In an exemplary embodiment, the respiratory rate of the preceding time point adjacent to the candidate time point is determined, and the absolute value of the difference between the respiratory rate of the candidate time point and the respiratory rate of the preceding adjacent time point is calculated. This absolute value of the difference is used as the degree of fluctuation of the respiratory rate at the candidate time point. It should be understood that if the degree of fluctuation of respiratory rate at the first time point in the time series cannot be obtained, then the degree of fluctuation of respiratory rate at the first time point in the time series is not calculated.

[0031] Based on the degree of fluctuation in respiratory rate at candidate times, it is determined whether a candidate time is a fluctuating time. A fluctuating time indicates a moment when the respiratory rate fluctuates to a certain extent, i.e., whether the degree of fluctuation is high. In an exemplary embodiment, this embodiment presets a fluctuation threshold. The fluctuation of the respiratory rate at a candidate time is compared with the fluctuation threshold. If the fluctuation of the respiratory rate at a candidate time is greater than or equal to the fluctuation threshold, the candidate time is determined to be a fluctuating time; if the fluctuation of the respiratory rate at a candidate time is less than the fluctuation threshold, the candidate time is determined not to be a fluctuating time. Each moment in the time sequence is compared one by one to determine the fluctuating time. In principle, the value range of the fluctuation threshold is 0 to 1. The specific value is set according to the actual judgment needs. If a more secure screening logic is required, the fluctuation threshold can be set slightly smaller, such as 0.2; or, the fluctuation threshold can be obtained through historical statistics, for example: manually calibrating the fluctuating times identified in the historical time period, determining the degree of fluctuation of the respiratory rate at these fluctuating times, and selecting the fluctuation corresponding to the 10th percentile as the fluctuation threshold.

[0032] Following the above method, each fluctuation moment within the reference period of the candidate moment is obtained, thereby determining the number of fluctuation moments within the reference period of the candidate moment. The more fluctuation moments there are, the more moments with respiratory rate fluctuations within the reference period of the candidate moment, and the higher the respiratory instability of the candidate moment. Respiratory instability is positively correlated with the number of fluctuation moments.

[0033] The fluctuation levels of each fluctuation moment within the reference period of the candidate time are obtained, and the maximum fluctuation level is taken as the maximum fluctuation level within the reference period of the candidate time. The larger the maximum fluctuation level, the higher the maximum amplitude of respiratory rate fluctuation in the reference period of the candidate time, and the higher the respiratory instability of the candidate time. Respiratory instability is positively correlated with the maximum fluctuation level. Therefore, the respiratory instability of the candidate time is obtained based on the number of fluctuation moments and the maximum fluctuation level within the reference period of the candidate time. Based on the above logical analysis, a specific calculation method for respiratory instability is given below: ; in, This indicates the instability of the respiratory state at the candidate time point. This indicates the number of fluctuating moments within the reference period for the candidate moment. This indicates the total number of times within the reference time period for the candidate time. This indicates the percentage of fluctuating moments within the reference period for the candidate moment. This represents the maximum fluctuation level within the reference period of the candidate time. This embodiment uses an averaging method to effectively integrate the proportion of fluctuating moments within the reference period of the candidate time with the maximum fluctuation level within the reference period, thereby obtaining the respiratory instability of the candidate time. It should be understood that if there are no fluctuating moments within the reference period of the candidate time, i.e., the number of fluctuating moments is 0, then the respiratory instability of the candidate time is directly set to 0. This is how the respiratory instability at each time point is obtained.

[0034] The higher the respiratory instability value of a candidate time point, the more unstable the respiratory state at that time point, and the more likely the candidate time point is to be unstable. Therefore, the instability of the respiratory state at a candidate time point is used to determine whether it is an unstable time point, thus obtaining the unstable time points in the time series. In an exemplary embodiment, this embodiment presets an instability threshold. The respiratory instability of a candidate time point is compared with this instability threshold. If the respiratory instability of a candidate time point is greater than or equal to the instability threshold, the candidate time point is determined to be unstable; if the respiratory instability of a candidate time point is less than the instability threshold, the candidate time point is determined not to be unstable. Each time point in the time series is compared one by one to determine the unstable time points in the time series. In principle, the value range of the instability threshold is 0 to 1. The specific value is set according to the actual judgment needs. If a more secure filtering logic is required, the instability threshold can be set slightly smaller, such as 0.6; or, the instability threshold can be obtained through historical statistics, for example: manually calibrating the unstable times identified in historical time periods, determining the respiratory instability of these unstable times, and selecting the respiratory instability corresponding to the 10th percentile as the instability threshold.

[0035] It should be understood that if the above judgment determines that there are no unstable moments, then the following filtering process for respiratory rate time series data will not be performed, and the collected respiratory rate time series data will be directly used as the subsequent updated respiratory rate time series data.

[0036] Step S2: Based on the changes in respiratory rate within the local time range of unstable moments and the instability of respiratory rate at unstable moments, obtain the filtering characteristics for each moment to filter the respiratory rate time series data and obtain updated respiratory rate time series data.

[0037] Pathological respiratory instability in the elderly is typically gradual, continuous, and trend-driven, with corresponding unstable moments occurring in clusters and consecutively, exhibiting a clear continuous trend rather than being scattered. Noise (such as equipment vaporization noise, pump vibration, and pipeline swaying) is an independent, transient disturbance, causing random and short-lived interference to respiratory rate. Therefore, analyzing the changes in respiratory rate within a local time range of unstable moments, and the instability of respiratory rate during these moments, to obtain filtering characteristics for each time point, can improve the accuracy of the filtering features. After obtaining the filtering characteristics for each time point, the respiratory rate time-series data can be filtered to obtain updated respiratory rate time-series data.

[0038] In one exemplary embodiment, such as Figure 2 As shown, one process for obtaining filtered features is presented: Step S21: Based on the changes in respiratory rate within the local time range of the unstable moment and the instability of the respiratory state at the unstable moment, obtain the noise interference intensity at the unstable moment.

[0039] For any unstable moment, it is designated as a candidate unstable moment. This embodiment takes the *a*th unstable moment as an example to determine its local time range. In this embodiment, the local time range of the *a*th unstable moment is the time period within a preset radius centered on the *a*th unstable moment. The specific value of the preset radius is set according to actual needs. If the value is too large, weakly correlated moments may be included in the local time range; if the value is too small, strongly correlated moments may be filtered out. In this embodiment, the preset radius is 15, meaning the time period consisting of the *a*th unstable moment, its 15 preceding moments, and its 15 following moments is considered the local time range of the *a*th unstable moment. It should be understood that if the *a*th unstable moment is located at several beginning or end positions in the time sequence, a complete local time range may not be obtained; in such cases, the actually obtained local time range shall prevail.

[0040] The changes in respiratory rate within the local time range of the a-th unstable moment and the instability of respiratory state at the a-th unstable moment can both characterize the noise interference intensity at the a-th unstable moment to a certain extent. Therefore, by comprehensively analyzing these two features, the noise interference intensity at the a-th unstable moment can be obtained.

[0041] In an exemplary embodiment, a reference time period for the a-th unstable moment is first determined, thereby obtaining unstable moments within the reference time period for the a-th unstable moment. At least two temporally consecutive unstable moments are combined to form an unstable time period, thus obtaining several unstable time periods within the reference time period for the a-th unstable moment. This embodiment does not treat an isolated unstable moment as a separate unstable time period because an isolated unstable moment is highly likely to be noisy data.

[0042] Obtain the number and length of each unstable period within the reference period of the a-th unstable moment, where length refers to the number of moments included. The fewer the number of unstable periods and the longer the duration of each unstable period (i.e., the longer the duration of the unstable period), the more concentrated the unstable moments are within the reference period of the a-th unstable moment, the more continuous the respiratory instability is in time, and the stronger the continuity of respiratory instability at the a-th unstable moment. Therefore, the continuity of respiratory instability at the a-th unstable moment is negatively correlated with the number of unstable periods and positively correlated with the length of the unstable periods. Specifically: Calculate the sum of the lengths of each unstable period within the reference period of the a-th unstable moment, as the total length of unstable periods within the reference period of the a-th unstable moment. Then, based on the number of unstable periods and the total length of unstable periods within the reference period of the a-th unstable moment, obtain the continuity of respiratory instability at the a-th unstable moment. Continuity is negatively correlated with the number of unstable periods and positively correlated with the total length of unstable periods. Based on the above logical analysis, the following is a specific calculation method for the continuity of respiratory instability at the a-th unstable moment: ; in, This indicates the continuity of the unstable respiratory state at the a-th unstable moment. This represents the total length of the unstable time interval within the reference time interval at the a-th unstable moment. This represents the percentage of the total length of the unstable period within the reference period at the a-th unstable moment. This represents the number of unstable periods within the reference period at the a-th unstable moment. The value 1 in the numerator has the same dimension as the number of unstable periods. This represents the dimensionless characteristic corresponding to the reciprocal of the number of unstable periods. It should be understood that if there are no unstable periods within the reference period of the a-th unstable moment, the above calculation is not performed, and the continuity of respiratory instability at the a-th unstable moment is directly set to 1.

[0043] The respiratory instability at each moment within the local time range of the a-th unstable moment is obtained, and then the changing trend of the respiratory instability at each moment is determined. In an exemplary embodiment, curve fitting is performed on the respiratory instability at each moment within the local time range of the a-th unstable moment to obtain a fitted curve. The least squares method is used for polynomial curve fitting, and the order of the polynomial can be determined empirically based on the specific changes in the respiratory instability; this embodiment uses order 3 as an example.

[0044] Determine the slope of the tangent line at each time step in the fitted curve of respiratory instability within the local time range of the a-th unstable time step. The sign of the tangent line slope at each time step represents the trend of respiratory instability at that time step. Specifically, when the tangent line slope is positive, the trend of respiratory instability is increasing; when the tangent line slope is negative, the trend of respiratory instability is decreasing; and when the tangent line slope is 0, the trend of respiratory instability is stable.

[0045] Within the local time range of the a-th unstable moment, at least two consecutive moments with the same trend are considered as a respiratory instability trend segment. For example, if three moments with increasing trends are consecutive, these three moments are considered as a respiratory instability trend segment. This results in several respiratory instability trend segments within the local time range of the a-th unstable moment. The respiratory instability trend segment in which the a-th unstable moment is located is determined based on these segments. It should be understood that if the a-th unstable moment is not within any of these respiratory instability trend segments (including situations where there is no respiratory instability trend segment within the local time range of the a-th unstable moment, or where there is a respiratory instability trend segment within the local time range of the a-th unstable moment, but the trend of the a-th unstable moment is isolated and it does not belong to any specific respiratory instability trend segment), then no further processing is performed, and the respiratory instability trend of the a-th unstable moment is directly set to 0.

[0046] Determine the length of the unstable trend segment of the respiratory state at the a-th unstable moment. The longer the length of the unstable trend segment of the respiratory state at the a-th unstable moment, the longer the duration of the same trend of respiratory state instability, and the stronger the trend of respiratory state instability at the a-th unstable moment. The two are positively correlated.

[0047] To determine the absolute value of the tangent slope of the respiratory instability at each time point within the unstable trend segment of the respiratory state at the a-th unstable time point, the absolute value of the tangent slope is normalized as follows for ease of calculation: A preset upper limit value for the slope is determined. This preset upper limit value can be the ratio of the theoretical maximum value of the normalized respiratory instability (e.g., a value of 1) to the time span of a single sampling period. The ratio of the absolute value of the tangent slope to this preset upper limit value is calculated as the absolute value of the normalized tangent slope. All absolute values ​​of tangent slopes mentioned below refer to the absolute value of the normalized tangent slope.

[0048] Calculate the average of the absolute values ​​of the slopes of the tangents to the respiratory instability at each time point within the unstable trend segment at time point a. The larger this average value, the stronger the trend of respiratory instability at time point a; the two are positively correlated.

[0049] Based on the above logical analysis, the following is a specific calculation method for the trend of respiratory instability at the a-th unstable moment: ; in, This indicates the trend of unstable respiratory state at the a-th unstable moment. This represents the length of the unstable trend segment of the respiratory state at the a-th unstable moment. This indicates the length of the local time range during the unstable moment. This represents the proportion of the unstable respiratory state at the a-th unstable moment. This represents the average absolute value of the slope of the tangent line representing the respiratory instability at each time point within the unstable trend segment of the respiratory state at the a-th unstable time point.

[0050] The stronger the instability of the respiratory state at the *a*th unstable moment, the more significant the noise interference on the respiratory rate at that moment, and the more it masks the true respiratory signal of the elderly. The stronger the noise interference intensity at the *a*th unstable moment, the stronger the correlation between the two. The stronger the continuity of the unstable respiratory state at the *a*th unstable moment, the more consistent it is with the respiratory characteristics of the elderly, and the less likely it is caused by interfering noise. The weaker the noise interference intensity at the *a*th unstable moment, the weaker the correlation between the two. The stronger the trend of the unstable respiratory state at the *a*th unstable moment, the more consistent it is with the respiratory characteristics of the elderly, and the less likely it is caused by interfering noise. The weaker the noise interference intensity at the *a*th unstable moment, the weaker the correlation between the two.

[0051] By comprehensively analyzing the instability, continuity, and trend of the breathing state at the a-th unstable moment, the noise interference intensity at the a-th unstable moment is obtained. Based on the above logical analysis, a specific calculation method for the noise interference intensity at the a-th unstable moment is given below: ; in, Let represent the noise interference intensity at the a-th unstable moment. This indicates the instability of the respiratory state at the a-th unstable moment. This can be characterized as the probability of noise interference at the a-th unstable moment. The more unstable the user's breathing state is at the a-th unstable moment, and the greater the probability of noise interference, indicating that the noise interference on the breathing frequency at the a-th unstable moment is more significant. The noise interference intensity at each unstable moment can then be obtained.

[0052] Step S22: Obtain the noise moment from the unstable moment based on the noise interference intensity.

[0053] The stronger the noise interference, the greater the degree of noise interference at the corresponding unstable moment, and the more likely the unstable moment is to be a noise moment. Therefore, based on the noise interference intensity at each unstable moment, noise moments are selected from the unstable moments.

[0054] In an exemplary embodiment, this embodiment presets a noise interference intensity threshold. The noise interference intensity at each unstable moment is compared with this threshold. Unstable moments with noise interference intensities greater than or equal to the threshold are taken as noise moments, thus obtaining the time-series noise moments. In principle, the noise interference intensity ranges from 0 to 1, and the specific value is set according to actual judgment needs. If a more secure filtering logic is required, the noise interference intensity threshold can be set slightly smaller, such as 0.6. Alternatively, the noise interference intensity threshold can be obtained through historical statistics, for example: manually calibrating the identified noise moments within a historical time period, determining the noise interference intensity of these noise moments, and selecting the noise interference intensity corresponding to the 10th percentile as the noise interference intensity threshold.

[0055] It should be understood that if the noise time cannot be obtained through the above analysis, the filtering process of the respiratory rate time series data will not be performed, and the collected respiratory rate time series data will be directly used as the subsequent updated respiratory rate time series data.

[0056] Step S23: Determine the filtering weight for the noise time based on the noise interference intensity at the noise time. The filtering weight is negatively correlated with the noise interference intensity. Use a preset constant as the filtering weight for other times besides the noise time.

[0057] For the elderly, large errors in oxygen demand can affect the accuracy of oxygen flow rate regulation, thus posing significant health risks. Therefore, it is necessary to filter out noise in the respiratory rate time series data to ensure the accuracy of oxygen demand obtained from the real respiratory signal, thereby achieving accurate oxygen flow rate regulation.

[0058] This embodiment employs a weighted moving average filtering algorithm to filter the collected respiratory rate time-series data. For any given noise moment, the filtering weight is determined based on the noise interference intensity at that moment. The greater the noise interference intensity at that moment, the more likely the respiratory rate data at that moment reflects noise rather than the true respiratory rate data, and the more likely the data at that moment is false information. Since the weighted moving average filtering algorithm aims to make the calculated result as close to the true value as possible when performing smoothing, a smaller filtering weight is assigned if the noise interference intensity at that moment is high. This pulls the calculated result closer to the true value, making the smoothed result closer to the true trend. Therefore, the filtering weight is negatively correlated with the noise interference intensity.

[0059] In an exemplary embodiment, any noise moment is set as the z-th noise moment, and the noise interference intensity at the z-th noise moment is... Then the filter weight at the z-th noise moment is: This yields the filtering weights for each noise moment. For moments other than the noise moment, since the noise interference intensity is lower, the filtering weights for these moments are higher. In this embodiment, a constant is preset and used as the filtering weight for all moments other than the noise moment. It should be understood that this constant is greater than the filtering weight for each noise moment; for example, this constant is set to the value 1. This yields the filtering weight for each moment, which characterizes the filtering features at each moment.

[0060] After obtaining the filter weights at each time point, the filter weights at each time point, along with the corresponding respiratory rate time-series data, are input into the weighted moving average filtering algorithm to smooth the respiratory rate time-series data, resulting in the smoothed respiratory rate at each time point, defined as the updated respiratory rate at each time point. The updated respiratory rate time-series data is then sorted. It should be understood that the smoothing process of the weighted moving average filtering algorithm is existing technology. The overall processing approach is as follows: Determine the sliding window width, i.e., the number of time points in the window participating in the smoothing process; normalize the filter weights at each time point within the sliding window so that the sum of all weights within the sliding window is 1; calculate the weighted moving average within the sliding window, i.e., the smoothing result, which is a weighted sum of the data at each time point within the sliding window based on the normalized weights; slide forward one time point and repeat the above smoothing process.

[0061] Step S3: Determine the activity time within the reference time period for the current moment based on the relationship between the updated respiratory rate and respiratory depth at each moment and the corresponding preset baseline.

[0062] In the process of obtaining oxygen demand, if only the current state of the elderly is considered and the intensity of subsequent activities is not taken into account, the oxygen supply will not be able to keep up with the oxygen consumption once the activity level increases. Therefore, after obtaining the updated respiratory rate time series data, the activity time is determined according to the changes in the activity intensity of the elderly.

[0063] This embodiment is used to obtain the user's real-time oxygen supply needs. Therefore, this embodiment takes the current moment as an example to determine the user's oxygen supply needs at the current moment. Accordingly, a reference time period is determined for the current moment, and the updated respiratory rate and respiratory depth are obtained for each moment within the reference time period.

[0064] Due to physical factors, older adults experience increased oxygen consumption with even slightly higher activity levels. Unlike younger people, they cannot increase ventilation by deepening their breaths; instead, they compensate by increasing their respiratory rate, resulting in rapid and shallow breathing. Therefore, the activity time within the reference time period is determined by comparing the updated respiratory rate and depth at each moment within the current reference time period with the corresponding preset baseline.

[0065] The preset baselines include a respiratory rate preset baseline and a respiratory depth preset baseline. The respiratory rate preset baseline represents the overall level of the user's respiratory rate in a resting state. In an exemplary embodiment, the first 3 minutes after the system is powered on can be used as the resting calibration period. The user is required to be in a sitting or lying position to rest, and the respiratory rate and respiratory depth are obtained at each moment during the resting calibration period. The average respiratory rate at each moment during the resting calibration period is calculated as the respiratory rate preset baseline, and the average respiratory depth at each moment during the resting calibration period is calculated as the respiratory depth preset baseline. Alternatively, if the environment does not have calibration conditions, the preset empirical benchmark values ​​can be used directly. It should be understood that, for ease of comparison, the respiratory rate preset baseline and the respiratory depth preset baseline need to be normalized according to the normalization method for respiratory rate and respiratory depth described above.

[0066] Because older adults tend to breathe rapidly and shallowly during exercise, this embodiment sets a criterion: the updated respiratory rate is greater than a preset baseline, and the respiratory depth is less than a preset baseline. The system checks each moment within a reference time period to see if this criterion is met. Moments that meet this criterion are designated as active moments within the reference time period, indicating that the user is in an active state. All moments within the reference time period other than active moments are defined as resting moments.

[0067] At least one active period is obtained by combining at least two consecutive active moments within the reference period of the current moment into an active period. Similarly, at least one inactive period is obtained by combining at least two consecutive resting moments within the reference period of the current moment into an inactive period. If an isolated active moment is flanked by inactive periods, the isolated active moment is treated as a resting moment, and the isolated active moment and its flanking inactive periods are connected to form one inactive period. Likewise, if an isolated resting moment is flanked by active periods, the isolated resting moment is treated as an active moment, and the isolated resting moment and its flanking active periods are connected to form one active period.

[0068] It should be understood that if there are no active periods within the current reference time period (e.g., no active moments or only isolated active moments, which are considered noise data), and the user is determined to be in a resting state throughout the current reference time period, then the following analysis will not be performed, and the user's oxygen demand at the current time will be directly set to 0. If the current reference time period is an active period, meaning there are no resting moments or only isolated resting moments, then the user is determined to be in an active state throughout the current reference time period, then the following analysis will not be performed, and the user's oxygen demand at the current time will be directly set to a preset upper limit value, such as 1. If there are alternating active and resting moments within the current reference time period, and the total number of alternating active and resting moments is greater than a preset value, such as 4, then the current reference time period is determined to be in an abnormal state, then the following analysis will not be performed, and an abnormal warning signal will be directly output to facilitate timely investigation of abnormalities by relevant personnel.

[0069] Step S4: Determine the rest effect of the user after the active period based on the correlation between the active period and the adjacent inactive period.

[0070] For any activity period within a reference time period at the current moment, determine the inactive period that follows and is adjacent to that activity period. The inactive period adjacent to the activity period represents the user's rest period after the activity. After any activity, the user experiences a certain degree of physical exertion, especially elderly people with abnormal lung function. The better the quality of rest, the more oxygen the user obtains during the rest period, the better the body's recovery after the activity, and the less impact the activity has on the user's body. Based on the correlation between the activity period and the inactive period adjacent to the activity period, determine the user's rest effect after the activity period. In an exemplary embodiment, such as... Figure 3 As shown, the following is a specific process for obtaining the effect of rest: Step S41: Determine the duration characteristics of inactive periods relative to active periods.

[0071] Any active period within the reference period at the current time is designated as the h-th active period, and the inactive period following and adjacent to the h-th active period is designated as the h-th inactive period. The duration characteristic of the h-th inactive period relative to the duration of the h-th active period is determined. In an exemplary embodiment, the duration characteristic of the h-th inactive period relative to the duration of the h-th active period is: ,in, This represents the duration of the h-th inactive period. This represents the duration of the h-th activity period. The larger the value of the duration feature of the h-th inactive period relative to the duration of the h-th activity period, the longer the rest time after the activity, and the better the rest effect. The rest effect is positively correlated with the duration feature.

[0072] Step S42: Determine the extent to which the overall respiratory rate during the active period exceeds the overall respiratory rate during the inactive period.

[0073] The degree to which the overall respiratory rate exceeds the overall respiratory rate during the h-th active period relative to the overall respiratory rate during the h-th inactive period is determined. In an exemplary embodiment, the degree to which the overall respiratory rate exceeds the overall respiratory rate is: ,in, This represents the overall respiratory rate during the h-th inactive period (i.e., the average respiratory rate at each moment within the h-th inactive period). This represents the overall respiratory rate during the h-th activity period (i.e., the average respiratory rate at each moment within the h-th activity period). `max` represents the maximum value function, used to ensure that the overall respiratory rate exceeding the limit is non-negative. A higher overall respiratory rate exceeding the limit indicates a greater decrease in respiratory rate during the h-th activity period compared to the h-th inactive period, indicating a better rest effect after the activity. The rest effect is positively correlated with the overall respiratory rate exceeding the limit.

[0074] Step S43: Based on the duration characteristics and the degree of excess in the overall respiratory rate, obtain the rest effect of the user after the activity period.

[0075] Based on the duration characteristic of the h-th inactive period relative to the h-th active period (referred to as the duration characteristic of the h-th active period), and the degree to which the overall respiratory rate of the h-th active period exceeds the overall respiratory rate of the h-th inactive period (referred to as the degree of overall respiratory rate exceedance of the h-th active period), the user's rest effect after the h-th active period is obtained. Based on the above logical analysis, the following is a method for calculating the user's rest effect after the h-th active period: Calculate the average of the duration characteristic of the h-th active period and the degree of overall respiratory rate exceedance of the h-th active period, and use the result as the user's rest effect after the h-th active period. The better the user's rest effect, the less oxygen is required, and the lower the oxygen supply demand.

[0076] It should be understood that if there is no non-activity period after the activity period within the reference time period at the current moment, for example, if the end time of the activity period within the reference time period at the current moment is the current moment, then the rest effect of the user after the activity period will be directly set to 0.

[0077] Step S5: Determine the activity intensity during the activity period and combine it with the rest effect to obtain the user's oxygen supply needs at the current moment.

[0078] To determine the user's activity intensity during the h-th activity period, specifically: determine the maximum respiratory rate from the respiratory rates at each time point in the h-th activity period, and determine the maximum respiratory depth from the respiratory depths at each time point in the h-th inactive period, and then determine the respiratory intensity characteristics of the h-th activity period. These characteristics characterize the extent to which the maximum respiratory rate of the h-th activity period exceeds the maximum respiratory depth of the h-th inactive period. Specifically, the respiratory intensity characteristics are as follows: ,in, This represents the maximum respiratory rate during the h-th activity period. This represents the maximum respiratory depth during the h-th inactive period. express Normalization function. (Through) The case where the maximum respiratory rate is less than the maximum respiratory depth is truncated, and 1 is subtracted to make the difference greater than or equal to 0. Finally, normalization is performed using a normalization function.

[0079] During activity, elderly individuals exhibit compensatory shallow and rapid breathing, unable to increase the volume of a single inspiratory breath, only able to increase the respiratory rate. The maximum respiratory rate during the h-th activity period represents the degree of tachypnea; the higher the value, the more acute the hypoxia. The maximum respiratory depth during the h-th inactive period represents the limit of the elderly person's expedited breathing during that period. Due to pathological limitations, the maximum respiratory depth during the h-th inactive period is usually small and cannot increase significantly with activity levels. Therefore, a higher respiratory intensity characteristic value during the h-th activity period indicates that the elderly person can only perform very shallow breathing at a higher respiratory rate, accurately quantifying the severity of abnormal compensation under ventilatory limitation. Thus, this characteristic is used to characterize respiratory intensity. A higher respiratory intensity characteristic value indicates greater activity intensity, and activity intensity and respiratory intensity characteristic are positively correlated.

[0080] The duration of the h-th activity period is determined. The longer the duration of the h-th activity period, the greater the activity intensity of the h-th activity period, and the two are positively correlated.

[0081] Therefore, based on the respiratory intensity characteristics and duration of the h-th activity period, the activity intensity of the h-th activity period is obtained.

[0082] ; in, This indicates the activity intensity during the h-th activity period. This indicates the total number of moments within the reference time period, i.e., the duration of the reference time period. This indicates the duration percentage of the h-th activity period.

[0083] Based on the activity intensity and rest effect of each activity period within the reference time frame at the current moment, the user's oxygen demand at the current moment is obtained. In an exemplary embodiment, for any activity period within the reference time frame at the current moment, taking the h-th activity period as an example, the oxygen demand index for the h-th activity period is obtained based on the activity intensity and rest effect of the h-th activity period. The oxygen demand index is positively correlated with activity intensity and negatively correlated with rest effect. Specifically, the following is a calculation method for the oxygen demand index of the h-th activity period: ; in, This represents the oxygen demand index for the h-th activity period. This indicates the rest effect during the h-th activity period.

[0084] By integrating the oxygen demand indicators of each activity period within the reference time period at the current moment, the user's oxygen demand at the current moment can be obtained. In an exemplary embodiment, the integration method is given below: ; in, H represents the user's oxygen demand at the current moment, and H represents the number of activity periods within the reference time period at the current moment.

[0085] After obtaining the user's current oxygen demand, the oxygen supply flow rate for the user after the current moment is adjusted according to the oxygen demand.

[0086] During oxygen supply flow regulation, oxygen demand represents the relative compensation weight, indicating the severity of additional oxygen consumption exceeding the baseline resting state for the user. Oxygen demand cannot be directly input into the PID formula; instead, it serves as a pre-variable used to dynamically generate the PID controller's target oxygen supply setpoint (i.e., the desired oxygen supply flow). The specific calculation process is as follows: The user's baseline static oxygen supply flow (e.g., the baseline oxygen supply flow required for rest) and the maximum allowable additional oxygen supply compensation flow of the oxygen supply device are pre-set. Then, a physical mapping calculation is performed: Current oxygen supply target setpoint = Baseline static oxygen supply flow + Oxygen demand × Maximum additional oxygen supply compensation flow, thus obtaining the user's current oxygen supply target setpoint at the current moment. This current oxygen supply target setpoint is used as the target setpoint for the PID controller. The real-time oxygen supply flow is obtained and used as the current measurement value for the PID controller. The difference between the target setpoint and the current measurement value is used to obtain the flow deviation. The PID controller then performs proportional, integral, and derivative calculations based on this flow deviation, ultimately outputting the opening degree of the flow regulating valve. This opening degree controls the flow regulating valve's action, thereby regulating the oxygen supply flow.

[0087] To reduce the amount of data processing required for adjustment, this embodiment can determine the oxygen supply demand at regular intervals, such as every 10 time intervals, and then adjust the user's oxygen supply flow rate based on the oxygen supply demand at that time. Then, during the time interval between the current time and the next oxygen supply flow rate adjustment time, and within the time interval between the next oxygen supply flow rate adjustment time and the current time, oxygen is supplied according to the oxygen supply flow rate obtained based on the current oxygen supply demand.

[0088] An embodiment of a portable oxygen supply device for the elderly with intelligent flow regulation system: This embodiment provides a flow intelligent adjustment system for a portable oxygen supply device for the elderly. The flow intelligent adjustment system has been described in detail in the above embodiment of the portable oxygen supply device for the elderly, and will not be repeated here.

Claims

1. A portable oxygen supply device for the elderly with intelligent flow regulation system, characterized in that, It includes a memory and a processor; the memory is connected to the processor; the memory is used to store program instructions; the processor is used to implement the following intelligent flow regulation method when the program instructions are executed: Determine the moment of instability based on the instability of the user's respiratory rate; Based on the changes in respiratory rate within a local time range during unstable moments, and the instability of respiratory rate during unstable moments, the filtering characteristics of each moment are obtained to filter the respiratory rate time series data and obtain updated respiratory rate time series data. Based on the relationship between the updated respiratory rate and respiratory depth at each moment and the corresponding preset baseline, the activity time within the reference time period at the current moment is determined; The rest effect of a user after an active period is determined based on the correlation between the active period and the adjacent inactive period that follows; the active period consists of at least two consecutive active moments in time. The activity intensity during the activity period is determined, and the rest effect is combined to obtain the user's oxygen supply demand at the current moment; the oxygen supply demand is used to indicate the adjustment of the user's oxygen supply flow.

2. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 1, characterized in that, The process of obtaining the unstable moment includes: Determine the degree of fluctuation in respiratory rate at each moment to obtain the time of respiratory rate fluctuation; The respiratory instability at each moment is obtained by considering the number of fluctuating moments within the reference period and the maximum degree of fluctuation within the reference period. The respiratory instability is positively correlated with both the number of fluctuating moments and the maximum degree of fluctuation. The unstable moments are derived from the instability of the respiratory state at each moment.

3. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 2, characterized in that, The process of obtaining the filtering features includes: The noise interference intensity at unstable moments is obtained based on the changes in respiratory rate within a local time range at unstable moments and the instability of respiratory state at unstable moments. The noise moment is obtained from the unstable moment based on the noise interference intensity. The filtering weight for the noise time is determined based on the noise interference intensity at the noise time, and the filtering weight is negatively correlated with the noise interference intensity; a preset constant is used as the filtering weight for other times besides the noise time; the filtering weight characterizes the filtering feature.

4. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 3, characterized in that, The process of obtaining the noise interference intensity includes: The continuity of respiratory instability at a candidate unstable moment is determined based on the number and length of unstable periods within a reference period of the candidate unstable moment; the candidate unstable moment is any unstable moment; the unstable period consists of at least two unstable moments that are sequentially consecutive; the continuity is negatively correlated with the number of unstable periods and positively correlated with the length of the unstable period. The trend of respiratory instability at each moment within a local time range of the candidate unstable moment is determined to obtain the respiratory instability trend segment in which the candidate unstable moment is located; the respiratory instability trend segment consists of at least two moments that are sequentially consecutive and have the same trend. Based on the length of the unstable trend segment of the respiratory state and the overall trend of the instability of the respiratory state, the trend of respiratory instability at the candidate unstable moment is obtained. The noise interference intensity at the candidate unstable moment is obtained based on the respiratory state instability, continuity, and trend. The noise interference intensity is positively correlated with the respiratory state instability at the candidate unstable moment and negatively correlated with both continuity and trend.

5. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 4, characterized in that, The process of obtaining the changing trend includes: Curve fitting is performed on the respiratory state instability within the local time range of the candidate unstable moment to obtain the fitted curve; Determine the slope of the tangent line for respiratory instability at each moment in the fitted curve; The trend of change is characterized by the sign of the slope of the tangent; The overall trend of respiratory instability in the unstable respiratory state trend segment is specifically the average value of the absolute value of the tangent slope at each moment in the unstable respiratory state trend segment. The process of acquiring the trend includes: The trend of respiratory instability at candidate unstable moments is obtained by averaging the length of the unstable trend segment and the absolute value of the tangent slope. The trend is positively correlated with both the length of the unstable trend segment and the average value of the tangent slope.

6. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 1, characterized in that, The process of obtaining the activity moment includes: The moment within the reference time period of the current moment that meets the following conditions is determined as the active moment within the reference time period of the current moment: the updated respiratory rate is greater than the preset baseline of respiratory rate, and the respiratory depth is less than the preset baseline of respiratory depth.

7. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 1, characterized in that, The process of obtaining the rest effect includes: Determine the duration characteristics of the inactive period relative to the duration of the active period; Determine the extent to which the overall respiratory rate during the active period exceeds the overall respiratory rate during the inactive period; Based on the duration characteristics and the degree of excess of the overall respiratory rate, the rest effect of the user after the activity period is obtained; the rest effect is positively correlated with both the duration characteristics and the degree of excess of the overall respiratory rate.

8. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 1, characterized in that, The process of obtaining the activity intensity includes: Determine the respiratory intensity characteristics of the active period, wherein the respiratory intensity characteristics characterize the extent to which the maximum respiratory rate of the active period exceeds the maximum respiratory depth of the inactive period; The activity intensity of the activity period is obtained based on the duration of the activity period and the respiratory intensity characteristics; the activity intensity is positively correlated with both the duration of the activity period and the respiratory intensity characteristics.

9. The intelligent flow regulation system for the portable oxygen supply device for the elderly as described in claim 1, characterized in that, The process of obtaining the oxygen supply demand includes: Based on the activity intensity and rest effect during the activity period, an oxygen demand index for the activity period is obtained; the oxygen demand index is positively correlated with the activity intensity and negatively correlated with the rest effect. By integrating the oxygen demand indicators of each activity period within the reference time period at the current moment, the user's oxygen demand at the current moment can be obtained.

10. A portable oxygen supply device for the elderly, characterized in that, include: The main body of a portable oxygen supply device for the elderly, and the intelligent flow regulation system as described in any one of claims 1-9.

Images