Ventilator output pressure control method and system

Abstract

This invention discloses a method and system for controlling the output pressure of a ventilator. The method includes: acquiring and determining, based on respiratory signals within a unit sampling period, that an obstructive sleep apnea event has occurred within the unit sampling period; calculating, using fuzzy inference, a sleep state corresponding to the unit sampling period based on the respiratory signals; determining a ventilator output pressure compensation amount corresponding to the sleep state based on the obstructive sleep apnea event and the sleep state; and adjusting and controlling the corresponding ventilator pressure output based on the ventilator output pressure compensation amount. This method can dynamically adjust the ventilator output pressure for different sleep states, adjust and control the corresponding ventilator pressure output in real time, has strong adaptability, and can improve more personalized and precise ventilator pressure adjustment, fully utilize the support role of the ventilator, and achieve automated analysis and intelligence of the ventilator control system.

Description

Technical Field

[0001] This invention relates to the field of medical equipment technology, and in particular to a method and system for controlling the output pressure of a ventilator. Background Technology

[0002] Sleep is essential for human life and an indispensable part of maintaining good health. Normally, human sleep is divided into two main stages: REM (Rapid Eye Movement) and NREM (Non-rapid Eye Movement), which alternate throughout the night. However, many individuals experience breathing-related problems during sleep, such as OSA (Obstructive Sleep Apnea), a common sleep disorder. The main characteristic of OSA is the collapse of the upper airway, leading to obstruction of airflow through the mouth and nose, reduced or even complete apnea. This phenomenon occurs repeatedly at night, causing hypoxemia and disrupted sleep structure, which in turn leads to symptoms such as daytime sleepiness.

[0003] Currently, PAP (Positive Airway Pressure Device) is used to address obstructive sleep apnea (OSA) events during sleep. This method maintains upper airway patency by applying positive pressure, effectively preventing upper airway obstruction during sleep, reducing the occurrence of obstructive sleep apnea events, and thus improving sleep quality. However, while PAP is primarily used to detect and assist users with sleep disorders in breathing smoothly, it cannot monitor sleep stages or dynamically adjust ventilator pressure when obstructive sleep apnea events occur at different sleep stages. This reduces the adaptability and accuracy of pressure adjustment, resulting in poor ventilator adaptability. Summary of the Invention

[0004] One of the objectives of this invention is to provide a method for controlling the output pressure of a ventilator, so as to solve the problem that the existing technology cannot dynamically adjust the ventilator pressure according to the sleep state when an obstructive sleep apnea event occurs, so as to make the ventilator adaptively low.

[0005] One of the objectives of this invention is to provide a ventilator output pressure control system.

[0006] To achieve one of the above-mentioned objectives, the present invention provides a method for controlling the output pressure of a ventilator, comprising: acquiring and determining, based on a respiratory signal within a unit sampling period, an obstructive sleep apnea event occurring within the unit sampling period; calculating, based on the respiratory signal, a sleep state corresponding to the unit sampling period using a fuzzy inference method; determining a ventilator output pressure compensation amount corresponding to the sleep state based on the obstructive sleep apnea event and the sleep state; and adjusting and controlling the output pressure of the corresponding ventilator based on the ventilator output pressure compensation amount.

[0007] As a further improvement of one embodiment of the present invention, the respiratory signal includes at least one of a respiratory flow signal, a respiratory effort signal, and a respiratory pressure signal.

[0008] As a further improvement of one embodiment of the present invention, the sleep state includes at least one of the following: a waking state, a rapid eye movement (REM) state, a light sleep state, and a deep sleep state; wherein, in the waking state, the respiratory signal has a higher respiratory rate and a higher fluctuation amplitude; in the REM state, the respiratory signal has a higher respiratory rate and a lower fluctuation amplitude; in the light sleep state, the respiratory signal has a higher respiratory rate and a higher fluctuation amplitude than in the deep sleep state; and in the deep sleep state, the respiratory signal has a lower respiratory rate and a lower fluctuation amplitude.

[0009] As a further improvement of one embodiment of the present invention, the step of "acquiring and determining, based on the respiratory signal within a unit sampling period, that an obstructive sleep apnea event has occurred within the unit sampling period" specifically includes: acquiring and calculating the corresponding respiratory impedance value based on the pressure signal and flow signal of the respiratory gas within the unit sampling period; determining whether the respiratory impedance value is greater than a preset respiratory impedance threshold; if so, determining that an obstructive sleep apnea event has occurred within the unit sampling period.

[0010] As a further improvement of one embodiment of the present invention, the step of "calculating the sleep state corresponding to the unit sampling period by using fuzzy inference method based on the breathing signal" specifically includes: analyzing and extracting signal features corresponding to the breathing signal to obtain several breathing features; using fuzzy inference method to perform fuzzy inference operation on the several breathing features, and determining the sleep state corresponding to the unit sampling period based on the inference operation result.

[0011] As a further improvement of one embodiment of the present invention, the respiratory characteristics include at least one of respiratory depth, peak-to-average ratio, minute ventilation, respiratory rate, expiratory flow rate amplitude, and inspiratory flow rate amplitude.

[0012] As a further improvement to one embodiment of the present invention, the step of "using fuzzy inference method to perform fuzzy inference operation on the several respiratory features and determining the sleep state corresponding to the unit sampling period based on the inference operation result" specifically includes: constructing a first fuzzy inference model, inputting each respiratory feature into the first fuzzy inference model, performing a fuzzification operation on each respiratory feature using a membership function to obtain several corresponding membership values; calculating several fuzzy sets of respiratory features based on the Mamdani fuzzy inference algorithm, according to the several membership values ​​and a preset fuzzy inference rule base; performing a defuzzification operation on the several fuzzy sets of respiratory features using the centroid method to obtain a first fuzzy inference coefficient; and determining the sleep state corresponding to the unit sampling period based on the first fuzzy inference coefficient.

[0013] As a further improvement of one embodiment of the present invention, the step of "determining the sleep state corresponding to the unit sampling period based on the first fuzzy inference coefficient" specifically includes: determining whether the first fuzzy inference coefficient is greater than or equal to a preset inference coefficient threshold; if so, determining that the sleep state corresponding to the unit sampling period is a first sleep state, and determining whether the first sleep state is a REM sleep state based on the first fuzzy inference coefficient and the plurality of breathing features; if not, determining that the sleep state corresponding to the unit sampling period is a second sleep state, and determining whether the second sleep state is a light sleep state based on the first fuzzy inference coefficient and the plurality of breathing features; wherein, the first sleep state includes at least one of an awake state and a REM sleep state; the second sleep state includes at least one of a light sleep state and a deep sleep state.

[0014] As a further improvement of one embodiment of the present invention, the step of "determining whether the first sleep state is a rapid eye movement state based on the first fuzzy inference coefficient and the plurality of breathing features" specifically includes: constructing a second fuzzy inference model, using the first fuzzy inference coefficient and the plurality of breathing features as inputs to the second fuzzy inference model, performing fuzzy inference operations to obtain the second fuzzy inference coefficient; if the second fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, then the first sleep state is determined to be a rapid eye movement state; if the second fuzzy inference coefficient is less than the preset inference coefficient threshold, then the first sleep state is determined to be a wakeful state.

[0015] As a further improvement of one embodiment of the present invention, the step of "determining whether the second sleep state is a light sleep state based on the second fuzzy inference coefficient and the plurality of breathing features" specifically includes: constructing a third fuzzy inference model, taking the first fuzzy inference coefficient and the plurality of breathing features as inputs to the third fuzzy inference model, performing fuzzy inference operations to obtain the third fuzzy inference coefficient; if the third fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, then the second sleep state is determined to be a light sleep state; if the third fuzzy inference coefficient is less than the preset inference coefficient threshold, then the second sleep state is determined to be a deep sleep state.

[0016] As a further improvement to one embodiment of the present invention, the step of "determining the ventilator output pressure compensation amount corresponding to the sleep state based on the obstructive sleep apnea event and the sleep state" specifically includes: acquiring and statistically analyzing the apnea duration T of the obstructive sleep apnea event occurring within the current unit sampling period; acquiring and calculating the ventilator output pressure P and the maximum output pressure P within the current unit sampling period. max And the apnea-hypopnea index (AHI) corresponding to the current sleep state, the corresponding ventilator output pressure adjustment parameter P is calculated. PR ; Obtain the fuzzy inference coefficients (Fuzzy) within the current sampling period, and adjust the parameter P based on the fuzzy inference coefficients (Fuzzy) and the ventilator output pressure. PR The ventilator output pressure compensation quantity P was calculated. MI .

[0017] As a further improvement of one embodiment of the present invention, the fuzzy inference coefficient includes at least one of a first fuzzy inference coefficient, a second fuzzy inference coefficient, and a third fuzzy inference coefficient.

[0018] As a further improvement to one embodiment of the present invention, the step of "adjusting parameter P according to the ventilator output pressure" is... PR The ventilator output pressure compensation amount P is calculated using the fuzzy inference coefficients. MI Specifically, it includes:

[0019] The ventilator output pressure compensation amount P MI The ventilator output pressure adjustment parameter P PR The ventilator output pressure P during the current sampling period, and the maximum output pressure P. max The duration T of the obstructive sleep apnea event, the apnea-hypopnea index (AHI), and the fuzzy inference coefficient (Fuzzy) must at least satisfy the following:

[0020] P MI =a1*P PR+a2*Fuzzy

[0021]

[0022] Wherein, a1 represents the weight of the ventilator output pressure adjustment parameter, a2 represents the weight of different sleep states, x1 represents the weight of the apnea-hypopnea index (AHI), x2 represents the weight of the apnea duration T during the obstructive sleep apnea event, and x3 represents the weight of the ventilator output pressure.

[0023] As a further improvement to one embodiment of the present invention, the step of "adjusting and controlling the output of the corresponding ventilator pressure according to the ventilator output pressure compensation amount" specifically includes: obtaining the ventilator output pressure P within the current unit sampling cycle; controlling the positive airway pressure ventilation system to increase the ventilator pressure output based on the ventilator output pressure P, with the increase being the ventilator output pressure compensation amount P corresponding to the current sleep state. MI .

[0024] As a further improvement of one embodiment of the present invention, the detection process of obstructive sleep apnea events occurring within a unit sampling period and the identification process of sleep stages are both configured to be real-time.

[0025] To achieve one of the above-mentioned objectives, one embodiment of the present invention provides a ventilator output pressure control system, comprising: a memory and a processor, wherein the memory has a computer program that can run on the processor, and when the processor executes the computer program, it implements the steps of any one of the above-described ventilator output pressure control methods.

[0026] Compared with the prior art, the embodiments of the present invention have at least one of the following beneficial effects:

[0027] This invention employs a method for controlling and adjusting ventilator output pressure. By combining information from both obstructive sleep apnea events and sleep states, it determines the ventilator output pressure compensation amount. This allows for dynamic adjustment of the ventilator output pressure for different sleep states, achieving personalized sleep apnea treatment and improving adaptability. Specifically, by identifying respiratory signals, the degree of airflow can be directly reflected, improving the accuracy of obstructive sleep apnea event identification. Simultaneously, the use of fuzzy reasoning methods can intuitively represent the uncertainty and gradual changes in sleep states, thereby better assessing the user's overall sleep quality, improving the accuracy of sleep state detection, and providing a basis for subsequent dynamic adjustment of ventilator output pressure, increasing the accuracy and reliability of pressure adjustment. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of a ventilator structure according to one embodiment of the present invention.

[0029] Figure 2 This is a schematic diagram of a ventilator output pressure control system according to one embodiment of the present invention.

[0030] Figure 3 This is a schematic diagram of the steps of a ventilator output pressure control method according to an embodiment of the present invention.

[0031] Figure 4 This is a schematic diagram of respiratory signal fluctuations corresponding to NREM-OSA and REM-OSA events in one embodiment of the present invention.

[0032] Figure 5 This is a detailed schematic diagram of step S1 of the ventilator output pressure control method in one embodiment of the present invention.

[0033] Figure 6 This is a detailed schematic diagram of step S2 of the ventilator output pressure control method in one embodiment of the present invention.

[0034] Figure 7 This is a detailed schematic diagram of step S22 of the ventilator output pressure control method in one embodiment of the present invention.

[0035] Figure 8(a) is a schematic diagram of the fuzzy member function based on respiratory depth features in the ventilator output pressure control method according to an embodiment of the present invention.

[0036] Figure 8(b) is a schematic diagram of the membership function based on the peak-to-average ratio feature in the ventilator output pressure control method according to an embodiment of the present invention.

[0037] Figure 8(c) is a schematic diagram of the membership function based on minute ventilation characteristics in the ventilator output pressure control method according to an embodiment of the present invention.

[0038] Figure 8(d) is a schematic diagram of the membership function based on respiratory rate characteristics in the ventilator output pressure control method according to an embodiment of the present invention.

[0039] Figure 8(e) is a schematic diagram of the membership function based on the expiratory flow rate amplitude characteristics in the ventilator output pressure control method according to an embodiment of the present invention.

[0040] Figure 8(f) is a schematic diagram of the membership function based on the inspiratory flow rate amplitude characteristics in the ventilator output pressure control method according to an embodiment of the present invention.

[0041] Figure 9 This is a detailed schematic diagram of step S224 of the ventilator output pressure control method in one embodiment of the present invention.

[0042] Figure 10 This is a schematic diagram of the sleep state division results of the ventilator output pressure control method in one embodiment of the present invention.

[0043] Figure 11 This is a detailed schematic diagram of step S3 of the ventilator output pressure control method in one embodiment of the present invention.

[0044] Figure 12 This is a flowchart illustrating a preferred embodiment of the ventilator output pressure control method according to one embodiment of the present invention. Detailed Implementation

[0045] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0046] It should be noted that the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. In the description of specific embodiments of the present invention, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0047] Obstructive sleep apnea (OSA) events can occur during sleep, easily leading to sleep structure disruption and thus causing health problems. These OSA events can occur at different sleep stages. Some OSA users have a low overall apnea-hypopnea index (AHI), but a high AHI during rapid eye movement (REM) sleep. This complicates the clinical treatment of OSA events. Therefore, accurately identifying OSA events associated with different sleep states is crucial for determining the type of OSA event and developing appropriate treatment strategies.

[0048] Based on this, the present invention provides a ventilator output pressure control system 100, such as... Figure 1 and Figure 2As shown, several sensors 200 monitor and collect respiratory signals within the current unit sampling period, wherein the respiratory signals may include respiratory flow signals and respiratory pressure signals; based on the respiratory signals, the user's sleep state within the current unit sampling period is determined, as well as the dynamic changes in the ventilator output pressure compensation amount when an obstructive sleep apnea event occurs; the ventilator pressure support module 500 in the ventilator is controlled to increase the output pressure based on the current output pressure value of the ventilator, with the increase being the ventilator output pressure compensation amount.

[0049] in, Figure 1 Flow monitoring and pressure monitoring can be performed separately through... Figure 2 The sensors 200 (including flow sensors and pressure sensors) monitor the respiratory flow and pressure-flow signals collected by the sensors 200. The ventilator pressure control system 400 can obtain the respiratory flow and pressure-flow signals from the storage device 300 and calculate and control the ventilator pressure output under different sleep states based on the respiratory flow and pressure-flow signals. The ventilator pressure support module 500 can output the ventilator output pressure calculated by the ventilator pressure control system 400 to the user's airway.

[0050] Furthermore, the present invention also provides a method for controlling the output pressure of a ventilator, such as... Figure 3 As shown, the ventilator output pressure control method specifically includes the following steps:

[0051] Step S1: Acquire and determine, based on the respiratory signals within a unit sampling period, that an obstructive sleep apnea event has occurred within the unit sampling period;

[0052] Step S2: Based on the breathing signal, the sleep state corresponding to the unit sampling period is calculated using a fuzzy inference method;

[0053] Step S3: Based on the obstructive sleep apnea event and the sleep state, determine the ventilator output pressure compensation amount corresponding to the sleep state, and adjust and control the output of the corresponding ventilator pressure according to the ventilator output pressure compensation amount.

[0054] Thus, by combining information from both obstructive sleep apnea events and sleep states, the amount of ventilator output pressure compensation can be determined, enabling personalized adjustment of ventilator output pressure and improving adaptability. At the same time, by employing fuzzy reasoning, the uncertainty and variability of sleep stages can be intuitively mapped to definite sleep states, improving the accuracy and reliability of sleep state detection.

[0055] The respiratory signal may include at least one of a respiratory flow signal, a respiratory effort signal, and a respiratory pressure signal. (Continue to refer to...) Figure 2 As shown, when an obstructive pause event is detected, the ventilator output pressure compensation amount mentioned in step S3 can be increased by controlling the ventilator pressure support module 500 on the ventilator to increase the ventilator pressure output based on the current ventilator output pressure. The increase is the ventilator output pressure compensation amount, so that the human airway is clear and the obstructive pause event no longer occurs.

[0056] In addition, normal human sleep can go through two cycles: NREM (Non-Rapid Eye Movement) and REM (Rapid Eye Movement). This sleep cycle repeats throughout the night, with each cycle lasting approximately 90 to 120 minutes. Specifically, the NREM period can be further divided into light sleep and deep sleep.

[0057] In one implementation, an obstructive sleep apnea event occurring during NREM can be termed NREM-OSA; in another implementation, an obstructive sleep apnea event occurring during REM can be termed REM-OSA. Figure 4 As shown, REM-OSA and NREM-OSA are two types of obstructive sleep apnea (OSA), which are associated with different sleep stages and have different characteristics and clinical manifestations.

[0058] On the one hand, the weakening of upper airway muscle tone during REM makes the upper airway more prone to collapse; on the other hand, the EELV (End-Expiratory Lung Volume) value is significantly reduced during REM, which reduces or weakens the longitudinal tension of the airway, making it easier for airway collapse to occur. Therefore, obstructive sleep apnea events are more likely to occur during REM, and the upper airway resistance is greatest during this stage. As a result, the respiratory signal corresponding to REM-OSA fluctuates more than that of NREM-OSA.

[0059] Furthermore, in one embodiment, the sleep state may include a waking state, in which the corresponding respiratory signal has a high respiratory rate and a high fluctuation amplitude. In this state, various areas of the brain remain active and sensory functions are normal. In another embodiment, the sleep state may include a rapid eye movement (REM) state, in which the corresponding respiratory signal has a high respiratory rate and a low fluctuation amplitude. This state is accompanied by active brain activity, rapid eye movements, and vivid dreams. In yet another embodiment, the sleep state may include a deep sleep state, in which the corresponding respiratory signal has a low respiratory rate and a low fluctuation amplitude. In this state, it is difficult to be awakened by external disturbances. In yet another embodiment, the sleep state may include a light sleep state, in which the corresponding respiratory signal has a higher respiratory rate and fluctuation amplitude than that of the deep sleep state. In other words, the respiratory rate and amplitude of the light sleep state are both lower, but compared to the deep sleep state, the frequency and amplitude are higher. In this state, it is easily awakened by external noise or other disturbances.

[0060] like Figure 5 As shown, in one embodiment, step S1 may specifically include the following steps:

[0061] Step S11: Obtain and calculate the corresponding respiratory impedance value based on the pressure signal and flow rate signal of the respiratory gas within the unit sampling period;

[0062] Step S12: Determine whether the respiratory impedance value is greater than a preset respiratory impedance threshold.

[0063] If so, proceed to step S13 to determine if an obstructive sleep apnea event has occurred within the unit sampling period.

[0064] Thus, by calculating respiratory impedance using only pressure and flow signals, the degree of breathing smoothness can be quantitatively reflected, which is helpful for accurately determining whether obstructive sleep apnea has occurred. This method directly uses basic measurements for calculation, without complicated calculation processes or intermediate steps, and is simple and easy to implement.

[0065] The respiratory impedance value refers to the degree to which airflow is obstructed or restricted in the airway. When the airway is blocked or narrowed, airflow encounters resistance as it passes through the respiratory passage, leading to difficulty or obstruction of breathing. Obstructive sleep apnea is a sleep-disordered breathing condition characterized by repeated partial or complete obstruction of the airway during sleep, resulting in pauses in breathing or extremely shallow breathing. This can lead to insufficient oxygen supply and affect sleep quality. Therefore, calculating the respiratory impedance value can directly determine whether the user's airway is obstructed during the current sleep stage.

[0066] like Figure 6 As shown, in one embodiment, for step S2, the present invention provides a refined step, which may specifically include:

[0067] Step S21: Analyze and extract the signal features corresponding to the respiratory signal to obtain several respiratory features;

[0068] Step S22: Using fuzzy inference, perform fuzzy inference operations on the several respiratory features, and determine the sleep state corresponding to the unit sampling period based on the inference operation results.

[0069] Thus, by employing a fuzzy inference method suitable for processing fuzzy and uncertain respiratory signals, fuzzy and uncertain respiratory features can be mapped to definite sleep states, making the judgment of sleep states more accurate and reliable.

[0070] It should be noted that the amplitude and pattern of respiratory signal fluctuations differ significantly between awake, REM (Rapid Eye Movement) sleep, and non-REM sleep (i.e., light or deep sleep). Based on this, signal features corresponding to these respiratory signals can be analyzed and extracted to classify different sleep stages.

[0071] Specifically, in one embodiment, the respiratory characteristic may include respiratory depth. The respiratory depth can be determined by monitoring the variability of the difference between inspiratory and expiratory flow rates within a unit sampling period. Specifically, the respiratory depth can be determined by calculating the median or variance of the difference between inspiratory and expiratory flow rates within a unit sampling period.

[0072] In one embodiment, the respiratory characteristic may include a peak-to-average flow rate (PAFR). The PAFR can be calculated and determined by the change in the ratio of the peak inspiratory flow rate to the average inspiratory flow rate within a unit sampling period. When the fluctuation amplitude of the inspiratory flow rate signal is large, the PAFR is large, that is, there is a large variability in inspiratory flow rate during that time period.

[0073] In one embodiment, the respiratory characteristics may include minute ventilation. The minute ventilation can be determined by calculating the variability of ventilation within a unit sampling period, specifically by calculating the variance or entropy of ventilation within a unit sampling period.

[0074] In one embodiment, the respiratory characteristic may include respiratory frequency. The respiratory frequency can be determined by calculating the variability of the respiratory frequency within a unit sampling period, specifically by calculating the variance or entropy value of the respiratory frequency within a unit sampling period.

[0075] In one embodiment, the respiratory characteristic may include the expiratory flow rate amplitude. The expiratory flow rate amplitude can be determined by calculating the variability of the expiratory flow rate amplitude within a unit sampling period, specifically by calculating the variance or entropy value of the expiratory flow rate amplitude within a unit sampling period.

[0076] In one embodiment, the respiratory characteristic may include the inspiratory flow rate amplitude. The inspiratory flow rate amplitude can be determined by calculating the variability of the inspiratory flow rate amplitude within a unit sampling period, specifically by calculating the average or median of the inspiratory flow rate amplitude within a unit sampling period.

[0077] In a preferred embodiment, the above six embodiments can be used in combination. Of course, other combinations are not excluded, and adjustments can be made freely according to the actual situation.

[0078] Furthermore, the essence of the fuzzy inference method is a computational process that maps a given input space to a specific output space using fuzzy logic. In other words, the fuzzy inference method can design different fuzzy sets and fuzzy rules based on different input features, thereby enabling corresponding inference based on different input features to obtain an output value. In this invention, by employing the fuzzy inference method, the scoring values ​​or probability values ​​of different sleep states can be analyzed and inferred, thereby determining the sleep state.

[0079] Based on this, such as Figure 7 As shown, in one embodiment, step S22 may specifically include the following steps:

[0080] Step S221: Construct a first fuzzy inference model, input each breathing feature into the first fuzzy inference model, and perform fuzzification operation on each breathing feature using a membership function to obtain several corresponding membership values;

[0081] Step S222: Based on the Mamdani fuzzy inference algorithm, several fuzzy sets of breathing features are calculated according to the membership values ​​and the preset fuzzy inference rule base.

[0082] Step S223: Perform defuzzification operation on the several fuzzy sets of breathing features using the centroid method to obtain the first fuzzy inference coefficients;

[0083] Step S224: Determine the sleep state corresponding to the unit sampling period based on the first fuzzy inference coefficient.

[0084] Thus, by employing a membership function, the uncertainty and ambiguity in respiratory features can be mapped to specific linguistic descriptions. Then, the Mamdani fuzzy inference algorithm is used for multi-rule inference to obtain the corresponding fuzzy set. Finally, the centroid method is used to defuzzify and determine the final sleep state. The centroid method can consider the overall features of the fuzzy set, resulting in higher accuracy.

[0085] The membership function determines the degree to which the input data belongs to an appropriate fuzzy set. In one embodiment, the membership function can be a combination of membership functions, such as the triangular membership function shown in formula (1) and the trapezoidal membership function shown in formula (2). The combination of membership functions can perform fuzzification operations on the breathing features mentioned above. Of course, this invention does not exclude other membership functions, which can be freely selected according to actual needs. This invention does not impose specific limitations.

[0086]

[0087]

[0088] Specifically, the membership value mentioned in step S221 can be understood as the degree to which each respiratory feature belongs to a certain appropriate fuzzy set, which can be divided into three levels: low, medium, and high. For each respiratory feature mentioned above, different fuzzy levels can be established according to different sleep states, as can be referred to... Figures 8(a) to 8(f) The diagram shows membership functions corresponding to different respiratory features; the corresponding output value is determined based on the input value of each respiratory feature, and the output value may include multiple values.

[0089] For example, referring to Figure 8(a), if the input value of the breathing depth is 3, then the value is less than 5, and the corresponding output value is "low", so the membership value of the breathing depth is "low"; if the input value of the breathing depth is 7.5, then the corresponding output value is "low, high", that is, the membership value of the breathing depth is "low, high".

[0090] Furthermore, the Mamdani fuzzy inference algorithm is a decision-making method based on fuzzy inference. The preset fuzzy inference rule base can be constructed independently based on experience and practical needs, combining all fuzzy rules to describe the relationship between input and output variables. In this invention, when executing the fuzzy inference algorithm on each breathing feature, the breathing feature can be used as input, and inference is performed according to the preset fuzzy inference rule base to obtain the inference result. Thus, using this algorithm can better handle the uncertainty and fuzziness of input data, its inference results are easy to interpret, and it has no complex calculation process or intermediate steps, therefore exhibiting high computational efficiency.

[0091] The centroid method can transform a fuzzy set into a uniquely determined numerical value. This method can fully consider the overall information of the fuzzy set and output a stable value. That is, in this invention, the first fuzzy inference coefficient can be output.

[0092] Furthermore, such as Figure 9 As shown, in one embodiment, step S224 may specifically include the following steps:

[0093] Step S2241: Determine whether the first fuzzy inference coefficient is greater than or equal to a preset inference coefficient threshold.

[0094] If so, proceed to step S2242A, determine that the sleep state corresponding to the unit sampling period is the first sleep state, and determine whether the first sleep state is a rapid eye movement state based on the first fuzzy inference coefficient and the several breathing features.

[0095] If not, proceed to step S2242B, determine that the sleep state corresponding to the unit sampling period is the second sleep state, and determine whether the second sleep state is a light sleep state based on the first fuzzy inference coefficient and the several breathing features.

[0096] In this way, by making full use of the correspondence between the inference coefficient and the sleep state, setting an inference coefficient threshold for division, and further subdividing different sleep sub-states, different sleep states can be effectively distinguished, and the judgment results are more accurate and reliable.

[0097] The first sleep state may include at least one of a wakeful state and a rapid eye movement (REM) state; the second sleep state may include at least one of a light sleep state and a deep sleep state.

[0098] It should be noted that the first sleep state (i.e., during NREM) and the second sleep state (i.e., during REM) can be easily distinguished by the fluctuation amplitude and frequency of the respiratory signal within a unit sampling period, because the fluctuation amplitude and frequency of the respiratory signal differ significantly between the two periods. The sub-sleep states of the first or second sleep state can be further determined based on the first fuzzy inference coefficients.

[0099] Based on this, in one embodiment, the part of step S2242A that "determines whether the first sleep state is a REM sleep state based on the first fuzzy inference coefficient and the plurality of breathing features" can specifically include the following steps:

[0100] Step S2242A1: Construct a second fuzzy inference model, take the first fuzzy inference coefficients and several breathing features as inputs to the second fuzzy inference model, perform fuzzy inference operations, and obtain the second fuzzy inference coefficients;

[0101] Step S2242A2: If the second fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, then the first sleep state is determined to be a rapid eye movement state.

[0102] Step S2242A3: If the second fuzzy inference coefficient is less than the preset inference coefficient threshold, then the first sleep state is determined to be an awake state.

[0103] In this way, by making full use of the combination of two fuzzy reasoning models, more complex reasoning and decision-making can be achieved, and the judgment results are more accurate and reliable.

[0104] In another implementation, the part of step S2242B that "determines whether the second sleep state is a light sleep state based on the second fuzzy inference coefficient and the plurality of breathing characteristics" can specifically include the following steps:

[0105] Step S2242B1: Construct a third fuzzy inference model, using the first fuzzy inference coefficients and several breathing features as inputs to the third fuzzy inference model, and perform fuzzy inference operations to obtain the third fuzzy inference coefficients.

[0106] Step S2242B2: If the third fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, then the second sleep state is determined to be a light sleep state.

[0107] Step S2242B3: If the third fuzzy inference coefficient is less than the preset inference coefficient threshold, then the second sleep state is determined to be a deep sleep state.

[0108] In this way, by making full use of the combination of two fuzzy reasoning models, more complex reasoning and decision-making can be achieved, and the judgment results are more accurate and reliable.

[0109] The first fuzzy inference coefficient serves as the input parameter for both the second and third fuzzy inference models. In a preferred embodiment, several respiratory features corresponding to the second fuzzy inference model may include minute ventilation, respiratory rate, and the first fuzzy inference coefficient; several respiratory features corresponding to the third fuzzy inference model may include inspiratory flow rate amplitude, expiratory flow rate amplitude, and the first fuzzy inference coefficient.

[0110] It should be noted that the first fuzzy inference coefficient can be used to distinguish between the first sleep state and the second sleep state; the second fuzzy inference coefficient can be used to distinguish between the sub-states of the first sleep state, namely, distinguishing between REM sleep and wakefulness; and the third fuzzy inference coefficient can be used to distinguish between the sub-states of the second sleep state, namely, distinguishing between light sleep and deep sleep.

[0111] Furthermore, steps S2242A1 to S2242A3 and steps S2242B1 to S2242B3 can be executed intermittently, either in parallel or sequentially. Step S2242A2 can be executed after step S2242A3, and step S2242B2 can also be executed after step S2242B3.

[0112] like Figure 10 As shown, in one embodiment, by employing the first to the third fuzzy inference models, the entire sleep stage can be divided into different sleep states and durations, which facilitates subsequent real-time monitoring of whether obstructive sleep apnea events occur in different sleep states.

[0113] like Figure 11 As shown, in one embodiment, for the part of step S3 described as "determining the ventilator output pressure compensation amount corresponding to the sleep state based on the obstructive sleep apnea event and the sleep state", the present invention provides detailed steps, which may specifically include:

[0114] Step S311: Obtain and count the duration T of the obstructive sleep apnea event occurring within the current unit sampling period;

[0115] Step S312: Obtain and calculate the ventilator output pressure P and maximum output pressure P within the current unit sampling cycle. max And the apnea-hypopnea index (AHI) corresponding to the current sleep state, the corresponding ventilator output pressure adjustment parameter P is calculated. PR ;

[0116] Step S313: Obtain the fuzzy inference coefficients (Fuzzy) within the current sampling period, and adjust the ventilator output pressure parameter P based on the fuzzy inference coefficients (Fuzzy) and the ventilator output pressure. PR The ventilator output pressure compensation quantity P was calculated. MI .

[0117] In this way, by comprehensively considering multiple parameters within the current unit sampling period, the ventilator output pressure compensation can be made more accurate and reasonable. Furthermore, by combining the inference coefficient corresponding to the current sleep state, the ventilator output pressure compensation can be dynamically adjusted, exhibiting strong adaptability. This can improve the more personalized and refined adjustment of the ventilator output pressure, thereby achieving the level of automated analysis and intelligence of the ventilator.

[0118] Wherein, the ventilator output pressure P refers to the current therapeutic pressure output by the ventilator; the maximum output pressure P maxIt is the maximum output pressure provided by the ventilator; the Apnea-Hypopnea Index (AHI) can be used to assess the severity of sleep apnea, which is determined by calculating the total number of apnea and hypopnea events that occur per hour.

[0119] Furthermore, the fuzzy inference coefficients can include at least one of a first fuzzy inference coefficient, a second fuzzy inference coefficient, and a third fuzzy inference coefficient. Specifically, the first fuzzy inference model can output the first fuzzy inference coefficient; the second fuzzy inference model can output the second fuzzy inference coefficient; and the third fuzzy inference model can output the third fuzzy inference coefficient.

[0120] Furthermore, based on the sleep state during which the obstructive sleep apnea event occurred within the current sampling period, the corresponding apnea-hypopnea index (AHI) is determined, and subsequently, the corresponding ventilator output pressure adjustment parameter P is determined. PR And the corresponding ventilator output pressure compensation amount P MI The ventilator output pressure compensation amount P MI The ventilator output pressure adjustment parameter P PR The ventilator output pressure P during the current sampling period, and the maximum output pressure P. max The duration T of the obstructive sleep apnea event, the apnea-hypopnea index (AHI), and the fuzzy inference coefficient (Fuzzy) must at least satisfy formulas (3) and (4):

[0121] P MI =a1*P PR +a2*Fuzzy (3)

[0122]

[0123] Wherein, a1 represents the weight of the ventilator output pressure adjustment parameter, a2 represents the weight of different sleep states, x1 represents the weight associated with the apnea-hypopnea index (AHI), x2 represents the weight of the apnea duration T during the obstructive sleep apnea event, and x3 represents the weight of the ventilator output pressure.

[0124] It should be noted that the fuzzy inference coefficient can be dynamically adjusted based on the sleep state in which the obstructive sleep apnea event occurs within the current sampling period. Specifically, in one embodiment, if the sleep state in which the obstructive sleep apnea event occurs within the current sampling period is deep sleep, then the fuzzy inference coefficient can be a third fuzzy inference coefficient; in another embodiment, if the sleep state in which the obstructive sleep apnea event occurs within the current sampling period is REM sleep, then the fuzzy inference coefficient can be a second fuzzy inference coefficient.

[0125] Based on the ventilator output pressure compensation amount during the current sleep state, the ventilator pressure output can be controlled and adjusted. Therefore, the part of step S3, "adjusting and controlling the corresponding ventilator pressure output according to the ventilator output pressure compensation amount," can specifically include the following steps:

[0126] Step S321: Obtain the ventilator output pressure P within the current sampling cycle;

[0127] Step S322: Control the positive airway pressure ventilation system to increase the ventilator's output pressure P based on the ventilator's output pressure, with the increase being the ventilator output pressure compensation amount P corresponding to the current sleep state. MI .

[0128] In this way, the ventilator pressure output can be dynamically adjusted according to the current ventilator pressure, resulting in high flexibility and good treatment effect.

[0129] The positive airway pressure (PAP) system may include at least one of continuous positive airway pressure (CPAP) and automatic positive airway pressure (APAP).

[0130] Specifically, the CPAP system can be used as an effective means of treating obstructive sleep apnea events. The user connects to a positive pressure source by wearing a nasal mask or face mask, and supports and stabilizes the upper airway by controlling and adjusting the pressure output of the ventilator, thereby eliminating upper airway obstruction. This can effectively eliminate obstructive sleep apnea events. The CPAP system can dynamically change the ventilator's output pressure based on the user's current respiratory data. Preferably, the CPAP system can be an APAP system. Furthermore, the ventilator output pressure can be the pressure that the ventilator uses to input into the user's airway to relieve or treat obstructive sleep apnea events.

[0131] It should be noted that regardless of the user's sleep state, the positive airway pressure system always has an initial ventilator output pressure P, and the ventilator output pressure compensation P corresponding to the current sleep state... MI It can be added to the initial ventilator output pressure P to generate something like P+P. MI The ventilator output pressure. Furthermore, the therapeutic pressure output of the positive airway pressure system does not increase indefinitely. When obstructive sleep apnea events are relieved by the ventilator output pressure, the respiratory rate of the patient will tend towards a normal respiratory rate, and the positive airway pressure system will be controlled to reduce the ventilator pressure output. When the respiratory rate remains normal, the positive airway pressure system will be controlled to gradually reduce the ventilator pressure output until it reaches the minimum ventilator output pressure.

[0132] Furthermore, the process of sleep state segmentation and the process of identifying obstructive sleep apnea events can be executed in parallel, or the sleep state segmentation process can be executed first, and then the obstructive sleep apnea event can be identified based on the segmentation results; alternatively, the obstructive sleep apnea event identification process can be executed first, and then the sleep state corresponding to the event can be determined. This invention does not impose specific limitations on these methods.

[0133] It should be noted that the detection process for obstructive sleep apnea events and the identification process for sleep stages within a unit sampling period are both configured to be real-time. In other words, by continuously monitoring respiratory signals and utilizing fuzzy inference methods and models, sleep states are determined in real time, and the occurrence of OSA events is detected in real time. Simultaneously, based on the detection results, the output pressure of the ventilator is adjusted promptly to ensure the user receives optimal respiratory support. This achieves real-time, continuous monitoring and adjustment, improving the ventilator's adaptability and efficiency.

[0134] The various embodiments, examples, or specific examples provided by this invention can be combined with each other to ultimately form multiple better embodiments.

[0135] For example, such as Figure 12 A flowchart illustrating a preferred embodiment of a ventilator output pressure control method is shown. The following will be combined with... Figure 12 This section summarizes the processing procedure of the preferred embodiment.

[0136] Within a unit sampling period, respiratory flow and respiratory pressure signals are collected by sensors; the collected signal data are processed to calculate several respiratory characteristics.

[0137] A first fuzzy inference model is constructed and based on it. The several breathing features are input into the first fuzzy inference model to perform fuzzy inference operations, so as to distinguish the first sleep state and the second sleep state.

[0138] A second fuzzy inference model and a third fuzzy inference model are constructed. The first fuzzy inference coefficient and the corresponding breathing features output by the first fuzzy inference model are respectively input into the second fuzzy inference model and the third fuzzy inference model. The sub-sleep states (wakefulness and rapid eye movement) and the second fuzzy inference coefficient in the first sleep state, as well as the sub-sleep states (including light sleep and deep sleep) and the third fuzzy inference coefficient in the second sleep state are calculated.

[0139] Based on the respiratory signal, it is identified whether obstructive sleep apnea events occur in different sleep states. If so, the ventilator output pressure compensation amount in different sleep states is calculated based on the first fuzzy inference coefficient, the second fuzzy inference coefficient, or the third fuzzy inference coefficient, and then the corresponding ventilator output pressure is obtained.

[0140] The present invention also provides a ventilator output pressure control system, comprising: a memory and a processor, wherein the memory has a computer program that can run on the processor, and when the processor executes the computer program, it implements the steps of any of the above-described ventilator output pressure control methods.

[0141] In summary, the ventilator output pressure control method provided by this invention, by acquiring and calculating the respiratory impedance value based on the respiratory signal within a unit sampling period, can quantitatively reflect the smoothness of breathing, which is beneficial for accurately determining whether obstructive sleep apnea has occurred. Furthermore, by employing fuzzy inference, uncertain respiratory characteristics or respiratory signals can be mapped to a definite sleep state, resulting in high accuracy and reliability. Finally, based on obstructive sleep apnea events occurring in different sleep stages, the corresponding ventilator output pressure is adjusted and controlled in real time, exhibiting strong adaptability. This allows for more personalized and precise adjustment of the ventilator output pressure, fully leveraging the support role of the ventilator and achieving automated analysis and intelligence in the ventilator control system.

[0142] It should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0143] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A ventilator output pressure control system, characterized by, include: A memory and a processor, the memory having a computer program executable on the processor, the processor executing the computer program to perform the following steps: Acquire and determine, based on the respiratory signals within a unit sampling period, that an obstructive sleep apnea event has occurred within the unit sampling period; Analyze and extract the signal features corresponding to the respiratory signal to obtain several respiratory features; Construct a first fuzzy inference model, input each breathing feature into the first fuzzy inference model, and perform a fuzzification operation on each breathing feature using a membership function to obtain several corresponding membership values; Based on the Mamdani fuzzy inference algorithm, several fuzzy sets of breathing features are calculated according to the aforementioned membership values ​​and a preset fuzzy inference rule base. The centroid method is used to perform a defuzzification operation on the several fuzzy sets of respiratory features to obtain the first fuzzy inference coefficients; Determine whether the first fuzzy inference coefficient is greater than or equal to a preset inference coefficient threshold; If so, the sleep state corresponding to the unit sampling period is determined to be the first sleep state, and a second fuzzy inference model is constructed. The first fuzzy inference coefficient and several breathing features are used as inputs to the second fuzzy inference model, and fuzzy inference operation is performed to obtain the second fuzzy inference coefficient. If the second fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, the first sleep state is determined to be a rapid eye movement state. If the second fuzzy inference coefficient is less than the preset inference coefficient threshold, the first sleep state is determined to be a wakeful state. If not, the sleep state corresponding to the unit sampling period is determined to be the second sleep state, and a third fuzzy inference model is constructed. The first fuzzy inference coefficient and several breathing features are used as inputs to the third fuzzy inference model, and fuzzy inference operation is performed to obtain the third fuzzy inference coefficient. If the third fuzzy inference coefficient is greater than or equal to the preset inference coefficient threshold, then the second sleep state is determined to be a light sleep state. If the third fuzzy inference coefficient is less than the preset inference coefficient threshold, then the second sleep state is determined to be a deep sleep state; Based on the obstructive sleep apnea event and the sleep state, determine the ventilator output pressure compensation amount corresponding to the sleep state, and adjust and control the output of the corresponding ventilator pressure according to the ventilator output pressure compensation amount.

2. The ventilator output pressure control system of claim 1, wherein, The respiratory signal includes at least one of the following: respiratory flow signal, respiratory effort signal, and respiratory pressure signal.

3. The ventilator output pressure control system of claim 1, wherein, The sleep state includes at least one of the following: wakefulness, REM sleep, light sleep, and deep sleep; wherein, in the wakefulness state, the respiratory signal has a high respiratory rate and a high fluctuation amplitude; in the REM sleep state, the respiratory signal has a high respiratory rate and a low fluctuation amplitude; in the light sleep state, the respiratory signal has a higher respiratory rate and a higher fluctuation amplitude than in the deep sleep state; and in the deep sleep state, the respiratory signal has a lower respiratory rate and a lower fluctuation amplitude.

4. The ventilator output pressure control system according to claim 1, characterized in that, When the processor executes the computer program, the step of determining whether an obstructive sleep apnea event occurs within the unit sampling period specifically includes: The corresponding respiratory impedance value is calculated based on the pressure signal and flow rate signal of the respiratory gas within the unit sampling period; Determine whether the respiratory impedance value is greater than a preset respiratory impedance threshold; If so, an obstructive sleep apnea event is determined to have occurred within the unit sampling period.

5. The ventilator output pressure control system according to claim 1, characterized in that, The respiratory characteristics include at least one of the following: respiratory depth, peak-to-average ratio, minute ventilation, respiratory rate, expiratory flow rate amplitude, and inspiratory flow rate amplitude.

6. The ventilator output pressure control system according to claim 1, characterized in that, When the processor executes the computer program, the step of determining the ventilator output pressure compensation amount corresponding to the sleep state based on the obstructive sleep apnea event and the sleep state specifically includes: Obtain and count the duration of obstructive sleep apnea events occurring within the current sampling period. ; Obtain and calculate the ventilator output pressure within the current sampling cycle. and maximum output pressure and the apnea-hypopnea index corresponding to the current sleep state. The corresponding ventilator output pressure adjustment parameters are calculated. ; Obtain the fuzzy inference coefficients within the current sampling period. According to the fuzzy inference coefficients and the ventilator output pressure adjustment parameters The ventilator output pressure compensation amount was calculated. .

7. The ventilator output pressure control system according to claim 6, characterized in that, When the processor executes the computer program, it implements the step of using the fuzzy inference coefficients. and the ventilator output pressure adjustment parameters The ventilator output pressure compensation amount was calculated. The steps specifically include: The ventilator output pressure compensation amount The ventilator output pressure adjustment parameters The ventilator output pressure within the current unit sampling cycle The maximum output pressure The duration of the obstructive sleep apnea event. The apnea-hypopnea index and the fuzzy inference coefficients At least the following conditions must be met: ； in, This indicates the weight of the ventilator output pressure adjustment parameter. Representing the weights of different sleep states, The apnea-hypopnea index indicates the apnea-hypopnea index. The weight, Indicates the duration of the obstructive sleep apnea event. The weight, This indicates the weight of the output pressure of the ventilator.

8. The ventilator output pressure control system according to claim 1, characterized in that, When the processor executes the computer program, the step of adjusting and controlling the output pressure of the corresponding ventilator based on the ventilator output pressure compensation amount specifically includes: Obtain the ventilator output pressure within the current sampling cycle. ; The positive airway pressure ventilation system controls the output pressure of the ventilator. Based on this, increase the pressure output of the ventilator by the amount of pressure compensation corresponding to the current sleep state. .

9. The ventilator output pressure control system according to claim 1, characterized in that, The detection process for obstructive sleep apnea events and the identification process for sleep stages within a unit sampling period are both configured to be real-time.

Images