A late rehabilitation respiratory training system for neurology nursing

Abstract

The present application relates to the technical field of respiratory training, in particular to a post-rehabilitation respiratory training system for neurology nursing, which comprises an airflow collection module, a cycle division module, a lagging segmentation module, a deviation grading module and a rhythm control module.In the present application, the start points of inhalation and exhalation are identified and the cycle is divided through positive and negative flow rate conversion, and at the same time, time normalization processing is performed, so that the breathing behavior under different states has a unified comparison scale, the phase difference is calculated in combination with the reference rhythm phase, and the interval is divided according to the change trend, so that the local lag and rhythm drift in the breathing process can be identified in detail, the interval average phase difference is continuously cycled and the deviation grade is divided, so that the rhythm evaluation takes into account stability and trend, the inhalation and exhalation time ratio is reconstructed according to the deviation result, and a control signal is output, so that the training rhythm and the actual breathing performance form a dynamic match, which helps to improve the training accuracy and individual adaptation level.

Description

Technical Field

[0001] This invention relates to the field of respiratory training technology, and in particular to a respiratory training system for post-rehabilitation in neurological nursing. Background Technology

[0002] The field of respiratory training technology involves methods and devices for restoring pulmonary ventilation capacity and training respiratory muscle function in individuals with respiratory dysfunction. Its core aspects include adjusting inspiratory and expiratory resistance, controlling respiratory rate and rhythm, monitoring airflow and vital capacity parameters, and guiding the training process based on the patient's condition. The overall technology employs airflow channel structure design, resistance adjustment component settings, respiratory rhythm prompts, and sensor-based respiratory parameter acquisition methods to provide users with phased respiratory training and data recording. Among these, traditional respiratory training systems for post-rehabilitation in neurological nursing refer to devices or systems applied to the recovery phase of patients with neurological diseases. These systems assist patients in respiratory function training through nursing interventions, addressing the problems of decreased respiratory muscle control and insufficient spontaneous breathing coordination in patients after nerve injury. Traditional methods use fixed-resistance respirators combined with verbal prompts for inspiratory and expiratory training, or record respiratory data using simple flow meters and pressure gauges, combined with timed training schedules and manual valve adjustments to change airflow resistance to complete the training process.

[0003] Existing technologies, in practical applications, largely rely on fixed rhythms and phased observations. Breathing processes are recorded using only limited discrete data or simple time-segment divisions, making it difficult to demonstrate the continuous correlation between flow rate changes and airway pressure response. This results in inaccurate localization of respiratory coordination problems, and training adjustments often remain at a holistic level. Verbal cues and artificial rhythms are prone to misalignment with actual breathing movements, especially in cases of slow response or rhythm drift, potentially leading to trainees chasing the beat or exhibiting uneven breathing. Resistance adjustment relies on manual operation and intermittent observation, making it difficult to promptly identify short-term anomalies or gradually accumulating deviations, posing a risk of delayed intervention. While simple instruments provide numerical values, they lack multi-period trend analysis, making it difficult to distinguish between continuous instability and occasional fluctuations, affecting judgment accuracy and potentially leading to training load mismatch, increased fatigue, and unstable rehabilitation progress. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a respiratory training system for post-rehabilitation in neurological nursing. The technical solution is as follows. On the one hand, a respiratory training system for post-rehabilitation in neurological nursing is provided, the system comprising:

[0005] The airflow acquisition module collects the flow velocity value output by the flow sensor and the airway pressure value output by the pressure sensor. It aligns the flow velocity value and the airway pressure value in chronological order, determines the positive and negative changes in flow velocity and the rise and fall of pressure, combines and marks the flow velocity state and pressure state at each time point, records the continuous changes, and generates a respiratory state time series.

[0006] Based on the respiratory state time series, the cycle division module calls the rhythm frequency of the rhythm prompting device, identifies the time point when the flow rate changes from negative to positive as the start of inhalation and the time point when the flow rate changes from positive to negative as the start of exhalation, divides the time interval between adjacent inhalation start points, performs time normalization processing, and marks the inhalation and exhalation segments to generate a respiratory cycle labeled sequence.

[0007] The lag segmentation module calls the normalized phase value of the reference time in the rhythm curve storage unit according to the respiratory cycle annotation sequence, calculates the phase difference value at time points to obtain the phase difference value sequence, determines the trend of phase difference value change between adjacent time points and records the direction change position as the segment boundary, divides the interval and calculates the average phase difference value of the interval and marks the time range of the interval and the average phase difference value of the interval, and generates a respiratory lag segment set.

[0008] The deviation grading module, based on the set of respiratory lag segments and combined with the operating time window of the resistance adjustment component, summarizes the average phase difference value of the interval to obtain the periodic phase deviation value, sorts the continuous periodic phase deviation values ​​and divides the grade interval, and matches the current periodic phase deviation value to obtain the rhythm deviation grade identifier.

[0009] The rhythm control module, based on the rhythm deviation level identifier, calls the airflow regulating valve assembly and the feedback display to set the inspiratory phase time ratio and the expiratory phase time ratio, reconstructs the segmented time range according to the ratio, arranges them to form a rhythm time sequence, and converts the control signal to generate a respiratory rhythm control sequence.

[0010] As a further aspect of the present invention, the respiratory state time series includes a flow rate symbol sequence, a pressure change label sequence, and a timestamp index sequence; the respiratory cycle labeling sequence includes a cycle boundary marker set, an inspiratory segment label set, an expiratory segment label set, and a normalized time axis; the respiratory lag segment set includes a segment interval list, an interval average phase difference set, a segment boundary index set, and an interval time range set; the rhythm deviation level identifier includes a deviation level label, an interval mapping identifier, and a level threshold interval; the respiratory rhythm control sequence includes a control time series, a valve drive signal set, and a display feedback parameter set.

[0011] As a further aspect of the present invention, the process of determining the positive and negative change state of the flow velocity includes comparing the flow velocity value output by the flow sensor with a preset zero reference threshold. When the flow velocity value output by the flow sensor is greater than the preset zero reference threshold, it is marked as a positive flow velocity state, and when the flow velocity value output by the flow sensor is less than or equal to the preset zero reference threshold, it is marked as a negative flow velocity state.

[0012] As a further aspect of the present invention, the process of determining the pressure rise and fall state includes calculating the difference between the airway pressure values ​​output by the pressure sensor at two consecutive time points. When the airway pressure value output by the pressure sensor at the later time point is greater than the airway pressure value output by the pressure sensor at the earlier time point, it is marked as a pressure rise state. When the airway pressure value output by the pressure sensor at the later time point is less than or equal to the airway pressure value output by the pressure sensor at the earlier time point, it is marked as a pressure fall state.

[0013] As a further aspect of the present invention, the airflow acquisition module includes:

[0014] The flow rate and pressure acquisition submodule acquires the flow rate value output by the flow sensor inside the breathing training mask and the airway pressure value output by the pressure sensor inside the airway interface tubing. The flow rate and airway pressure values ​​are recorded synchronously according to the sampling time. The direction of the flow rate value at the sampling point is determined to form a positive or negative status indicator, and the trend of the change of the airway pressure value is determined to form an up or down status indicator, generating a flow and pressure synchronous sampling matrix.

[0015] The timing alignment processing submodule reconstructs the time sequence based on the sampling time identifier in the flow-pressure synchronization sampling matrix, processes sampling points with time interval differences into a unified time step and forms a continuous sequence, calls the flow velocity status identifier and pressure status identifier in the flow-pressure synchronization sampling matrix for position matching, performs corresponding verification of the flow velocity status and pressure status at each time point and completes the combined marking to obtain the flow-pressure status alignment sequence.

[0016] The state sequence generation submodule sequentially encodes the combined markers of flow velocity and pressure states at time points in the flow-pressure state alignment sequence, converts each time point combined marker into a unified state identifier value, and arranges them in chronological order to form a sequence structure. It then performs a continuity check on the state identifiers in the sequence and fills in any missing identifiers to generate a respiratory state time sequence.

[0017] As a further aspect of the present invention, the period division module includes:

[0018] The start and end point identification submodule, based on the breathing state time series, calls the current rhythm frequency of the breathing training rhythm prompting device, detects the direction change of the flow rate state at continuous time points, determines the start of inhalation when the flow rate state changes from negative to positive, and determines the start of exhalation when the flow rate state changes from positive to negative, and records the time index corresponding to the start point in chronological order to form a set of start and end point identifiers, thus obtaining the breathing start and end point index sequence.

[0019] The cycle interval construction submodule performs adjacent matching based on the inspiratory start time index in the respiratory start and end point index sequence, divides the time range between adjacent inspiratory start points into independent cycle intervals, calls the expiratory start time index for location within the cycle interval, confirms the boundaries of the cycle intervals and establishes interval correspondence, and obtains a set of respiratory cycle intervals.

[0020] The cycle labeling generation submodule performs time normalization processing on the time points within the cycle based on the time range of the cycle interval in the respiratory cycle interval set, and divides the time points into stages according to the positions of the inspiratory start point and expiratory start point in the respiratory start and end point index sequence. The time points are marked as inspiratory segments or expiratory segments and arranged in chronological order to generate a respiratory cycle labeling sequence.

[0021] As a further aspect of the present invention, the hysteresis segmentation module includes:

[0022] The phase difference construction submodule calls the reference time normalized phase value in the rehabilitation breathing training rhythm curve storage unit according to the respiratory cycle annotation sequence, calculates the difference between the time point annotation phase and the reference phase, forms a phase difference record arranged in time order, and stores the time point phase difference value with the corresponding time index to obtain the phase difference sequence.

[0023] The boundary recognition processing submodule determines the trend based on the relationship between the phase difference changes of adjacent time points in the phase difference sequence, compares the consistency of the phase difference direction of continuous time points, marks the position where the change direction changes as the segment boundary, and records the time index corresponding to the boundary and arranges them in chronological order to obtain the segment boundary index sequence.

[0024] The segmentation result generation submodule divides the single-cycle time range into intervals based on the segmentation boundary index sequence, defines the time intervals between adjacent boundaries as independent intervals, and calls the phase difference value sequence within the interval for statistical processing to form the interval average phase difference value. The interval time range and the interval average phase difference value are marked accordingly and arranged in chronological order to generate a respiratory lag segment set.

[0025] As a further aspect of the present invention, the deviation grading module includes:

[0026] The deviation summary calculation submodule, based on the set of respiratory lag segments, calls the current operating time window of the respiratory resistance adjustment component to summarize and calculate the average phase difference value of the segment interval, integrates the phase difference value within the same cycle in chronological order, and combines it with the corresponding segment data within the time window to perform periodic aggregation, forming a periodic hierarchical phase deviation record, and obtaining the periodic phase deviation value.

[0027] The interval division processing submodule sorts the continuous periodic phase deviation values ​​according to the periodic phase deviation values, divides the sorting results into intervals according to the numerical distribution relationship, establishes the level interval boundaries according to the interval range division rules, and identifies the interval ranges in sequence and constructs the corresponding relationship to obtain the deviation level interval set.

[0028] The grade matching output submodule, based on the set of deviation grade intervals, calls the periodic phase deviation value for interval matching processing, determines the position of the current periodic phase deviation value and the grade interval range, selects the corresponding grade identifier according to the matching result, and outputs the matching result in chronological order to generate rhythm deviation grade identifiers.

[0029] As a further aspect of the present invention, the rhythm control module includes:

[0030] The ratio setting processing submodule calls the airflow regulating valve component and the breathing training feedback display screen to set the inhalation phase time ratio and the exhalation phase time ratio according to the rhythm deviation level identifier. The set ratio is matched with the rhythm deviation level identifier to form a set of inhalation phase and exhalation phase time ratio parameters. The ratio parameters are sorted in chronological order to obtain the phase time ratio parameter sequence.

[0031] The time reconstruction processing submodule, based on the phase time ratio parameter sequence, calls the segmented time range in the respiratory lag segment set, performs proportional mapping processing on the segmented time range and performs time reconstruction, adjusts the segmented time according to the ratio of the inspiratory phase and the expiratory phase, and rearranges the reconstructed segmented time range in chronological order to obtain the rhythm time sequence.

[0032] The control sequence output submodule encodes the control state corresponding to the time point according to the rhythm time sequence, converts the encoding result into a control signal and outputs it to the airflow regulating valve assembly and the breathing training feedback display screen. At the same time, it organizes the control signals in chronological order to form a sequence structure and generates a breathing rhythm control sequence.

[0033] As a further aspect of the present invention, the process of combining and marking the flow rate and pressure states at each time point and recording continuous changes includes performing binary combination encoding on the positive or negative flow rate state and the pressure rising or falling state at the same time point, and recording the number of changes in the combination encoding of adjacent time points in chronological order. When the combination encoding of three or more consecutive time points is consistent, a stable interval marker is formed and written into the respiratory state time series.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the starting points of inhalation and exhalation are identified and cycles are divided by positive and negative flow rate conversion. At the same time, time normalization is performed to provide a unified comparative scale for respiratory behavior under different states. The phase difference is calculated by combining the reference rhythm phase and the intervals are divided according to the trend of change, which can refine the identification of local lag and rhythm drift in the respiratory process. The average phase difference of the intervals is statistically analyzed continuously and the deviation level is divided, so that the rhythm assessment takes into account both stability and trend. Based on the deviation results, the ratio of inhalation and exhalation time is reconstructed and a control signal is output, so that the training rhythm and the actual respiratory performance are dynamically matched, which helps to improve the training accuracy and individual adaptation level. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a system block diagram of the present invention; Figure 3 This is a flowchart of the airflow acquisition module in this invention; Figure 4 This is a flowchart of the period division module in this invention; Figure 5 This is a flowchart of the hysteresis segmentation module in this invention; Figure 6 This is a flowchart of the deviation grading module in this invention; Figure 7 This is a flowchart of the rhythm control module in this invention. Detailed Implementation

[0037] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0038] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0039] This invention provides a respiratory training system for post-rehabilitation in neurological nursing, such as... Figure 1-2 The diagram shows a respiratory training system for post-rehabilitation in neurology nursing. The system includes:

[0040] The airflow acquisition module acquires the flow velocity value output by the flow sensor inside the breathing training mask and the airway pressure value output by the pressure sensor inside the airway interface tubing. The flow velocity value and the airway pressure value are aligned in chronological order. The positive and negative changes in flow velocity and the rise and fall of pressure are recorded. The flow velocity state and pressure state corresponding to each time point are combined and marked to form a continuous breathing state recording sequence and generate a breathing state time series.

[0041] The cycle division module is based on the respiratory state time series. It calls the current rhythm frequency of the respiratory training rhythm prompting device, identifies the time point when the flow rate changes from negative to positive as the inspiratory start point, and identifies the time point when the flow rate changes from positive to negative as the expiratory start point. It takes the time interval between adjacent inspiratory start points as a single respiratory cycle, performs time normalization on the time points within the cycle, and marks the inspiratory and expiratory segments to obtain the respiratory cycle labeling sequence.

[0042] The lag segmentation module, based on the respiratory cycle annotation sequence, calls the reference time normalized phase value in the rehabilitation breathing training rhythm curve storage unit, calculates the phase difference value sequence for each time point, determines the trend of change of the phase difference value between adjacent time points, records the position where the change direction changes as the segment boundary, divides the single cycle into multiple continuous intervals according to the segment boundary, statistically processes the phase difference value within the interval to obtain the interval average phase difference value, and marks the interval time range and the interval average phase difference value accordingly to form a structured segmented data set and obtain the respiratory lag segment set;

[0043] The deviation grading module is based on the respiratory lag segment set and combined with the current operating time window of the respiratory resistance adjustment component. It summarizes the average phase difference value of the segment interval to obtain the periodic phase deviation value, sorts multiple consecutive periodic phase deviation values, constructs the deviation interval range and divides the level interval, and performs interval matching on the current periodic phase deviation value to obtain the rhythm deviation level identifier.

[0044] Based on the rhythm deviation level identifier, the rhythm control module calls the airflow regulating valve assembly and the breathing training feedback display to set the inspiratory phase time ratio and the expiratory phase time ratio. The ratio is applied to the time range of the segment in the respiratory lag segment set for time reconstruction processing. The reconstructed segment time sequence is arranged to form the current rhythm time sequence, and the time sequence is converted into a control signal and output to the airflow regulating valve assembly and the breathing training feedback display to generate a breathing rhythm control sequence.

[0045] The respiratory state time series includes a flow rate symbol sequence, a pressure change label sequence, and a timestamp index sequence; the respiratory cycle label sequence includes a cycle boundary marker set, an inspiratory segment label set, an expiratory segment label set, and a normalized time axis; the respiratory lag segment set includes a segment interval list, an interval average phase difference set, a segment boundary index set, and an interval time range set; the rhythm deviation level identifier includes a deviation level label, an interval mapping identifier, and a level threshold interval; the respiratory rhythm control sequence includes a control time series, a valve drive signal set, and a display feedback parameter set.

[0046] Specifically, such as Figure 2 , 3 As shown, the airflow acquisition module includes:

[0047] The flow rate and pressure acquisition submodule acquires the flow rate value output by the flow sensor inside the breathing training mask and the airway pressure value output by the pressure sensor inside the airway interface tubing. The flow rate and airway pressure values ​​are recorded synchronously according to the sampling time. The direction of the flow rate value at the sampling point is determined to form a positive or negative status indicator, and the trend of the change of the airway pressure value is determined to form an up or down status indicator, generating a flow and pressure synchronous sampling matrix.

[0048] The flow sensor and pressure sensor were connected to the same sampling clock with a sampling period of 10 milliseconds. 300 static points were collected consecutively to determine the zero bias of flow velocity and pressure. Subsequent real-time values ​​were then subtracted point by point. The corrected flow velocity was smoothed using a 5-point sliding motion, and the pressure was suppressed using a 3-point median. Flow velocities above 2 liters per minute were marked as positive inhalation, below 2 liters per minute as negative exhalation, and in between as pauses. Pressure was marked as rising if it cumulatively increased by 0.15 cmH2O within the last 30 milliseconds, and falling if it cumulatively decreased by 0.15 cmH2O; otherwise, it was marked as flat. For example, if a flow rate is continuously 18, 21, 24, 22, 19 L / min and the pressure is continuously 6.8, 7.0, 7.3, 7.5, 7.6 cmH2O, then this flow rate is recorded as positive inhalation with an upward trend. If another flow rate is continuously -14, -18, -22, -25 L / min and the pressure is continuously 7.4, 7.1, 6.8, 6.5 cmH2O, then this flow rate is recorded as negative exhalation with a downward trend. Five records are generated time-by-time: sampling time, corrected flow rate, flow rate status, corrected pressure, and pressure status, forming a flow-pressure synchronous sampling matrix.

[0049] The timing alignment processing submodule reconstructs the time sequence based on the sampling time identifier in the flow-pressure synchronous sampling matrix. It processes sampling points with time interval differences into a unified time step and forms a continuous sequence. It calls the flow velocity status identifier and pressure status identifier in the flow-pressure synchronous sampling matrix for position matching. It performs corresponding verification on the flow velocity status and pressure status at each time point and completes the combined marking to obtain the flow-pressure status alignment sequence.

[0050] After reading the flow-pressure synchronous sampling matrix, it is first rearranged in ascending order of sampling time, and then a continuous time axis is established with a uniform step size of 10 milliseconds. If the interval between adjacent raw points exceeds 10 milliseconds, a standard time is inserted into the gap. When inserting points, the flow velocity and pressure transition according to the time distance between the two points, and the state label is inherited according to the proximity priority rule. If the states before and after are consistent, the inserted point directly inherits the state; if the states before and after are inconsistent, the closer side determines the state. Subsequently, the position matching of the flow velocity state and pressure state is performed for each uniform time, and combined into a single label. Positive inhalation combined with an ascending state is written as inhalation propulsion, negative exhalation combined with a descending state is written as exhalation release, and a pause combined with a flat state is written as a transitional pause. Other combinations are encoded sequentially according to the same rule. If two points before and after a certain time are both inhalation propulsion, but the point is isolated and exhalation release occurs, the absolute value of the flow velocity and the change in pressure at that point are reviewed. If the pause and trend criteria are not exceeded, it is corrected to a transitional pause. This yields a time-continuous, position-one-to-position aligned sequence of flow pressure states, providing direct input for subsequent unified encoding.

[0051] The state sequence generation submodule sequentially encodes the combined markers of flow velocity and pressure states at time points in the flow-pressure state alignment sequence, converts each time point combined marker into a unified state identifier value, and arranges them in chronological order to form a sequence structure. It also performs continuity checks on the state identifiers in the sequence and fills in missing position identifiers to generate a respiratory state time series.

[0052] Sequential encoding is performed on the combined markers in the flow-pressure state alignment sequence, converting inspiratory propulsion, inspiratory transition, expiratory release, expiratory obstruction, and pause transition into fixed state values, which are then written into the state chain in chronological order. During writing, the duration of each state is recorded synchronously. The sampling step size is fixed at 10 milliseconds, so the duration directly corresponds to the actual duration. Continuity checks first identify time index gaps. If the gap length is no more than 3 points, it is filled with consistent results from preceding and following states. If the preceding and following states are inconsistent, the interpolated flow rate and interpolated pressure change at that position are reviewed. If the absolute value of the flow rate is less than 2 liters per minute and the pressure change is less than 0.15 cmH2O, it is filled with a pause state; otherwise, it is filled with a state on the side of the adjacent longer duration segment. If a theoretical cycle should have 429 points but only 421 points are actually present, the missing 8 points are preferentially filled into state segments with longer durations and gentler changes. If there are more points than expected, they are uniformly compressed from similar gentle segments. A respiratory state time series consisting of time index, state value, and duration point is formed, providing a single-value state chain for start and end point identification.

[0053] Specifically, such as Figure 2 , 4 As shown, the periodic division module includes:

[0054] The start and end point identification submodule is based on the respiratory state time series. It calls the current rhythm frequency of the respiratory training rhythm prompt device, detects the direction change of the flow rate state at continuous time points, determines the start of inhalation when the flow rate state changes from negative to positive, and determines the start of exhalation when the flow rate state changes from positive to negative. It records the time index corresponding to the start point in chronological order to form a set of start and end point identifiers and obtains the respiratory start and end point index sequence.

[0055] The flow direction changes are scanned point-by-point on the respiratory time series. The time point from negative exhalation to positive inhalation is marked as the candidate for the inspiratory start point, and the time point from positive inhalation to negative exhalation is marked as the candidate for the expiratory start point. After the candidate points are formed, the stability of the state is checked for 50 milliseconds before and after them. If at least two of the three consecutive points before the candidate point are expiratory-related states, and at least two of the three consecutive points after the candidate point are inspiratory-related states, then the point is confirmed as the inspiratory start point; the reverse rule is used to confirm the expiratory start point. To avoid duplicate identification caused by short-term jitter, a minimum boundary interval of 300 milliseconds is set. If the interval is insufficient, only candidate points with a higher proportion of target states within the last 20 milliseconds are retained. For example, if the inspiratory state continues to enter after 4210 milliseconds, then 4210 milliseconds is confirmed as the inspiratory start point; if the expiratory state continues to enter after 6330 milliseconds, then 6330 milliseconds is confirmed as the expiratory start point. After identification, the inspiratory start index and expiratory start index are output in chronological order. Then, the interval between adjacent inspiratory start points is compared with the theoretical cycle of 4290 milliseconds corresponding to the current 14 cycles per minute. If it still falls within the effective fluctuation window, it is retained as the basis for subsequent cycle construction.

[0056] The cycle interval construction submodule performs adjacent matching based on the inspiratory start time index in the respiratory start and end point index sequence, divides the time range between adjacent inspiratory start points into independent cycle intervals, calls the expiratory start time index for location within the cycle interval, confirms the boundaries of the cycle intervals and establishes the interval correspondence, and obtains a set of respiratory cycle intervals.

[0057] The time range between two adjacent inspiratory initiation points is used as an independent cycle parent interval. The expiratory initiation point is then located within this parent interval to confirm the cycle boundary. Taking 4210 milliseconds and 8500 milliseconds as examples, the first parent interval is formed between them. The expiratory initiation point of 6330 milliseconds is then found within this interval. Therefore, 4210 milliseconds to 6320 milliseconds is recorded as the inspiratory segment, and 6330 milliseconds to 8490 milliseconds as the expiratory segment. If two expiratory initiation point candidates appear within a parent interval, the distances of each candidate to the midpoint of the parent interval are compared, and the continuous length of the states on both sides is checked. The one with the closer distance and more stable continuous segment is retained. If no expiratory initiation point is detected within the parent interval, the first stable expiratory point after the pause and transition states is retrieved and recorded as the expiratory initiation point. Each cycle outputs seven data items: cycle number, start time, expiratory start time, end time, total cycle duration, inspiratory segment duration, and expiratory segment duration. If the total cycle duration is less than 3600 milliseconds or more than 5000 milliseconds, it is considered an abnormal cycle and will not proceed to the next processing stage.

[0058] The cycle labeling generation submodule performs time normalization on the time points within the cycle based on the time range of the cycle interval in the respiratory cycle interval set, and divides the time points into stages according to the positions of the inspiratory start point and expiratory start point in the respiratory start and end point index sequence. The time points are marked as inspiratory segments or expiratory segments and arranged in chronological order to generate a respiratory cycle labeling sequence.

[0059] Time normalization and phase labeling are performed on all time points within each respiratory cycle interval. Normalization maps the cycle start point to 0 and the cycle end point to 1. The normalized phase value for each intermediate time point is written as the proportion of its elapsed time from the cycle start point to the total cycle duration. For example, in cycle 1, which runs from 4210 ms to 8490 ms (total duration 4280 ms), 5280 ms, being at approximately one-quarter of the total duration, is written as 0.25; 7420 ms, being approximately three-quarters of the total duration, is written as 0.75. Phase labeling is directly based on the positions of the inspiratory and expiratory start points: 4210 ms to 6320 ms are all marked as the inspiratory phase, and 6330 ms to 8490 ms are all marked as the expiratory phase. If a time point falls exactly at the expiratory start point, it is directly rewritten as the expiratory phase from that point onwards, without retaining double labels. To enable point-by-point recall for subsequent phase difference construction, the submodule writes four fields for each unified sampling point: cycle number, absolute time, normalized phase value, and stage label, forming a respiratory cycle annotation sequence so that the same phase position in different cycles can be directly compared.

[0060] Specifically, such as Figure 2 , 5 As shown, the lag segmentation module includes:

[0061] The phase difference construction submodule calls the reference time normalized phase value in the rehabilitation breathing training rhythm curve storage unit according to the respiratory cycle annotation sequence, calculates the difference between the time point annotation phase and the reference phase, forms a phase difference record arranged in time order, and stores the time point phase difference value with the corresponding time index to obtain the phase difference sequence.

[0062] Normalized phase values ​​are read point-by-point from the respiratory cycle annotation sequence, and the reference normalized phase value corresponding to the current training frequency is retrieved from the rhythm curve storage unit. The current example rhythm frequency is 14 times per minute. The reference sequence and the actual sequence use the same sampling step size and the same number of points, so they can be paired one-to-one by index. The phase difference value at each time point is obtained by subtracting the reference normalized phase from the actual normalized phase, and then written into the phase difference value sequence along with the absolute time and cycle number. For example, if the actual phase at a certain point is 0.25 and the reference phase is 0.27, then the phase difference at that point is recorded as -0.02; if the actual phase at another point is 0.76 and the reference phase is 0.74, then the phase difference at that point is recorded as +0.02. To reduce the impact of single-point jumps, the submodule performs a 7-point moving average smoothing on the phase difference value. After smoothing, the magnitude of the phase difference before and after is compared. When the current value is greater than the previous point, it is recorded as upward; when it is less than the previous point, it is recorded as downward; and when the change is very small, it is recorded as parallel. The resulting phase difference sequence retains the time position, numerical value, and direction of change, providing a direct basis for subsequent segment boundary identification.

[0063] The boundary recognition processing submodule determines the trend based on the relationship between the phase difference changes of adjacent time points in the phase difference sequence, compares the consistency of the phase difference direction of continuous time points, marks the position where the change direction changes as the segment boundary, and records the time index corresponding to the boundary and arranges them in chronological order to obtain the segment boundary index sequence.

[0064] The direction of change in the phase difference sequence is compared point by point, and the position where the direction changes stably is determined as the segment boundary. The criterion for a stable transition is that at least 4 out of 5 consecutive points before a candidate point have the same direction, and at least 4 out of 5 consecutive points after the candidate point also have the same direction, but the directions before and after are opposite. If the preceding segment is downward and the following segment is upward, the boundary is recorded as the local minimum point boundary; if the preceding segment is upward and the following segment is downward, it is recorded as the local maximum point boundary. For example, if the phase difference sequence shows a continuous downward movement from 4210 ms to 5120 ms, and a continuous upward movement after 5130 ms, then 5130 ms is written as the first boundary; if the phase difference sequence shows a continuous upward movement from 6240 ms to 7060 ms, and a downward movement after 7070 ms, then 7070 ms is written as the second boundary. The start and end points of the cycle are automatically incorporated into the boundary sequence. If only a single point of reversal occurs, or the distance between two candidate boundaries is less than 200 milliseconds, retain the one with the larger absolute value of the phase difference and discard the short false boundary. Generate a segmented boundary index sequence arranged by period number and time order.

[0065] The segmented result generation submodule divides the single-cycle time range into intervals based on the segment boundary index sequence, defines the time intervals between adjacent boundaries as independent intervals, calls the phase difference value sequence for statistical processing within the interval to form the interval average phase difference value, marks the interval time range and the interval average phase difference value accordingly and organizes them in chronological order to generate a respiratory lag segment set.

[0066] Based on the segment boundary index sequence, the single-cycle time range is divided into adjacent pairs. Taking 4210 ms, 5130 ms, 7070 ms, and 8490 ms as examples, three intervals are formed between adjacent boundaries: 4210 ms to 5130 ms, 5140 ms to 7070 ms, and 7080 ms to 8490 ms, respectively. All smoothed phase differences within each sub-interval are accumulated and then divided by the number of points in the interval to obtain the average phase difference for that interval. Simultaneously, the duration of each interval, the dominant direction of change, and the stage classification are statistically analyzed. For example, the first interval has 93 points, and the accumulated smoothed phase difference is -1.209, which, when divided by 93, yields an average phase difference of -0.013; the second interval has an accumulated difference of -0.388, yielding an average phase difference of -0.002; and the third interval has an accumulated difference of +1.022, yielding an average phase difference of +0.007. The durations are 920 ms, 1930 ms, and 1410 ms, respectively. If the duration of a certain interval is less than 300 milliseconds, it is merged into the adjacent longer interval and is not retained separately. After processing, each interval is recorded with its cycle number, start time, end time, average phase difference, duration, and stage affiliation, forming a set of respiratory lag segments.

[0067] Specifically, such as Figure 2 , 6 As shown, the deviation grading module includes:

[0068] The deviation summary calculation submodule is based on the respiratory lag segment set. It calls the current running time window of the respiratory resistance regulation component to summarize and calculate the average phase difference value of the segment interval. The phase difference value within the same cycle is integrated and processed in chronological order. Combined with the corresponding segment data within the time window, the cycle is aggregated to form a cycle-level phase deviation record and obtain the cycle phase deviation value.

[0069] The respiratory lag segments are periodically aggregated according to the current running time window. The current example uses the most recent 6 complete cycles as one running window, corresponding to approximately 25.7 seconds. For each cycle, the average phase difference and duration of all intervals are read. Instead of a simple average, the values ​​are aggregated by duration. Taking the first cycle as an example, the average phase differences of the three intervals are -0.013, -0.002, and +0.007, with durations of 920 milliseconds, 1930 milliseconds, and 1410 milliseconds, respectively. The interval values ​​are then weighted by duration and divided by the total cycle duration of 4280 milliseconds, yielding a cycle phase deviation of -0.0014 for the first cycle. The second and third cycles are obtained in the same way, yielding -0.0021 and +0.0018, respectively. The deviation values ​​of the most recent 6 cycles are then combined to obtain the distribution of positive and negative deviations within the window, the average absolute deviation, and the maximum fluctuation amplitude. If the absolute deviation of a single period exceeds 0.020, it will not be included in the current window statistics. This will create a period-level phase deviation record consisting of the period number, the period phase deviation value, the window average absolute deviation, and the window fluctuation amplitude.

[0070] The interval division processing submodule sorts the continuous periodic phase deviation values ​​according to the periodic phase deviation values, divides the sorting results into intervals according to the numerical distribution relationship, establishes the level interval boundaries according to the interval range division rules, and identifies the interval ranges in sequence and constructs the corresponding relationship to obtain the deviation level interval set.

[0071] The periodic phase deviation values ​​of continuous cycles are sorted, and level interval boundaries are established based on the numerical distribution. In implementation, three levels of absolute deviation intervals obtained from preliminary experiments are used first: absolute values ​​less than 0.003 are classified as Level 1, absolute values ​​not less than 0.003 and less than 0.008 as Level 2, and absolute values ​​not less than 0.008 as Level 3. Then, based on the positive and negative directions, they are further divided into negative and positive levels. The deviation values ​​of the current window's six cycles are sorted as -0.0048, -0.0021, -0.0015, +0.0018, +0.0034, and +0.0061. Therefore, -0.0048 is classified as negative Level 2, -0.0021 and -0.0015 as negative Level 1, +0.0018 as positive Level 1, and +0.0034 and +0.0061 as positive Level 2. The boundaries were not arbitrarily set, but rather derived from the statistical distribution of 40 trainees and 4800 valid cycles. After comparison with manual annotation, the consistency rate reached 91.8%. The submodule outputs a set of level intervals, specifying the lower limit, upper limit, deviation direction, and corresponding level number of each interval, providing a benchmark for subsequent single-cycle matching.

[0072] The grade matching output submodule is based on the set of deviation grade intervals. It calls the periodic phase deviation value to perform interval matching processing, determines the position of the current periodic phase deviation value and the grade interval range, selects the corresponding grade identifier according to the matching result, and outputs the matching result in time order to generate rhythm deviation grade identifiers.

[0073] The current cycle phase deviation value is read and compared sequentially with the level boundaries output by the interval division processing submodule to determine the rhythm deviation level of each cycle. In the current example, the deviation value of the first cycle is -0.0015, and the absolute value is less than 0.003, so it outputs a negative level 1; the second cycle is -0.0021, still outputting a negative level 1; the third cycle is positive 0.0018, outputting a positive level 1; the fourth cycle is positive 0.0034, entering a positive level 2; the fifth cycle is positive 0.0061, continuing to maintain a positive level 2; the sixth cycle is negative 0.0048, transitioning to a negative level 2. To avoid frequent level changes due to occasional fluctuations, the submodule adds a continuous level count field and sets confirmation rules. If a level higher than 1 appears only once and falls back to level 1 before and after, then that cycle can be rewritten to a lower level; if two consecutive cycles maintain the same high level, it is directly retained. Since the 4th and 5th cycles are consecutively positive level 2, both officially output positive level 2. This forms a rhythm deviation level identifier sequence arranged in cycle order, which can be directly called by the proportional setting processing submodule.

[0074] Specifically, such as Figure 2 , 7 As shown, the rhythm control module includes:

[0075] The ratio setting processing submodule calls the airflow regulating valve component and the breathing training feedback display screen to set the inhalation phase time ratio and the exhalation phase time ratio according to the rhythm deviation level identifier. It matches the set ratio with the rhythm deviation level identifier and forms a set of inhalation phase and exhalation phase time ratio parameters. The ratio parameters are sorted in chronological order to obtain the phase time ratio parameter sequence.

[0076] Based on the rhythm deviation level identifier, the inhalation and exhalation time ratios are retrieved from a preset ratio mapping table. The mapping table specifies that negative level 1 corresponds to 49% inhalation and 51% exhalation, negative level 2 corresponds to 47% inhalation and 53% exhalation, positive level 1 corresponds to 51% inhalation and 49% exhalation, and positive level 2 corresponds to 53% inhalation and 47% exhalation. The remaining levels are configured according to the same rule. In the current example, cycles 1 and 2 are both negative level 1, so the ratio is 49:51; cycle 3 is positive level 1, so it's 51:49; cycles 4 and 5 are positive level 2, so it's 53:47; cycle 6 is negative level 2, but since the previous cycle was positive level 2, a direct switch would result in a 6 percentage point jump, exceeding the 4 percentage point limit. Therefore, this round transitions to 49:51 first. For each cycle, write down the cycle number, inhalation ratio, exhalation ratio, and source level to form a sequence of phase time ratio parameters, providing a clear duration allocation target for subsequent time reconstruction.

[0077] The time reconstruction processing submodule, based on the phase time ratio parameter sequence, calls the segment time range in the respiratory lag segment set, performs ratio mapping processing on the segment time range and performs time reconstruction, adjusts the segment time according to the ratio of the inspiratory phase and the expiratory phase, and rearranges the reconstructed segment time range in chronological order to obtain the rhythm time sequence.

[0078] Based on the phase time ratio parameter sequence, the time ranges of each segment in the respiratory lag segment set are redistributed. The total cycle length remains unchanged at 4270 milliseconds corresponding to the current rhythm; only the proportions of the inspiratory and expiratory segments in the total cycle are changed. Taking cycle 4, positive level 2 as an example, the inspiratory proportion is 53% and the expiratory proportion is 47%, corresponding to a new target inspiratory duration of 2274 milliseconds and a new target expiratory duration of 2016 milliseconds. If the original inspiratory segment was 2110 milliseconds, it needs to be lengthened by 164 milliseconds; if the original expiratory segment was 2160 milliseconds, it needs to be compressed by 144 milliseconds. The submodule does not distribute the time evenly across all points; instead, it first checks the original duration of each inspiratory and expiratory segment within the cycle and then allocates the increase or decrease based on the length ratio. For example, two inspiratory segments are increased by 72 milliseconds and 92 milliseconds respectively, and two expiratory segments are shortened by 70 milliseconds and 74 milliseconds respectively. After completion, the new starting point and new ending point of each segment are rewritten in sequence to ensure that the ending point of the previous segment is continuously connected with the starting point of the next segment, without time overlap or gaps, thus forming a new rhythmic time sequence.

[0079] The control sequence output submodule encodes the control state corresponding to the time point according to the rhythm time sequence, converts the encoding result into a control signal and outputs it to the airflow regulating valve assembly and the breathing training feedback display screen. At the same time, it organizes the control signals in time sequence to form a sequence structure and generates a breathing rhythm control sequence.

[0080] Based on the rhythmic time sequence, the control state corresponding to each time point is encoded and output to the airflow regulating valve assembly and the breathing training feedback display screen respectively. The encoding uses fixed Chinese character states: inhalation is written as state 1, exhalation as state 2, the inhalation-exhalation switching buffer as state 3, and the hold phase as state 4. Taking the 4th cycle as an example, after reconstruction, the inhalation duration is 2274 milliseconds and the exhalation duration is 2016 milliseconds. Therefore, all outputs within the range from the start point to 2274 milliseconds are state 1, followed by state 3 inserted around 20 milliseconds, and then back to state 2. The valve opening target value is synchronously bound to the level. At positive level 2, the target opening for the inhalation phase is set to 68% and the target opening for the exhalation phase is set to 52%; at negative level 1, they are set to 60% and 56% respectively. The display screen shows an inhalation prompt during the inhalation phase, an exhalation prompt during the exhalation phase, and a commutation prompt during the buffer phase. At the end of each cycle, the submodule compares the total duration of the control sequence with the total duration of the rhythmic time sequence; execution is only sent after they match. The output respiratory rhythm control sequence is arranged in the order of cycle number, time point, control state, valve opening degree, and display content, and corresponds one-to-one with the aforementioned sampling, segmentation, and grading results.

[0081] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A respiratory training system for post-rehabilitation in neurological nursing, characterized in that, The system includes: The airflow acquisition module collects the flow velocity value output by the flow sensor and the airway pressure value output by the pressure sensor. It aligns the flow velocity value and the airway pressure value in chronological order, determines the positive and negative changes in flow velocity and the rise and fall of pressure, combines and marks the flow velocity state and pressure state at each time point, records the continuous changes, and generates a respiratory state time series. Based on the respiratory state time series, the cycle division module calls the rhythm frequency of the rhythm prompting device, identifies the time point when the flow rate changes from negative to positive as the start of inhalation and the time point when the flow rate changes from positive to negative as the start of exhalation, divides the time interval between adjacent inhalation start points, performs time normalization processing, and marks the inhalation and exhalation segments to generate a respiratory cycle labeled sequence. The lag segmentation module calls the normalized phase value of the reference time in the rhythm curve storage unit according to the respiratory cycle annotation sequence, calculates the phase difference value at time points to obtain the phase difference value sequence, determines the trend of phase difference value change between adjacent time points and records the direction change position as the segment boundary, divides the interval and calculates the average phase difference value of the interval and marks the time range of the interval and the average phase difference value of the interval, and generates a respiratory lag segment set. The deviation grading module, based on the set of respiratory lag segments and combined with the operating time window of the resistance adjustment component, summarizes the average phase difference value of the interval to obtain the periodic phase deviation value, sorts the continuous periodic phase deviation values ​​and divides the grade interval, and matches the current periodic phase deviation value to obtain the rhythm deviation grade identifier. The rhythm control module, based on the rhythm deviation level identifier, calls the airflow regulating valve assembly and the feedback display to set the inspiratory phase time ratio and the expiratory phase time ratio, reconstructs the segmented time range according to the ratio, arranges them to form a rhythm time sequence, and converts the control signal to generate a respiratory rhythm control sequence.

2. The respiratory training system for post-rehabilitation in neurology nursing according to claim 1, characterized in that: The respiratory state time series includes a flow rate symbol sequence, a pressure change label sequence, and a timestamp index sequence; the respiratory cycle label sequence includes a cycle boundary marker set, an inspiratory segment label set, an expiratory segment label set, and a normalized time axis; the respiratory lag segment set includes a segment interval list, an interval average phase difference set, a segment boundary index set, and an interval time range set; the rhythm deviation level identifier includes a deviation level label, an interval mapping identifier, and a level threshold interval; the respiratory rhythm control sequence includes a control time series, a valve drive signal set, and a display feedback parameter set.

3. The respiratory training system for post-rehabilitation in neurology nursing according to claim 1, characterized in that: The process of determining the positive and negative changes in flow velocity includes comparing the flow velocity value output by the flow sensor with a preset zero reference threshold. When the flow velocity value output by the flow sensor is greater than the preset zero reference threshold, it is marked as a positive flow velocity state. When the flow velocity value output by the flow sensor is less than or equal to the preset zero reference threshold, it is marked as a negative flow velocity state.

4. The respiratory training system for post-rehabilitation in neurology nursing according to claim 1, characterized in that: The process of determining the pressure rise and fall state includes calculating the difference between the airway pressure values ​​output by the pressure sensor at two consecutive time points. When the airway pressure value output by the pressure sensor at the later time point is greater than the airway pressure value output by the pressure sensor at the earlier time point, it is marked as a pressure rise state. When the airway pressure value output by the pressure sensor at the later time point is less than or equal to the airway pressure value output by the pressure sensor at the earlier time point, it is marked as a pressure fall state.

5. The respiratory training system for post-rehabilitation in neurological nursing according to claim 1, characterized in that, The airflow acquisition module includes: The flow rate and pressure acquisition submodule acquires the flow rate value output by the flow sensor inside the breathing training mask and the airway pressure value output by the pressure sensor inside the airway interface tubing. The flow rate and airway pressure values ​​are recorded synchronously according to the sampling time. The direction of the flow rate value at the sampling point is determined to form a positive or negative status indicator, and the trend of the change of the airway pressure value is determined to form an up or down status indicator, generating a flow and pressure synchronous sampling matrix. The timing alignment processing submodule reconstructs the time sequence based on the sampling time identifier in the flow-pressure synchronization sampling matrix, processes sampling points with time interval differences into a unified time step and forms a continuous sequence, calls the flow velocity status identifier and pressure status identifier in the flow-pressure synchronization sampling matrix for position matching, performs corresponding verification of the flow velocity status and pressure status at each time point and completes the combined marking to obtain the flow-pressure status alignment sequence. The state sequence generation submodule sequentially encodes the combined markers of flow velocity and pressure states at time points in the flow-pressure state alignment sequence, converts each time point combined marker into a unified state identifier value, and arranges them in chronological order to form a sequence structure. It then performs a continuity check on the state identifiers in the sequence and fills in any missing identifiers to generate a respiratory state time sequence.

6. The respiratory training system for post-rehabilitation in neurological nursing according to claim 1, characterized in that, The periodic division module includes: The start and end point identification submodule, based on the breathing state time series, calls the current rhythm frequency of the breathing training rhythm prompting device, detects the direction change of the flow rate state at continuous time points, determines the time point when the flow rate state changes from negative to positive as the start of inhalation, and determines the time point when the flow rate state changes from positive to negative as the start of exhalation, and records the time index corresponding to the start point in chronological order to form a set of start and end point identifiers, thus obtaining the breathing start and end point index sequence; The cycle interval construction submodule performs adjacent matching based on the inspiratory start time index in the respiratory start and end point index sequence, divides the time range between adjacent inspiratory start points into independent cycle intervals, calls the expiratory start time index for location within the cycle interval, confirms the boundaries of the cycle intervals and establishes interval correspondence, and obtains a set of respiratory cycle intervals. The cycle labeling generation submodule performs time normalization processing on the time points within the cycle based on the time range of the cycle interval in the respiratory cycle interval set, and divides the time points into stages according to the positions of the inspiratory start point and expiratory start point in the respiratory start and end point index sequence. The time points are marked as inspiratory segments or expiratory segments and arranged in chronological order to generate a respiratory cycle labeling sequence.

7. The respiratory training system for post-rehabilitation in neurological nursing according to claim 1, characterized in that, The lag segmentation module includes: The phase difference construction submodule calls the reference time normalized phase value in the rehabilitation breathing training rhythm curve storage unit according to the respiratory cycle annotation sequence, calculates the difference between the time point annotation phase and the reference phase, forms a phase difference record arranged in time order, and stores the time point phase difference value with the corresponding time index to obtain the phase difference sequence. The boundary recognition processing submodule determines the trend based on the relationship between the phase difference changes of adjacent time points in the phase difference sequence, compares the consistency of the phase difference direction of continuous time points, marks the position where the change direction changes as the segment boundary, and records the time index corresponding to the boundary and arranges them in chronological order to obtain the segment boundary index sequence. The segmentation result generation submodule divides the single-cycle time range into intervals based on the segmentation boundary index sequence, defines the time intervals between adjacent boundaries as independent intervals, and calls the phase difference value sequence within the interval for statistical processing to form the interval average phase difference value. The interval time range and the interval average phase difference value are marked accordingly and arranged in chronological order to generate a respiratory lag segment set.

8. The respiratory training system for post-rehabilitation in neurology nursing according to claim 1, characterized in that, The deviation grading module includes: The deviation summary calculation submodule, based on the set of respiratory lag segments, calls the current operating time window of the respiratory resistance adjustment component to summarize and calculate the average phase difference value of the segment interval, integrates the phase difference value within the same cycle in chronological order, and combines it with the corresponding segment data within the time window to perform periodic aggregation, forming a periodic hierarchical phase deviation record, and obtaining the periodic phase deviation value. The interval division processing submodule sorts the continuous periodic phase deviation values ​​according to the periodic phase deviation values, divides the sorting results into intervals according to the numerical distribution relationship, establishes the level interval boundaries according to the interval range division rules, and identifies the interval ranges in sequence and constructs the corresponding relationship to obtain the deviation level interval set. The grade matching output submodule, based on the set of deviation grade intervals, calls the periodic phase deviation value for interval matching processing, determines the position of the current periodic phase deviation value and the grade interval range, selects the corresponding grade identifier according to the matching result, and outputs the matching result in chronological order to generate rhythm deviation grade identifiers.

9. The respiratory training system for post-rehabilitation in neurological nursing according to claim 1, characterized in that, The rhythm control module includes: The ratio setting processing submodule calls the airflow regulating valve component and the breathing training feedback display screen to set the inhalation phase time ratio and the exhalation phase time ratio according to the rhythm deviation level identifier. The set ratio is matched with the rhythm deviation level identifier to form a set of inhalation phase and exhalation phase time ratio parameters. The ratio parameters are sorted in chronological order to obtain the phase time ratio parameter sequence. The time reconstruction processing submodule, based on the phase time ratio parameter sequence, calls the segmented time range in the respiratory lag segment set, performs proportional mapping processing on the segmented time range and performs time reconstruction, adjusts the segmented time according to the ratio of the inspiratory phase and the expiratory phase, and rearranges the reconstructed segmented time range in chronological order to obtain the rhythm time sequence. The control sequence output submodule encodes the control state corresponding to the time point according to the rhythm time sequence, converts the encoding result into a control signal and outputs it to the airflow regulating valve assembly and the breathing training feedback display screen. At the same time, it organizes the control signals in chronological order to form a sequence structure and generates a breathing rhythm control sequence.

10. The respiratory training system for post-rehabilitation in neurology nursing according to claim 1, characterized in that: The process of combining and marking the flow rate and pressure status at each time point and recording continuous changes includes performing binary combination encoding on the positive or negative flow rate status and the pressure increase or decrease status at the same time point, and recording the number of changes in the combination encoding of adjacent time points in chronological order. When the combination encoding of three or more consecutive time points is consistent, a stable interval is formed and written into the respiratory state time series.

Images