Anaesthesia machine control system, control method and anaesthesia machine

Abstract

The invention relates to the technical field of anesthesia devices, in particular to an anesthesia machine control system and method and an anesthesia machine, and the system comprises an acquisition conversion module, a ladder segmentation module, a stage correction module, an integral sorting module and a valve control generation module. According to the method, the continuous pressure sequence is constructed by collecting the compression electric signals of the breathing bag and is smoothed, synchronous difference and gradient recognition is carried out on pressure changes and gas flow changes, refined stage division and dynamic position correction are formed, the boundary of the ventilation stage is clear and stable, and the accuracy is high. The valve opening reference relation is established through stage pressure integration and section sorting, the collaboration between pressure fluctuation and flow fluctuation is strengthened, deviation caused by manual experience dependence is reduced, valve dynamic adjustment is more real-time and consistent, and the gas conveying stability degree and the pressure control reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of anesthesia device technology, and in particular to an anesthesia machine control system, control method and anesthesia machine. Background Technology

[0002] The field of anesthesia device technology encompasses device systems for delivering anesthetic gases to a patient's respiratory system on demand and for monitoring and regulating the infusion process. Its core aspects include the generation of anesthetic gas ratios, the regulation of pressure and flow rate, the establishment and switching of breathing circuits, the monitoring of the ventilation process, and the airway connection with the patient. This technical field typically involves components such as anesthetic gas supply mechanisms, breathing circuit assemblies, pressure and flow rate regulation mechanisms, monitoring and sensing components, and corresponding control and execution structures, forming a comprehensive technical system for managing the entire anesthesia supply process.

[0003] Traditional anesthesia machine control systems and anesthesia machines refer to the control structure and process built around the supply and ventilation management of anesthetic gases. Typically, this involves installing mechanical regulators within the anesthesia machine to adjust the flow of oxygen and volatile anesthetic agents, maintaining stable gas pressure through mechanical pressure regulators, forming inhalation and exhalation circuits through fixed tubing, switching between different ventilation modes by using a dial-type flow regulator and a ventilation switching valve, and using mechanical pressure gauges and observation windows to read airway pressure and gas output, thereby achieving routine control of the overall operating status of the anesthesia machine.

[0004] Traditional control methods rely on mechanical adjustment and dial reading for parameter setting. Gas pressure and flow regulation are discrete and lack continuous feedback capability. Changes in breathing bag compression are difficult to quantify accurately. Ventilation stage division depends on human experience. There is a lack of effective correlation analysis between pressure fluctuations and flow changes, resulting in valve response lag, insufficient gas ratio stability, difficulty in timely identification of pressure anomalies, and ventilation consistency is significantly affected by operator differences, thereby increasing ventilation deviation and safety risks. Summary of the Invention

[0005] To address the technical problems existing in the prior art, embodiments of the present invention provide an anesthesia machine control system, control method, and anesthesia machine. The technical solution is as follows: On the one hand, an anesthesia machine control system is provided, the system comprising: The acquisition and conversion module acquires the electrical signal of the breathing bag compression degree and converts it into a pressure sequence from analog to digital. It then smooths the pressure sequence and acquires the corresponding gas flow rate for synchronous differential analysis, generating pressure-flow differential data and transmitting it to the gradient segmentation module. The gradient segmentation module performs gradient change detection on the pressure-flow differential data, identifies the location of gradient abrupt changes, performs segmentation on the smoothed pressure sequence, generates compression stage data, and transmits it to the stage correction module. The stage correction module acquires the corresponding gas flow rate based on the compression stage data and analyzes the flow rate fluctuation. It then combines the compression stage data to determine and correct the stage position, generates the stage correction result, and transmits it to the integration sorting module. The integral sorting module calculates the stage pressure integral based on the stage correction result and the smoothed pressure sequence, compares it with the stage correction result and sorts it by segment, generates the opening basic data and transmits it to the valve control generation module. The valve control generation module obtains the corresponding gas flow rate based on the opening basic data and calculates the opening adjustment factor by normalization. It then linearly combines the opening adjustment factor with the opening basic data to generate valve control commands for the anesthesia machine.

[0006] As a further embodiment of the present invention, the pressure-flow differential data includes a pressure differential sequence, a flow differential sequence, and a differential corresponding time index; the compression stage data includes a compression stage start point set, a compression stage end point set, and compression stage pressure characteristic parameters; the stage correction results include a stage position correction amount, a stage boundary correction index, and a stage validity judgment mark; the opening degree basic data includes a stage pressure integral, a valve opening degree initial ratio set, and stage sorting identifier information; and the anesthesia machine valve control command includes a target valve opening degree sequence, an opening degree adjustment control coefficient set, and a valve drive execution sequence.

[0007] As a further aspect of the present invention, the acquisition and conversion module specifically comprises: The signal acquisition submodule acquires the electrical signal of the breathing bag compression degree and performs discrete quantization on the electrical signal based on the voltage amplitude through analog-to-digital conversion. The quantized multi-point voltage sequence is arranged in time order and mapped into a pressure sequence to generate a pressure sequence value group. The sequence smoothing submodule calls the pressure sequence value group and performs smoothing processing based on the amplitude difference of adjacent sampling points. It weights the amplitudes of adjacent points in the sequence according to the weighted average coefficient and replaces the original amplitudes. It combines all the processed points in the original order to generate a smoothed pressure sequence value group. The differential generation submodule performs amplitude differential analysis on adjacent points based on the smoothed pressure sequence value group and the sequential relationship of the sequence points. It obtains the corresponding gas flow rate from the amplitude change obtained by differential analysis and performs synchronous pairing. It then organizes the differential components to generate pressure-flow differential data.

[0008] As a further aspect of the present invention, the stepped segmentation module specifically comprises: The gradient detection submodule, based on the pressure-flow differential data, performs gradient change detection according to adjacent differential components, compares the differential components with the gradient median value and calculates the numerical difference between the differential components and the gradient median value, and combines all the difference quantities in the original sequence order to generate a gradient change sequence value group. The mutation identification submodule calls the gradient change sequence value group, performs amplitude comparison based on adjacent differences in the sequence, marks and records the sequence positions of points where the difference exceeds the mutation benchmark value, and arranges all recorded positions in sequence order to generate a gradient mutation position sequence. The segmentation submodule performs segmentation based on the gradient mutation location sequence and the smoothed pressure sequence. It sequentially combines and divides the pressure points between adjacent mutation locations into multiple pressure segments, and then organizes the pressure segment sequence according to the segmentation order to generate compression stage data.

[0009] As a further aspect of the present invention, the mutation benchmark value is determined by sorting all the differential quantities from low to high according to their values ​​after statistically analyzing the overall distribution characteristics of multiple differential quantities in the gradient change sequence, and then obtaining the median differential quantity and the dispersion of adjacent differential quantities.

[0010] As a further aspect of the present invention, the stage correction module specifically comprises: The flow extraction submodule extracts the corresponding gas flow sequence based on the compression stage data, arranges the flow of multiple time sampling points in the sequence according to time, calculates the change in flow difference between adjacent sampling points, analyzes the range of change in multiple intervals in the sequence, and generates the flow change range. The fluctuation determination submodule calls the flow change range and the compression stage data time position to determine the flow change range segment, compares the flow change range within the segment with the change range of adjacent segments to determine the fluctuation position, and corresponds it with the compression stage data to generate a stage fluctuation indicator. The position correction submodule calls the stage fluctuation indicator, obtains the stage start and end position sequence in the compressed stage data, performs position offset calculation on the stage fluctuation indicator, and adds the offset calculation value to the stage start and end position sequence for correction, generating the stage correction result.

[0011] As a further aspect of the present invention, the integral sorting module specifically comprises: The pressure integration submodule, based on the stage correction results and the smoothed pressure sequence, obtains and accumulates the pressure sequence of continuous time points within multiple stages, multiplies the accumulated time interval with the pressure value and merges them according to the stage as the integration amount corresponding to the stage, judges the segment alignment at the stage boundary position, and generates the stage pressure integration amount. The pressure relationship submodule calls the stage pressure integral quantity and the stage order in the stage correction result to perform a segment comparison on the stage pressure integral quantity, uses the difference of the integral quantity in the segment comparison to perform an inter-segment difference calculation with the adjacent stage, and arranges them according to the stage sequence to generate stage pressure relationship data. The segment sorting submodule performs sorting calculations on the differences in multiple segments within the stage pressure relationship data, maps the segment arrangement order obtained from the sorting calculation to the stage number in the stage correction result, and converts it into a quantified opening sequence to generate basic opening data.

[0012] As a further aspect of the present invention, the valve control generation module specifically comprises: The flow normalization submodule obtains the corresponding gas flow sequence based on the opening basic data and performs a difference operation on all gas flow values ​​in the sequence and the minimum gas flow value in the sequence. Then, it performs a ratio operation on the difference between the difference and the difference between the maximum and minimum gas flow values ​​in the sequence to generate a gas flow normalization sequence. The adjustment factor submodule calls the gas flow normalization sequence and the opening degree basic data to perform difference calculation, performs ratio analysis on the difference and the opening degree basic data, and arranges them according to the position of the gas flow normalization sequence to generate the opening degree adjustment factor sequence. The instruction combination submodule calls the opening adjustment factor sequence, performs linear combination calculations on multiple factors in the sequence and the opening basic data, reassembles the quantification results obtained by linear combination in sequence order and transcribes them into control parameters, and generates anesthesia machine valve control instructions.

[0013] On the other hand, an anesthesia machine control method, which is based on the aforementioned anesthesia machine control system, includes the following steps: S1: Acquire electrical signals of the compression degree of the breathing bag and convert them from analog to digital to a pressure sequence. Smooth the pressure sequence and collect the corresponding gas flow rate to calculate the difference, generating pressure-flow rate difference data. S2: Perform gradient change detection on the pressure-flow differential data, identify the location of gradient abrupt changes, and perform segmentation on the smoothed pressure sequence to generate compression stage data; S3: Based on the compression stage data, obtain the corresponding gas flow rate and analyze the flow rate fluctuation. Combine the compression stage data to determine the stage position and correct the position, and generate the stage correction result. S4: Based on the stage correction results and the smoothed pressure sequence, calculate the stage pressure integral and compare it with the stage correction results and perform segment sorting to generate basic opening data; S5: Based on the opening baseline data, obtain the corresponding gas flow rate and normalize it to calculate the opening adjustment factor. Linearly combine the opening adjustment factor with the opening baseline data to generate anesthesia machine valve control instructions.

[0014] An anesthesia machine includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement an anesthesia machine control system as described above.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: A continuous pressure sequence is constructed based on the respirator compression electrical signal and smoothed. Pressure changes and gas flow rate changes are synchronously differentially identified and gradients are used to form a refined stage division and dynamic position correction, making the ventilation stage boundary clear and stable. By integrating the stage pressure and sorting the segments, a valve opening benchmark relationship is established, which strengthens the synergy between pressure fluctuations and flow rate fluctuations, reduces the deviation caused by human experience, and makes the dynamic adjustment of the valve more real-time and consistent, thereby improving the stability of gas delivery and the reliability of pressure control. Attached Figure Description

[0016] 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.

[0017] Figure 1 This is a system schematic diagram of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the acquisition and conversion module in this invention; Figure 4 This is a flowchart of the gradient segmentation module in this invention; Figure 5 This is a flowchart of the stage correction module in this invention; Figure 6 This is a flowchart of the integral sorting module in this invention; Figure 7 This is a flowchart of the valve control generation module in this invention; Figure 8 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] 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.

[0023] This invention provides an anesthesia machine control system, such as... Figure 1-2 The diagram shown illustrates an anesthesia machine control system, which includes: The acquisition and conversion module acquires the electrical signal of the breathing bag compression degree and converts it into a pressure sequence from analog to digital. It then smooths the pressure sequence and acquires the corresponding gas flow rate for synchronous differential analysis, generating pressure-flow differential data and transmitting it to the gradient segmentation module. The gradient segmentation module detects gradient changes in the differential pressure-flow data, identifies locations of gradient abrupt changes, divides the smoothed pressure sequence into segments, generates compression stage data, and transmits it to the stage correction module. The stage correction module acquires the corresponding gas flow rate based on the compression stage data and analyzes the flow rate fluctuation. It combines the compression stage data to determine and correct the stage position, generates the stage correction result, and transmits it to the integration and sorting module. The integral sorting module calculates the stage pressure integral based on the stage correction results and the smoothed pressure sequence, compares it with the stage correction results and sorts it by section, generates the basic opening data and transmits it to the valve control generation module. The valve control generation module obtains the corresponding gas flow rate based on the opening basic data and calculates the opening adjustment factor by normalization. It then linearly combines the opening adjustment factor with the opening basic data to generate valve control commands for the anesthesia machine.

[0024] The pressure-flow differential data includes pressure differential sequence, flow differential sequence, and differential time index; the compression stage data includes compression stage start point set, compression stage end point set, and compression stage pressure characteristic parameters; the stage correction results include stage position correction amount, stage boundary correction index, and stage validity judgment mark; the opening basic data includes stage pressure integral, valve opening initial proportion set, and stage sorting identifier information; and the anesthesia machine valve control commands include target valve opening sequence, opening adjustment control coefficient set, and valve drive execution timing.

[0025] Specifically, such as Figure 2 , 3 As shown, the acquisition and conversion module specifically consists of: The signal acquisition submodule acquires the electrical signal of the breathing bag compression degree and performs discrete quantization on the electrical signal based on the voltage amplitude through analog-to-digital conversion. The quantized multi-point voltage sequence is arranged in time order and mapped into a pressure sequence to generate a pressure sequence value group. For real-time monitoring of the respirator's compression status, a piezoelectric pressure sensor tightly attached to the outer wall of the respirator first performs continuous analog electrical signal capture. The sensor is set to a sampling frequency of 100 Hz, meaning that a voltage amplitude capture operation is performed every 10 milliseconds. The acquired raw analog voltage signal range is strictly limited by the hardware circuit to between 0 volts and 5 volts. This voltage range linearly corresponds to the deformation of the respirator's outer wall, and is then mapped to the pressure variation range of 0 kPa to 10 kPa inside the respirator through a calibrated conversion coefficient. The analog signal then enters the analog-to-digital conversion unit via a transmission line. This unit uses a 12-bit precision quantization standard to discretize the continuously varying analog voltage from 0 volts to 5 volts into digital quantized values ​​from 0 to 4095. The digital quantized value at each sampling moment is read, and the voltage-to-pressure conversion coefficient stored in the register is called to perform calculations. This conversion coefficient was obtained through previous standard pressure source experiments and set to 0.00244 kPa per quantization unit (i.e., 10 kPa divided by 4095). The quantized value collected at each moment is multiplied by the coefficient to obtain the corresponding instantaneous pressure value. The calculated pressure values ​​are then stored sequentially in the pressure sequence storage space, strictly following the order of the sampling timestamps. For example, a quantized value of 820 collected at 10 milliseconds yields a pressure value of 2.00 kPa; a quantized value of 845 collected at 20 milliseconds yields a pressure value of 2.06 kPa; and a quantized value of 810 collected at 30 milliseconds yields a pressure value of 1.98 kPa. This conversion and arrangement process continues until a complete breathing bag compression cycle is completed, forming a pressure sequence value group containing multiple discrete pressure points. To verify the specific distribution of the collected data, data from the first five sampling points in a specific monitoring process are shown in Table 1.

[0026] Table 1 Initial pressure data during respirator compression Sampling time sequence number Sampling time raw quantization value Converted voltage Calculate pressure value 1 10 820 1.00 2.00 2 20 845 1.03 2.06 3 30 810 0.99 1.98 4 40 860 1.05 2.10 5 50 835 1.02 2.04 As shown in Table 1, the states of five consecutive sampling points from 10 milliseconds to 50 milliseconds are recorded. The original quantized values ​​are used to calculate the specific computational pressure values ​​through conversion coefficients. This set of pressure values ​​(2.00, 2.06, 1.98, 2.10, 2.04) constitutes the initial pressure sequence value set, which serves as the basic data input for subsequent smoothing processing.

[0027] The sequence smoothing submodule calls the pressure sequence value group and performs smoothing processing based on the amplitude difference of adjacent sampling points. It weights the amplitudes of adjacent points in the sequence according to the weighted average coefficient and replaces the original amplitudes. It combines all the processed points in the original order to generate a smoothed pressure sequence value group. The generated pressure sequence value set is invoked, and a weighted smoothing operation based on the amplitude difference between adjacent points is performed on the minor fluctuations in the sequence caused by hand tremors or circuit noise. First, a sliding window is created in memory, sequentially covering the current point, the previous sampling point, and the next sampling point in the sequence. The values ​​of these three points are read, and the absolute value of the amplitude difference between the previous sampling point and the current point, and the absolute value of the amplitude difference between the current point and the next sampling point are calculated respectively. A preset amplitude difference judgment interval is established, which is determined through tensile tests based on the elastic hysteresis characteristics of the respirator material: when the absolute value of the amplitude difference is less than 0.1 kPa, it is judged as a stable fluctuation interval; when the absolute value of the amplitude difference is greater than or equal to 0.1 kPa, it is judged as a drastic change interval. Based on the judgment results, a weighted average coefficient is dynamically assigned. If both amplitude differences are within a stable fluctuation range, the previous, current, and next points are assigned similar weight coefficients: 0.33, 0.34, and 0.33, respectively. If either amplitude difference is within a rapidly changing range, to preserve the peak characteristics of the signal, the weight coefficient of the current point is increased to 0.6, while the weight coefficients of the previous and next points are decreased to 0.2. Taking the second sampling point (20ms, 2.06kPa) in Table 1 as an example, smoothing calculations are performed: the previous point is 2.00kPa, and the next point is 1.98kPa. The difference at the previous point is calculated as 2.06 minus the absolute value of 2.00, which is 0.06 kPa, and the difference at the next point is 2.06 minus the absolute value of 1.98, which is 0.08 kPa. Since both 0.06 kPa and 0.08 kPa are less than the threshold of 0.1 kPa, they are determined to be within a stable fluctuation range. The weighted summation of the recombined weights 0.33, 0.34, and 0.33 is calculated as follows: 2.00*0.33 + 2.06*0.34 + 1.98*0.33 = 0.66 + 0.7004 + 0.6534 = 2.0138, which is 2.0138 kPa. This calculated result of 2.0138 kPa replaces the original amplitude of the second point in the original sequence. Next, the calculation is performed on the third sampling point (30 ms, 1.98 kPa), with the previous point being the original value of 2.06 kPa and the subsequent point being 2.10 kPa. The difference before the calculation is 0.08 kPa, and the difference after is 0.12 kPa. Since the difference after 0.12 kPa is greater than 0.1 kPa, the weight allocation of the drastic change interval is triggered at 0.2, 0.6, and 0.2. The calculation is as follows: 2.06*0.2 + 1.98*0.6 + 2.10*0.2 = 0.412 + 1.188 + 0.42 = 2.02 kPa. This logic is then applied sequentially to all points in the sequence except for the beginning and end points. The calculated new pressure values ​​are then recombined according to their original time sequences, ultimately generating a smooth pressure sequence value set.

[0028] The differential generation submodule performs amplitude differential analysis on adjacent points based on the smoothed pressure sequence value group and the sequential relationship of the sequence points. It obtains the corresponding gas flow rate from the amplitude change obtained by differential analysis and performs synchronous pairing. It then organizes the differential components to generate pressure-flow differential data. The system receives a smoothed pressure sequence and performs differential calculations on adjacent points according to the chronological order of the time sequence to obtain the pressure change per unit time. It reads the amplitude of the i-th point and the amplitude of the i+1-th point in the smoothed sequence, and performs a subtraction operation (subtracting the previous point from the subsequent point) to obtain the pressure difference component. This difference component reflects the compression or rebound state of the respirator within a 10-millisecond sampling interval. Subsequently, a preset gas flow conversion model is invoked to pair this difference component. This conversion model is based on fluid dynamics experiments, setting a linear correlation coefficient between flow rate and pressure difference. The experimentally measured flow resistance coefficient of the respirator under standard atmospheric pressure is 50 liters per minute per kilopascal (i.e., a pressure change of 1 kilopascal corresponds to an instantaneous flow rate change of 50 liters per minute). The calculated pressure difference component is multiplied by this flow resistance coefficient to obtain the gas flow rate value corresponding to that time interval (Note: Since the pressure difference component ΔP implicitly contains a 10ms time interval, its physical meaning is equivalent to the instantaneous pressure change rate ΔP / Δt, so it can be directly matched with the flow coefficient). Taking the smoothed data as an example: the smoothed value for the second point is 2.0138 kPa, and the smoothed value for the third point is 2.02 kPa. The difference is calculated as: 2.02 - 2.0138 = 0.0062 kPa. This positive value indicates that the pressure is increasing and the breathing bag is in the compression and decompression phase. Substituting the difference of 0.0062 kPa into the flow rate calculation: 0.0062 * 50 = 0.31 L / min. This calculated result of 0.31 L / min is the paired gas flow rate for that time period. Taking the third point (2.02 kPa) and the set smoothed value of the fourth point (2.05 kPa) as another example, the difference is 2.05 - 2.02 = 0.03 kPa, and the corresponding flow rate is 0.03 * 50 = 1.5 L / min. The pressure difference calculated for each sampling interval and the calculated gas flow rate are integrated and packaged in a one-to-one correspondence, constructing data entries containing timestamps, pressure difference components, and gas flow rates in chronological order. By traversing the entire smooth pressure sequence and completing the calculation and pairing of all adjacent point pairs, complete pressure-flow differential data is finally generated. The calculation result of 0.31 liters per minute indicates the instantaneous intensity of the gas output by the breathing bag under the current small pressure rise rate, revealing the quantitative relationship between compression action and actual gas supply.

[0029] Specifically, such as Figure 2 , 4 As shown, the stepped segmentation module is specifically as follows: The gradient detection submodule is based on pressure and flow differential data and performs gradient change detection according to adjacent differential components. It compares the differential components with the gradient median value and calculates the numerical difference between the differential components and the gradient median value. All the difference values ​​are combined in the original sequence order to generate a gradient change sequence value group. First, the output pressure-flow differential data is retrieved, and the pressure differential sequence arranged in chronological order is extracted. Gradient calculation logic is set up to obtain the gradient of adjacent differential components by subtracting the pressure differential component at the current sampling time from the pressure differential component at the previous sampling time. This gradient characterizes the acceleration of pressure change, i.e., the rate of compression or rebound of the breathing bag. The calculation logic is to subtract the previous pressure differential component from the current pressure differential component. To eliminate non-motion-related random noise interference and extract effective trend features, a sliding window with a length of 5 sampling points is constructed, covering the current point and the two points before and after it. The 5 gradient values ​​within the window are stored in a temporary buffer and sorted. The third value after sorting is selected as the median gradient value for the window. Subsequently, a comparison calculation between the differential component and the median gradient value is performed. The gradient value calculated at the current time is subtracted from the corresponding median gradient value, and the absolute value is taken to obtain the numerical difference between the two. This calculation process can effectively separate abrupt changes in the signal because the median value represents the local trend, while the numerical difference reflects the degree to which the current point deviates from the local trend. Taking actual monitoring data as an example, a continuous pressure differential sequence (unit: kPa) is extracted, with the following data from the 20th to the 60th millisecond: 0.010 for the 20th millisecond, 0.012 for the 30th millisecond, 0.015 for the 40th millisecond, 0.040 for the 50th millisecond (simulating sudden acceleration of the pressure), and 0.042 for the 60th millisecond. First, the gradient of change is calculated: the gradient for the 30th millisecond is 0.012 - 0.010 = 0.002; the gradient for the 40th millisecond is 0.015 - 0.012 = 0.003; the gradient for the 50th millisecond is 0.040 - 0.015 = 0.025; and the gradient for the 60th millisecond is 0.042 - 0.040 = 0.002. If the gradient data is set to a stable 0.002 before the 30th millisecond, then at the 50th millisecond, the data within the sliding window could be 0.002, 0.003, 0.025, 0.002, 0.002 (and even at the last point, it would still be 0.002). Sort this set of data to obtain 0.002, 0.002, 0.002, 0.003, 0.025, with a median value of 0.002. The numerical difference at the 50th millisecond is calculated as 0.025 minus the absolute value of 0.002, resulting in 0.023. This result of 0.023 is significantly greater than the background noise level, indicating a drastic change in action state at the 50th millisecond. The advantage of this method is that by introducing the median value as a dynamic reference, it can adaptively filter out uniform gradient fluctuations caused by slight hand tremors, accurately pinpointing acceleration abrupt changes caused by altered force applied manually. Perform the above operation on the entire sequence, and combine all the calculated numerical differences in the original time order to generate a gradient change sequence value group.

[0030] The mutation identification submodule calls the gradient change sequence value group, performs amplitude comparison based on adjacent differences in the sequence, marks and records the sequence position of points where the difference exceeds the mutation benchmark value, and arranges all recorded positions in sequence order to generate a gradient mutation position sequence. The generated gradient change sequence value set is used to accurately pinpoint the critical time points where the bag compression state changes from continuous numerical variations. Each numerical difference point in the gradient change sequence is traversed and compared with a preset mutation benchmark value. The mutation benchmark value is set based on statistical analysis of a large amount of bag compression experimental data. The experiment selected 50 groups of compression operation data from different medical personnel in a calm state, calculating the mean and standard deviation of the gradient change sequence during the steady-state compression phase (excluding start and end points). Experimental data showed that the average background fluctuation under steady-state conditions was 0.0015 kPa per 10 milliseconds, and the standard deviation was 0.0008. Based on the 3-standard-deviation criterion and considering the need to capture weak signals, the mutation benchmark value was set as the mean plus 4 times the standard deviation, i.e., 0.0015 + 4 * 0.0008 = 0.0047, rounded down to 0.005 kPa per 10 milliseconds. The comparison logic is as follows: If the numerical difference at the current point is greater than 0.005, the point is determined to be a gradient abrupt change point, and the corresponding timestamp or sequence index position is immediately read and labeled. Taking the numerical difference of 0.023 at the 50th millisecond calculated in the above steps as an example, 0.023 is compared with the baseline value of 0.005. Since 0.023 is greater than 0.005, it is determined that there is an action abrupt change (such as pressing to start or accelerating) at the 50th millisecond. Conversely, if the calculation result at the 40th millisecond is 0.001 (the set value), which is less than 0.005, it is considered a non-abrupt change point. To prevent misjudgment of a single noise point, a continuity verification logic is further introduced: only when the numerical difference between two consecutive sampling points exceeds the baseline value, or the difference of a single point exceeds three times the baseline value (i.e., 0.015), is it confirmed as a valid abrupt change position. In the case of a numerical difference of 0.023 at the 50th millisecond, its value has exceeded three times the baseline value, so it is directly confirmed as a valid abrupt change. All verified mutation point locations were indexed sequentially according to time to generate a gradient mutation location sequence. Partial processed data and results are shown in Table 2.

[0031] Table 2. Gradient mutation detection process data table time Gradient Change gradient median Numerical differences Mutation determination 30 0.002 0.002 0.000 no 40 0.003 0.002 0.001 no 50 0.025 0.002 0.023 Yes (location record) 60 0.002 0.002 0.000 no As shown in Table 2, by comparing the calculated numerical difference with the mutation baseline value of 0.005, the critical node of the state change at 50 milliseconds was successfully identified. This result shows that the moment of action mode switching can be sensitively captured from a smooth pressing process.

[0032] The segmentation submodule performs segmentation based on the gradient mutation location sequence and the smoothed pressure sequence. It combines and divides the pressure points between adjacent mutation locations into multiple pressure segments in sequence, and then organizes the pressure segment sequence according to the segmentation order to generate compressed stage data. Based on the generated gradient abrupt change sequence and the smoothed pressure sequence, a fine segmentation of the respirator compression cycle is performed. The index values ​​in the gradient abrupt change sequence are read; these index values ​​represent the moments when the respirator's operational state physically changes, such as the start of compression from rest, the transition from compression to hold, or the end of release to rest. These abrupt change positions are used as segmentation points to divide the complete pressure sequence into several independent subsequences on the time axis. The abrupt change sequence is set to include two index points: 50 milliseconds and 150 milliseconds. The set of points from the start point to 50 milliseconds in the extracted smoothed pressure sequence is defined as the first segment (e.g., the resting baseline segment), the set of points between 50 milliseconds and 150 milliseconds is defined as the second segment (e.g., the effective compression segment), and the set of points after 150 milliseconds is defined as the third segment (e.g., the release and rebound segment). After segmentation, the data within each segment is processed for attributes. For the second segment (50 milliseconds to 150 milliseconds), the total pressure change and duration within this segment are calculated. Setting the pressure value at 50 milliseconds to 2.04 kPa and the pressure value at 150 milliseconds to 3.54 kPa, the pressure increment for this segment is 3.54 - 2.04 = 1.50 kPa, with a duration of 150 - 50 = 100 milliseconds. These statistical characteristics are packaged with all the original pressure point data within the segment and arranged according to the time sequence of the segmentation (segment 1, segment 2, segment 3) to generate structured compression phase data. This compression phase data not only includes segmented pressure waveforms but also includes start and end time markers for each phase, providing semantically segmented basic data units for subsequent analysis of tidal volume output and compression frequency. Through this gradient-mutation-based segmentation processing, continuous analog waveforms are transformed into discrete action event sequences with clinical operational significance.

[0033] Specifically, such as Figure 2 , 5 As shown, the phase correction module specifically includes: The flow extraction submodule extracts the corresponding gas flow sequence based on the compression stage data, arranges the flow of multiple time sampling points in the sequence according to time, calculates the change in flow difference between adjacent sampling points, analyzes the range of change in multiple intervals in the sequence, and generates the flow change range. The generated compression phase data is retrieved, and the gas flow rate sequence for that time period is extracted. Each flow rate point in the sequence is arranged according to the sampling timestamp order. A time step of 10 milliseconds is set, and the flow rate value of the current sampling point and the flow rate value of the next sampling point are read. A subtraction operation is performed to obtain the change in flow rate difference between adjacent sampling points. This difference reflects the acceleration or deceleration characteristics of the gas flow velocity per unit time. To analyze the stability of flow rate changes and generate amplitude ranges, the entire compression phase time series is divided into multiple equal-length analysis intervals, with each interval length set at 30 milliseconds (i.e., containing 3 sampling point intervals). Within each analysis interval, all changes in flow rate difference are iterated, and the maximum and minimum differences within the interval are selected. These two values ​​are then combined to form the flow rate change amplitude range for that interval. Taking the effective compression phase (50 milliseconds to 140 milliseconds) from a specific monitoring instance as an example, data from the first two analysis intervals are selected for illustration. Interval A (50 ms to 80 ms): Sampling point flow rate (liters per minute): 0.31 for 50 ms, 0.35 for 60 ms, 0.38 for 70 ms, and 0.42 for 80 ms. Calculate the difference between adjacent values: the difference at 50 ms is 0.35 - 0.31 = 0.04; the difference at 60 ms is 0.38 - 0.35 = 0.03; and the difference at 70 ms is 0.42 - 0.38 = 0.04. Within this interval, the maximum difference is 0.04, the minimum difference is 0.03, and the resulting flow rate variation range is 0.03 to 0.04. Interval B (80 ms to 110 ms): Sampling point flow rate (liters per minute): 0.42 for 80 ms, 1.52 for 90 ms, 2.72 for 100 ms, and 3.82 for 110 ms. Calculate the adjacent differences: the difference at 80 milliseconds is 1.52 - 0.42 = 1.10; the difference at 90 milliseconds is 2.72 - 1.52 = 1.20; and the difference at 100 milliseconds is 3.82 - 2.72 = 1.10. Within this interval, the maximum difference is 1.20, the minimum difference is 1.10, and the generated flow rate variation range is 1.10 to 1.20. Perform the above calculations on all time periods within the compression phase to generate a series of flow rate variation ranges arranged in chronological order. Some data are shown in Table 3.

[0034] Table 3 Extraction of Flow Rate Variation Range During Compression Phase Analysis of interval numbers Time range Flow difference set Extracting amplitude range 1 50-80 {0.04，0.03，0.04} [0.03，0.04] 2 80-110 {1.10，1.20，1.10} [1.10，1.20] 3 110-140 {1.05，1.15，1.08} [1.05，1.15] As shown in Table 3, by statistically analyzing the differences between each sub-interval, the continuous flow waveform is transformed into discrete interval characteristics. The interval 0.03 to 0.04 indicates that the flow velocity changes gradually during this stage, while the interval 1.10 to 1.20 shows that the flow velocity is in a state of sharp increase. This result provides a quantitative basis for subsequent judgment of substantial changes in gas output.

[0035] The fluctuation determination submodule calls the range of traffic change amplitude and the time position of the data in the compression stage to determine the segment of the traffic change amplitude range. It compares the traffic change amplitude in the segment with the change amplitude of the adjacent segment to determine the fluctuation position and corresponds it with the data in the compression stage to generate the stage fluctuation indicator. The system retrieves the generated flow rate variation range and the time position information of the compression phase data, and performs fluctuation position locking based on interval differences. A comparison logic is set up to sequentially read the flow rate variation ranges of two adjacent analysis intervals, calculate the interval mean for each interval, and calculate the absolute value of the difference between the means of adjacent intervals. The calculation method is as follows: the interval mean equals the sum of the maximum and minimum values ​​of the interval divided by 2, and the interval difference equals the absolute value of the difference between the means of adjacent intervals. This difference represents the degree of abrupt change in gas flow acceleration within adjacent time periods. The calculated difference is compared with a preset fluctuation judgment benchmark value. The fluctuation judgment benchmark value is set based on a fluid dynamics experiment on the airway opening characteristics of a respirator. The experiment selected a standard adult respirator, and under the condition of simulating lung compliance of 0.05 liters per centimeter of water column, tested the jump in the rate of flow change during the process from static push to effective opening of the duckbill valve. Experimental data shows that at the instant the valve opens, the mean fluctuation of the flow rate is between 0.4 L / min and 0.8 L / min, while during the non-opening, uniform pressing or stationary phases, the fluctuation is below 0.1 L / min. To ensure effective identification and elimination of operational noise, the midpoint between the lower limit of the opening fluctuation (0.4 L / min) and the upper limit of the stationary fluctuation (0.1 L / min) is taken, and considering a certain safety margin, the fluctuation judgment benchmark is set at 0.25 L / min. Judgment: If the difference is greater than 0.25 L / min, it is determined that a substantial flow fluctuation has occurred at the boundary between the two analysis intervals, meaning the breathing bag has completed airway opening and entered an effective air delivery state. Taking the data in Table 3 as an example: the amplitude range of interval 1 (50 to 80 milliseconds) is 0.03 to 0.04, and the calculated mean is (0.03 + 0.04) / 2 = 0.035, resulting in 0.035 L / min. The amplitude range of interval 2 (80 to 110 milliseconds) is 1.10 to 1.20. The calculated mean is (1.10 + 1.20) / 2 = 1.15, resulting in 1.15 liters per minute. The difference between adjacent intervals is calculated as 1.15 minus the absolute value of 0.035, which equals 1.115 liters per minute. This result of 1.115 liters per minute is compared with the baseline value of 0.25 liters per minute. Since 1.115 is greater than 0.25, the boundary time point between interval 1 and interval 2 (i.e., 80 milliseconds) is determined as the location of the flow fluctuation. This time point (80 milliseconds) is recorded as the stage fluctuation indicator. If the difference between subsequent intervals 2 and 3 (with a mean of approximately 1.10 liters per minute) is 1.10 minus the absolute value of 1.15, i.e., 0.05 liters per minute, which is less than the baseline value, the indicator is not updated. Finally, 80 milliseconds was output as a fluctuation indicator for this compression phase, which indicates that the gas output characteristics of the breathing bag changed fundamentally at this moment, from a pre-compression state to an explosive exhaust state.

[0036] The position correction submodule calls the stage fluctuation indicator, obtains the stage start and end position sequence in the compressed stage data, performs position offset calculation on the stage fluctuation indicator, and adds the offset calculation value to the stage start and end position sequence for correction, generating the stage correction result. The generated stage fluctuation indicator (e.g., 80 milliseconds) is invoked, and the original start and end position sequence of the compression stage data (e.g., 50 milliseconds for the start point and 150 milliseconds for the end point) is obtained. A precise correction is then performed for the start position. Since gradient detection is primarily based on pressure changes, and pressure build-up often precedes actual airflow output (due to mechanical travel and valve opening resistance), the original start point is often earlier than the effective ventilation point. The deviation between the stage fluctuation indicator and the original start position is calculated by subtracting the original start position from the fluctuation indicator. Based on the aforementioned data, the deviation is calculated as: 80 - 50 = 30 milliseconds. This indicates that the originally identified compression start point is 30 milliseconds earlier than the actual effective flow output point. To obtain a stage division more consistent with clinical ventilation, a weighted offset superposition method is used to correct the original position. A correction coefficient is introduced, reflecting the degree of confidence in the hysteresis of the flow signal. This coefficient is typically set to 1.0 based on sensor response synchronicity experiments, meaning the effective start point is entirely based on the flow fluctuation point. The correction calculation is then performed: the corrected start position equals the original start position plus the deviation multiplied by the correction coefficient. Substituting the values: 50 + 30 * 1.0 = 80 milliseconds. For the correction of the termination position, the fluctuation indication of the flow rate falling edge is also detected. It is set that the flow rate drops from the high flow range to the low flow range at 140 milliseconds (fluctuation indication value is 140 milliseconds), while the original pressure termination point is 150 milliseconds. The calculated deviation is 140 - 150 = -10 milliseconds. The correction is performed: 150 + (-10 * 1.0) = 140 milliseconds. The corrected start position of 80 milliseconds and the termination position of 140 milliseconds are recombine, replacing the original start and end sequences, to generate the final stage correction result. This correction result adjusts the effective duration of the compression phase from the original 100 milliseconds (50 to 150 milliseconds) to 60 milliseconds (80 to 140 milliseconds). This adjustment eliminates the invalid pressure build-up period and residual pressure decay period, making the generated data more accurately correspond to the actual volume of gas pumped into the patient's lungs. The final output contains corrected data with a time window of 80 to 140 milliseconds, which serves as the sole time-domain reference for subsequent calculations of tidal volume and respiratory mechanics parameters.

[0037] Specifically, such as Figure 2 , 6 As shown, the integral sorting module is as follows: The pressure integration submodule, based on the stage correction results and the smoothed pressure sequence, obtains and accumulates the pressure sequence of continuous time points within multiple stages, multiplies the accumulated time interval with the pressure value and merges them according to the stage as the integration amount corresponding to the stage, judges the segment alignment at the stage boundary position, and generates the stage pressure integration amount. The system retrieves the stage correction results and the smoothed pressure sequence generated during the signal acquisition stage. Based on the corrected time window (e.g., 80 to 140 milliseconds for the first stage), it precisely indexes all discrete sampling points within that time period in the smoothed pressure sequence. Since the sampling frequency is fixed at 100 Hz, the time interval between adjacent sampling points is constant at 0.01 seconds. Discrete integration is performed to calculate the area under the pressure curve, which physically represents the pressure impulse during a single compression, i.e., the total driving force exerted by the respirator on the gas. The pressure amplitude at each sampling moment within the stage is read sequentially, and each pressure value is multiplied by the time interval 0.01 seconds to generate the instantaneous pressure integral at that moment. Subsequently, the instantaneous pressure integral components at all moments within the stage are accumulated to obtain the stage pressure integral. To ensure the consistency and comparability of the calculations, three consecutive compression processes are simulated, defined as stage 1, stage 2, and stage 3, respectively. The time window for stage 1 is 80 to 140 milliseconds. Extract the specific pressure data within Phase 1 and perform calculations. Detailed data is shown in Table 4.

[0038] Table 4. Data table for pressure integral calculation process in stage 1. Sampling time Smooth pressure value Time interval (s) Instantaneous integral components 80 2.05 0.01 0.0205 90 2.50 0.01 0.0250 100 3.20 0.01 0.0320 110 3.80 0.01 0.0380 120 4.10 0.01 0.0410 130 4.25 0.01 0.0425 140 4.30 0.01 0.0430 The instantaneous integral components are summed (0.0205, 0.0250, 0.0320, 0.0380, 0.0410, 0.0425, 0.0430), resulting in 0.0205 + 0.0250 + 0.0320 + 0.0380 + 0.0410 + 0.0425 + 0.0430 = 0.242, yielding a stage pressure integral of 0.242 kPa·s for stage 1. Similarly, the same boundary alignment and integration calculations are performed on subsequent monitored stages 2 (with a time window of 800 ms to 860 ms, indicating higher average pressure) and 3 (with a time window of 1500 ms to 1560 ms, indicating pressure decline). The integral calculated for stage 2 is set to 0.280 kPa·s, and the integral calculated for stage 3 is set to 0.265 kPa·s. Finally, a sequence of stage pressure integrals containing three consecutive values ​​(0.242, 0.280, 0.265) is generated, which quantifies the evolution trend of the operator's force intensity during continuous pressing.

[0039] The pressure relationship submodule calls the stage pressure integral quantity and the stage order in the stage correction result to perform a segment comparison of the stage pressure integral quantity. It uses the difference of the integral quantity in the segment comparison to perform an inter-segment difference calculation with the adjacent stage and arranges them according to the stage sequence to generate stage pressure relationship data. The generated sequence of stage pressure integrals and the stage order information from the stage correction results are used to perform dynamic difference analysis between adjacent stages. A sliding comparison window is constructed according to the chronological order of occurrence (stage 1, stage 2, stage 3), and the integral of the current stage is sequentially compared with the integral of the previous stage to perform a difference calculation. This calculation aims to capture changes in the force applied to the breathing bag, thereby determining the stability or trend of the ventilation process. The difference calculation logic is defined as subtracting the integral of the previous stage from the integral of the next stage to obtain the inter-segment difference value. First, stage 2 is compared with stage 1: the integral of stage 2 (0.280 kPa·s) and the integral of stage 1 (0.242 kPa·s) are read, and the subtraction operation 0.280 - 0.242 = 0.038 is performed, resulting in a difference value of positive 0.038 kPa·s. This positive value indicates that from stage 1 to stage 2, the operator's compression work significantly increases, meaning that the ventilation volume or driving pressure is in an upward process. Next, stage 3 was compared with stage 2: the integral value of stage 3 (0.265 kPa·s) and stage 2 (0.280 kPa·s) were read, and a subtraction operation was performed: 0.265 - 0.280 = -0.015, yielding a difference of -0.015 kPa·s. This negative value indicates that the work done by pressure begins to decrease from stage 2 to stage 3. The calculated difference values ​​were arranged in the sequence of comparison occurrences to generate stage pressure relationship data (+0.038, -0.015). This data not only reflects the magnitude of a single change but also constitutes the dynamic trajectory of the ventilation operation. To ensure the usability of the data for subsequent control, the validity of the difference values ​​needs to be verified by benchmark. The difference neglect dead zone value was set to 0.005 kPa·s. This value was set based on the sensor's own integral drift error experiment. Under static conditions, the sensor's integral drift amount does not exceed 0.003 kPa·s within 1 minute, and a safety factor of 0.005 was adopted. In this example, the absolute values ​​of 0.038 and 0.015 are both greater than 0.005, so they are both retained as valid pressure relationship data, accurately revealing the operational characteristic of "increase first and then decrease".

[0040] The segment sorting submodule performs sorting calculations on the differences in multiple segments within the stage pressure relationship data, matches the segment arrangement order obtained from the sorting calculation with the stage number in the stage correction result, and converts it into a quantified opening sequence to generate basic opening data. The system retrieves multi-segment difference values ​​from the stage pressure relationship data, performs quantitative grading and sorting based on the difference amplitude, and maps the grading results to a quantitative opening sequence for the actuators. The core logic is to dynamically adjust the opening of the airway valves based on the rate of change of the pressure integral to compensate for or respond to changes in operator force and maintain stable ventilation. A mapping table between difference values ​​and opening adjustment steps is pre-defined. This mapping relationship was determined through simulated lung compliance experiments: under standard compliance, for every 0.005 kPa·s increase in pressure integral difference, a 1-unit valve opening needs to be increased to balance the flow rate; conversely, it decreases. First, quantification calculations are performed on the difference value sequence (+0.038, -0.015). The baseline reference opening is set to 50 units (corresponding to standard ventilation). For the first difference value +0.038 kPa·s: it is divided by the quantization step size of 0.005 kPa·s, i.e., 0.038 / 0.005 = 7.6. A rounding down logic is then performed to determine 7 positive step levels. The corresponding opening value is calculated as follows: Basic opening 50 + (7 * 1) = 57, resulting in 57 units. This indicates that in the face of increased pressure, the opening is expanded to 57 to accommodate high flow demand. For the second difference value -0.015 kPa·s: divide it by the quantization step size of 0.005 kPa·s, i.e., -0.015 / 0.005 = -3. This is determined to be 3 negative step levels. The corresponding opening value is calculated as follows: Basic opening 50 + (-3 * 1) = 50 - 3 = 47, resulting in 47 units. This indicates that in the face of decreased pressure, the opening is contracted to 47 to maintain pressure. The calculated quantized opening values ​​(57, 47) are associated and combined with the corresponding subsequent stage numbers (Stage 3 and Stage 4) to generate basic opening data. The results of 57 units and 47 units indicate that, based on the pressure work changes in the previous stage, the precise target position that the valve should reach in the next moment has been calculated. Through this sorting and quantification mapping, abstract pressure trends are transformed into specific mechanical control commands, achieving closed-loop compensation control of the respirator's output characteristics. The final generated opening baseline data is then directly transmitted to the motor controller as a drive signal to execute real-time mechanical adjustments.

[0041] Specifically, such as Figure 2 , 7 As shown, the valve control generation module specifically consists of: The flow normalization submodule obtains the corresponding gas flow sequence based on the opening basic data, performs difference calculation on all gas flow values ​​in the sequence and the minimum gas flow value in the sequence, and then performs ratio calculation on the difference between the difference and the difference between the maximum and minimum gas flow values ​​in the sequence to generate a gas flow normalization sequence. The generated opening baseline data and the acquired gas flow sequence are retrieved. Based on the stage time window marked in the opening baseline data, a set of discrete gas flow values ​​for the corresponding time period is extracted. Taking stage 2 (high-pressure push stage) as an example, the gas flow sequence for this time period is read, and the sequence is set to contain four key sampling points with flow values ​​of 35.20 L / min, 36.50 L / min, 37.80 L / min, and 38.00 L / min, respectively. The sequence is traversed, and an extreme value filtering operation is performed to identify the minimum gas flow value of 35.20 L / min and the maximum gas flow value of 38.00 L / min. Subsequently, the global dynamic range of the sequence is calculated, that is, the difference between the maximum and minimum values, and the result is 38.00 - 35.20 = 2.80 L / min. This dynamic range represents the maximum amplitude of flow fluctuation within this compression stage. Based on this, a normalization operation is performed on each flow point in the sequence. The operation logic is as follows: first, the difference between the current flow value and the minimum flow value is calculated, and then the difference is divided by the global dynamic range. For the first sampling point of 35.20 liters per minute: the difference is 35.20 - 35.20 = 0.00, the ratio is 0.00 / 2.80 = 0.000, and the result is 0.000; for the second sampling point of 36.50 liters per minute: the difference is 36.50 - 35.20 = 1.30, the ratio is 1.30 / 2.80 = 0.464; for the third sampling point of 37.80 liters per minute: the difference is 37.80 - 35.20 = 2.60, the ratio is 2.60 / 2.80 = 0.929; for the fourth sampling point of 38.00 liters per minute: the difference is 38.00 - 35.20 = 2.80, the ratio is 2.80 / 2.80 = 1.000. The dimensionless values ​​obtained from the above calculations are arranged in their original time order to generate a normalized gas flow rate sequence (0.000, 0.464, 0.929, 1.000). This sequence eliminates the dimensional influence of the absolute flow rate values ​​and purely reflects the relative strength distribution of gas output over time, providing a unified scaling standard for subsequent waveform matching of flow characteristics with valve opening. Some calculation data are shown in Table 5.

[0042] Table 5. Gas Flow Rate Normalization Data Table Sampling point number raw traffic Difference operation result Dynamic range Normalization results 1 35.20 0.00 2.80 0.000 2 36.50 1.30 2.80 0.464 3 37.80 2.60 2.80 0.929 4 38.00 2.80 2.80 1.000 As shown in Table 5, the waveform fingerprint of the flow rate change was successfully extracted through normalization. The values ​​of 0.464 and 0.929 in the results accurately depict the intermediate state characteristics during the flow rate increase process.

[0043] The adjustment factor submodule calls the gas flow normalization sequence and the opening degree basic data to perform difference calculation, performs ratio analysis on the difference and the opening degree basic data, and arranges them according to the position of the gas flow normalization sequence to generate the opening degree adjustment factor sequence. The generated gas flow normalization sequence and the generated opening baseline data are used to perform a deviation analysis between the opening and flow waveforms. First, a unified comparison benchmark needs to be established, converting the opening baseline data to have the same dimensions as the normalization sequence. Given that the opening baseline data for stage 2, determined in the previous steps, is 57 units, and the full-stroke opening range of the solenoid valve is set to 100 units, the opening baseline data is calculated as 57 / 100 = 0.570. This value represents the average target opening percentage set for the current stage. Each value in the gas flow normalization sequence is read sequentially, and a difference calculation is performed with the opening baseline data to assess the deviation of the current instantaneous flow pattern from the average target opening. Taking the second point in the sequence, 0.464, as an example: the difference calculation is 0.464 - 0.570 = -0.106. This negative value means that at the current moment, the relative intensity of the flow is lower than the target opening level. Subsequently, a ratio analysis is performed between this difference and the opening baseline data to calculate the relative deviation rate. The calculation logic is to divide the difference by the base opening data. That is, -0.106 / 0.570 = -0.186. This value is defined as the opening adjustment factor, indicating that at the current moment, approximately 18.6% negative correction or damping compensation is needed for the base opening to match the lag characteristics of the actual flow. Similarly, for the third point 0.929: the difference is 0.929 - 0.570 = 0.359; the adjustment factor is 0.359 / 0.570 = 0.630. This indicates that the flow intensity at this moment is significantly higher than the average level, requiring the intervention of a positive adjustment factor. Perform the above calculation on all points in the sequence to generate the opening adjustment factor sequence (e.g., -1.000, -0.186, +0.630, +0.754). The advantage of this approach is that it does not only set a fixed opening based on average pressure, but also introduces a dynamic factor of the flow waveform, which allows the control to sense the "starting force" and "ending force" during the breathing bag compression process, providing a refined correction coefficient for achieving biomimetic breathing control.

[0044] The instruction combination submodule calls the opening adjustment factor sequence, performs linear combination calculation on the multiple factors and opening basic data within the sequence, reassembles the quantitative results obtained by linear combination in sequence order and transcribes them into control parameters, and generates anesthesia machine valve control instructions. The system calls the opening adjustment factor sequence and the opening base data to perform the final control command synthesis calculation. Linear combination logic is introduced to superimpose the dynamic adjustment factor onto the static base opening, generating a dynamic valve control curve that changes over time. The calculation formula for the linear combination is set as follows: the target proportion equals the opening base data plus (opening adjustment factor multiplied by opening base data multiplied by response weight coefficient). The setting of the response weight coefficient is crucial, as it determines the sensitivity to changes in the flow waveform. The coefficient is determined based on a ventilator-simulated lung ventilation stability experiment: different coefficient values ​​within the range of 0.1 to 0.5 were tested while maintaining airway pressure fluctuations of less than 2 cmH2O. Experimental data shows that when the coefficient is set to 0.2, it effectively tracks flow changes while avoiding valve oscillations caused by over-adjustment. Therefore, this embodiment sets the response weight coefficient to 0.2. Point-by-point calculations are performed on the data in the sequence: for the second point (adjustment factor is -0.186): the target proportion equals 0.570 + (-0.186 * 0.570 * 0.2). Intermediate calculation: -0.186 * 0.570 * 0.2 = -0.0212. Target ratio: 0.570 + (-0.0212) = 0.5488. Subsequently, this target ratio is converted into specific hardware control parameters. Based on the full range of 0 to 100, the control parameter equals the target ratio multiplied by 100 and rounded down. That is, 0.5488 * 100 = 54.88, which, after rounding down, generates the instruction value 54. For the third point (adjustment factor is 0.630): the target ratio equals 0.570 + (0.630 * 0.570 * 0.2). Intermediate calculation: 0.630 * 0.570 * 0.2 = 0.0718. Target ratio: 0.570 + 0.0718 = 0.6418. The control parameter was transcribed as 0.6418 * 100 = 64.18, which was rounded down to generate the command value 64. All calculated command values ​​were then rearranged in their original time sequence to generate the final anesthesia machine valve control command sequence (e.g., 46, 55, 64, 66). This result shows that although the basic opening setting for stage 2 is 57, after dynamic modulation of the flow waveform, the actual command output to the valve exhibits a dynamic trajectory that smoothly increases from 46 to 66. This dynamic adjustment strategy ensures that the anesthesia machine's gas supply valve can accurately match the actual hydrodynamic characteristics of manual compression of the breathing bag by medical personnel, achieving a digital replication of mechanical ventilation and manual techniques.

[0045] Please see Figure 8 An anesthesia machine control method is executed based on the aforementioned anesthesia machine control system, and includes the following steps: S1: Acquire electrical signals of the compression degree of the breathing bag and convert them from analog to digital to a pressure sequence. Smooth the pressure sequence and collect the corresponding gas flow rate to calculate the difference, generating pressure-flow rate difference data. S2: Perform gradient change detection on the pressure-flow differential data, identify the location of gradient abrupt changes, and perform segmentation on the smoothed pressure sequence to generate compression stage data; S3: Based on the compression stage data, obtain the corresponding gas flow rate and analyze the flow rate fluctuation. Combine the compression stage data to determine and correct the stage position, and generate the stage correction result. S4: Based on the stage correction results and the smoothed pressure sequence, calculate the stage pressure integral and compare it with the stage correction results and sort the segments to generate basic opening data; S5: Based on the opening baseline data, obtain the corresponding gas flow rate and normalize it to calculate the opening adjustment factor. Then, linearly combine the opening adjustment factor with the opening baseline data to generate the valve control command for the anesthesia machine.

[0046] The above are merely specific embodiments 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. An anesthesia machine control system, characterized in that, The system includes: The acquisition and conversion module acquires the electrical signal of the breathing bag compression degree and converts it into a pressure sequence from analog to digital. It then smooths the pressure sequence and acquires the corresponding gas flow rate for synchronous differential analysis, generating pressure-flow differential data and transmitting it to the gradient segmentation module. The gradient segmentation module performs gradient change detection on the pressure-flow differential data, identifies the location of gradient abrupt changes, performs segmentation on the smoothed pressure sequence, generates compression stage data, and transmits it to the stage correction module. The stage correction module acquires the corresponding gas flow rate based on the compression stage data and analyzes the flow rate fluctuation. It then combines the compression stage data to determine and correct the stage position, generates the stage correction result, and transmits it to the integration sorting module. The integral sorting module calculates the stage pressure integral based on the stage correction result and the smoothed pressure sequence, compares it with the stage correction result and sorts it by segment, generates the opening basic data and transmits it to the valve control generation module. The valve control generation module obtains the corresponding gas flow rate based on the opening basic data and calculates the opening adjustment factor by normalization. It then linearly combines the opening adjustment factor with the opening basic data to generate valve control commands for the anesthesia machine.

2. The anesthesia machine control system according to claim 1, characterized in that, The pressure-flow differential data includes a pressure differential sequence, a flow differential sequence, and a differential time index. The compression stage data includes a compression stage start point set, a compression stage end point set, and compression stage pressure characteristic parameters. The stage correction results include a stage position correction amount, a stage boundary correction index, and a stage validity judgment mark. The opening basic data includes a stage pressure integral, a valve opening initial ratio set, and stage sorting identifier information. The anesthesia machine valve control commands include a target valve opening sequence, an opening adjustment control coefficient set, and a valve drive execution sequence.

3. The anesthesia machine control system according to claim 1, characterized in that, The acquisition and conversion module is specifically: The signal acquisition submodule acquires the electrical signal of the breathing bag compression degree and performs discrete quantization on the electrical signal according to the voltage amplitude through analog-to-digital conversion. The quantized multi-point voltage sequence is arranged in time order and mapped into a pressure sequence to generate a pressure sequence value group. The sequence smoothing submodule calls the pressure sequence value group and performs smoothing processing based on the amplitude difference of adjacent sampling points. It weights the amplitudes of adjacent points in the sequence according to the weighted average coefficient and replaces the original amplitudes. It combines all the processed points in the original order to generate a smoothed pressure sequence value group. The differential generation submodule performs amplitude differential analysis on adjacent points based on the smoothed pressure sequence value group and the sequential relationship of the sequence points. It obtains the corresponding gas flow rate from the amplitude change obtained by differential analysis and performs synchronous pairing. It then organizes the differential components to generate pressure-flow differential data.

4. The anesthesia machine control system according to claim 1, characterized in that, The stepped segmentation module is specifically as follows: The gradient detection submodule, based on the pressure-flow differential data, performs gradient change detection according to adjacent differential components, compares the differential components with the gradient median value and calculates the numerical difference between the differential components and the gradient median value, and combines all the difference quantities in the original sequence order to generate a gradient change sequence value group. The mutation identification submodule calls the gradient change sequence value group, performs amplitude comparison based on adjacent differences in the sequence, marks and records the sequence positions of points where the difference exceeds the mutation benchmark value, and arranges all recorded positions in sequence order to generate a gradient mutation position sequence. The segmentation submodule performs segmentation based on the gradient mutation location sequence and the smoothed pressure sequence. It sequentially combines and divides the pressure points between adjacent mutation locations into multiple pressure segments, and then organizes the pressure segment sequence according to the segmentation order to generate compression stage data.

5. The anesthesia machine control system according to claim 4, characterized in that, The mutation baseline value is determined by sorting all the differences from low to high values ​​according to the overall distribution characteristics of multiple differences in the statistical gradient change sequence, and then obtaining the median difference and the dispersion of adjacent differences.

6. The anesthesia machine control system according to claim 1, characterized in that, The stage correction module specifically includes: The flow extraction submodule extracts the corresponding gas flow sequence based on the compression stage data, arranges the flow of multiple time sampling points in the sequence according to time and calculates the change in flow difference between adjacent sampling points, analyzes the range of change in multiple intervals in the sequence, and generates the flow change range. The fluctuation determination submodule calls the flow change range and the compression stage data time position to determine the flow change range segment, compares the flow change range within the segment with the change range of adjacent segments to determine the fluctuation position, and corresponds it with the compression stage data to generate a stage fluctuation indicator. The position correction submodule calls the stage fluctuation indicator, obtains the stage start and end position sequence in the compressed stage data, performs position offset calculation on the stage fluctuation indicator, and adds the offset calculation value to the stage start and end position sequence for correction, generating the stage correction result.

7. The anesthesia machine control system according to claim 1, characterized in that, The integral sorting module is specifically as follows: The pressure integration submodule, based on the stage correction results and the smoothed pressure sequence, obtains and accumulates the pressure sequence of continuous time points within multiple stages, multiplies the accumulated time interval with the pressure value and merges them according to the stage as the integration amount corresponding to the stage, judges the segment alignment at the stage boundary position, and generates the stage pressure integration amount. The pressure relationship submodule calls the stage pressure integral quantity and the stage order in the stage correction result to perform a segment comparison on the stage pressure integral quantity, uses the difference of the integral quantity in the segment comparison to perform an inter-segment difference calculation with the adjacent stage, and arranges them according to the stage sequence to generate stage pressure relationship data. The segment sorting submodule performs sorting calculations on the differences in multiple segments within the stage pressure relationship data, maps the segment arrangement order obtained from the sorting calculation to the stage number in the stage correction result, and converts it into a quantified opening sequence to generate basic opening data.

8. The anesthesia machine control system according to claim 1, characterized in that, The valve control generation module is specifically: The flow normalization submodule obtains the corresponding gas flow sequence based on the opening basic data and performs a difference operation on all gas flow values ​​in the sequence and the minimum gas flow value in the sequence. Then, it performs a ratio operation on the difference between the difference and the difference between the maximum and minimum gas flow values ​​in the sequence to generate a gas flow normalization sequence. The adjustment factor submodule calls the gas flow normalization sequence and the opening degree basic data to perform difference calculation, performs ratio analysis on the difference and the opening degree basic data, and arranges them according to the position of the gas flow normalization sequence to generate the opening degree adjustment factor sequence. The instruction combination submodule calls the opening adjustment factor sequence, performs linear combination calculations on multiple factors in the sequence and the opening basic data, reassembles the quantification results obtained by linear combination in sequence order and transcribes them into control parameters, and generates anesthesia machine valve control instructions.

9. A method for controlling an anesthesia machine, characterized in that, An anesthesia machine control system according to any one of claims 1-9 is executed, comprising the following steps: S1: Acquire electrical signals of the compression degree of the breathing bag and convert them from analog to digital to a pressure sequence. Smooth the pressure sequence and collect the corresponding gas flow rate to calculate the difference, generating pressure-flow rate difference data. S2: Perform gradient change detection on the pressure-flow differential data, identify the location of gradient abrupt changes, and perform segmentation on the smoothed pressure sequence to generate compression stage data; S3: Based on the compression stage data, obtain the corresponding gas flow rate and analyze the flow rate fluctuation. Combine the compression stage data to determine the stage position and correct the position, and generate the stage correction result. S4: Based on the stage correction results and the smoothed pressure sequence, calculate the stage pressure integral and compare it with the stage correction results and perform segment sorting to generate basic opening data; S5: Based on the opening baseline data, obtain the corresponding gas flow rate and normalize it to calculate the opening adjustment factor. Linearly combine the opening adjustment factor with the opening baseline data to generate anesthesia machine valve control instructions.

10. An anesthesia machine, comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor executes the computer program, it implements an anesthesia machine control system according to any one of claims 1 to 8.

Images