Wearable ventricular assist device for pediatric dcm and control method thereof

Abstract

The present application relates to the technical field of ventricular assist, in particular to a wearable ventricular assist device for children DCM and a control method thereof, the device comprising: a sensing feedback module, a flow rate acquisition module, a data alignment module, a perfusion determination module and an output control module.In the present application, by periodically collecting chest flexible pressure and temperature sensor signals, calculating the pressure stress and temperature fluctuation amplitude, and combining the data for dynamic comparison, the real-time evaluation of the perfusion state is carried out, the sensing characteristics are linked with the auxiliary pump parameters, the output liquid quantity is calculated and matched with the time period, the perfusion-pump speed correlation data is formed, the abnormal period is identified based on the double-index threshold value judgment, the output target is dynamically adjusted combining the flow rate data and the adjustment rule, the closed-loop control logic is overall constructed, the sensing, judgment and control are integrated and linked, the perfusion abnormality identification accuracy and response speed are improved, and the intelligent adjustment ability and individual adaptability of the device in dynamic circulation management are enhanced.

Description

Technical Field

[0001] This invention relates to the field of ventricular assist technology, and in particular to a wearable ventricular assist device for children with diabetic pulmonary vascular disease (DCM) and its control method. Background Technology

[0002] The field of ventricular assist technology involves medical devices used to assist or replace some functions of the heart. These devices primarily include various equipment that mechanically assist the heart's ejection function. Core aspects include blood pumping mechanisms, physiological parameter monitoring, power transmission methods, and in vivo / external coordinated control systems. This technology utilizes fluid dynamics methods, electrically driven structures, and structural designs compatible with the human circulatory system to achieve blood circulation support. Methodologically, this field focuses on the research and implementation of device implantation methods, power control strategies, and coordinated response mechanisms with human signals. Traditional ventricular assist devices (VADs) are primarily developed for adult patients to replace or assist ventricular ejection function. These devices address circulatory disorders caused by heart failure, employing externally driven electric centrifugal or axial flow pumps as the means of blood propulsion. They collect static parameters of the circulatory system through pressure sensors and adjust the drive in a constant-speed or intermittent control manner to help patients maintain a basic hemodynamic state.

[0003] Existing technologies primarily target adult patients, employing externally driven centrifugal or axial flow pumps. They collect static circulation parameters via pressure sensors and control output based on constant-rate or intermittent drive strategies, which presents several limitations. Under dynamic physiological conditions, the limited number of sensing parameters and the lack of linkage mechanisms with microscopic changes such as temperature and stress make it difficult to promptly capture trends in microcirculation perfusion, leading to low accuracy in identifying localized inadequate perfusion. Constant pump rate control lags in response to frequent changes in patient activity, failing to match actual perfusion needs and causing inconsistencies between short-term output fluid volume and tissue oxygen demand, leading to the risk of over- or under-perfusion. Flow rate regulation relies primarily on fixed rules, lacking refined control based on real-time sensing, easily resulting in non-individualized output strategies. The lack of a period numbering matching mechanism causes temporal misalignment between sensing data and flow rate data, affecting the system's accurate assessment of perfusion status. These shortcomings reduce the adaptability and intervention effectiveness of ventricular assist devices under frequently changing physiological conditions, especially in pediatric patients, where they are more prone to perfusion disturbances. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a wearable ventricular assist device for pediatric diabetic coronary syndrome (DCM) and its control method. The technical solution is as follows:

[0005] On the one hand, a wearable ventricular assist device for pediatric DCM is provided, the system comprising:

[0006] The sensing feedback module collects signal changes from the flexible pressure sensor and the thermal temperature sensor that fit the chest area, calculates the pressure stress amplitude and temperature fluctuation amplitude respectively, and compares them with the average values ​​of the previous several cycles to construct perfusion state response characteristic items and generate microcirculation perfusion sensing data set.

[0007] The flow rate acquisition module extracts the response data tag of the current cycle based on the feature items in the microcirculation perfusion sensing data set, and calculates the output liquid volume by combining the rotor speed and pump cavity parameters in the auxiliary pump drive, matching the sensing data cycle, and generating the auxiliary pump response perfusion data set.

[0008] The data alignment module, based on the flow rate information in the auxiliary pump response infusion data group and combined with the feature items in the sensing data group, completes the pairing through the cycle number of the control chip to form a wearable infusion-pump speed association group.

[0009] The perfusion determination module extracts the pressure stress amplitude and temperature fluctuation amplitude from the wearable perfusion-pump speed correlation group, and makes a bidirectional determination based on the preset perfusion threshold. If both indicators are lower than the threshold at the same time, it is marked as an abnormal perfusion cycle and a perfusion response insufficient label is output.

[0010] The output control module identifies abnormal cycles based on the insufficient perfusion response label, reads the flow rate value of the corresponding cycle, determines the output target in combination with the adjustment rules, and generates ventricular assist output adjustment control commands.

[0011] As a further embodiment of the present invention, the microcirculation perfusion sensing data set includes pressure stress amplitude, temperature fluctuation amplitude, and a comparison value between the current cycle sensing signal change value and the average level value; the auxiliary pump response perfusion data set includes execution cycle data tags, rotor speed, pump chamber parameters, and cycle output volume; the wearable perfusion-pump speed association set includes pressure stress amplitude, temperature fluctuation amplitude, cycle number, and time mapping result; the insufficient perfusion response tag includes cycle abnormality markers, low pressure stress value markers, and low temperature fluctuation value markers; and the ventricular auxiliary output adjustment control command includes the current cycle flow rate value, adjustment rule, and target pump speed output value.

[0012] As a further aspect of the present invention, the sensing feedback module includes:

[0013] The signal acquisition submodule acquires signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle. Based on the two types of sensor signals acquired in the current cycle, it calls the time series and change amplitude parameters of the signal data, filters out signal segments that meet the change trend, and generates the signal change segment value of the current cycle.

[0014] The amplitude calculation submodule calls the pressure change data and temperature change data in the current cycle signal change segment value, calculates the difference value of the fluctuation amplitude and change trend of the two types of signals based on the corresponding data sequence recorded in the previous few times of the current cycle, extracts the core parameters reflecting the amplitude state, and obtains the amplitude difference value of the current cycle.

[0015] The feature construction submodule determines the state classification standard to which the current signal feature belongs based on the changes in pressure and temperature amplitudes in the current cycle amplitude difference value, according to the set perfusion judgment parameters and state reference interval, establishes data feature items corresponding to the perfusion state, and generates microcirculation perfusion sensing data set.

[0016] As a further aspect of the present invention, the flow rate acquisition module includes:

[0017] The tag extraction submodule extracts data tags with the same timestamp in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data group, calls the time stamp and signal type parameters corresponding to the data tags, filters all tag information belonging to the current cycle range, and generates the current cycle response tag group.

[0018] The liquid volume calculation submodule calls the corresponding time period range in the current cycle response tag group, combines the rotor speed and pump cavity structure parameters recorded in the auxiliary pump drive unit, and calculates the output liquid volume of the pump cavity within the cycle based on the correlation time marker and structure comparison relationship of the two types of parameters to obtain the cycle liquid output volume.

[0019] The data matching submodule determines the correspondence between the period of the sensing data and the liquid volume data based on the periodic liquid output and the time tag recorded in the current periodic response tag group. It performs corresponding merging processing according to the periodic mark, establishes periodic matching data items between the auxiliary pump working output and the sensing tag, and generates the auxiliary pump response injection data group.

[0020] As a further aspect of the present invention, the process of extracting data tags with correlation in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data set is specifically defined as follows: the continuously collected perfusion state response feature items are time-aligned according to a fixed sampling time interval; when adjacent perfusion state response feature items are in the same time mark range and have the same signal type parameters, they are determined to have a correlation and the corresponding data tags are extracted.

[0021] The process of calculating the output liquid volume of the pump cavity within a cycle based on the correlation time markers and structural comparison relationships of the two types of parameters is specifically defined as follows: within the time period, the rotor speed recorded in the auxiliary pump drive unit is integrated at equal time intervals to obtain the cumulative rotor rotations within the corresponding time period, and based on the single-rotor output liquid volume or equivalent displacement parameters preset in the pump cavity structural parameters, the cumulative rotor rotations are converted into liquid volume to obtain the cycle liquid output volume.

[0022] The specific limitation of the merging process based on the periodic marker is that a periodic matching data item between the auxiliary pump working output and the sensing tag is only established when the time tag of the periodic liquid output completely falls within the time period range corresponding to the current periodic response tag group.

[0023] As a further aspect of the present invention, the data alignment module includes:

[0024] The information extraction submodule, based on the flow rate information recorded in the auxiliary pump response injection data group, calls the time tag and data identifier number parameters to extract the complete flow rate content recorded within the current cycle range, and organizes it into a structural sequence according to the cycle order to generate a flow rate cycle tag group.

[0025] The cycle pairing submodule calls the corresponding infusion response feature items in the flow rate cycle tag group and the sensing data group, performs tag comparison processing according to the cycle number in the control chip, filters data combinations with the same number, establishes data comparison items according to the number order, and obtains the cycle pairing mapping table.

[0026] The time mapping submodule extracts the time period information corresponding to the number according to the period numbering relationship marked in the period pairing mapping table, and synchronously maps the flow rate data with the perfusion feature items according to the time dimension, integrates them into a unified structure output format, and generates a wearable perfusion-pump rate association group.

[0027] As a further aspect of the present invention, the infusion determination module includes:

[0028] The index extraction submodule extracts the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, calls the number information and corresponding amplitude value in each cycle, filters the cycle data of the signal record, combines and constructs a dual index data structure, and generates a perfusion fluctuation parameter group.

[0029] The status judgment submodule calls the pressure stress amplitude and temperature fluctuation amplitude in the perfusion fluctuation parameter group. Based on the set pressure stress threshold and temperature fluctuation threshold, it performs interval judgment on the values ​​of the two indicators in each cycle. When the pressure stress amplitude is less than the corresponding pressure stress threshold and the temperature fluctuation amplitude is less than the corresponding temperature fluctuation threshold, it records the corresponding cycle number and obtains the perfusion insufficient cycle number group.

[0030] The anomaly marking submodule locates the corresponding cycle data in the wearable perfusion-pump speed association group based on the number content recorded in the insufficient perfusion cycle number group, establishes an anomaly marking field and writes it to the end of each cycle data item, outputs the marked data structure, and generates an insufficient perfusion response tag.

[0031] As a further aspect of the present invention, the output control module includes:

[0032] The status recognition submodule identifies the periodic status marked in the insufficient perfusion response label, calls the period number and status mark field, filters the periodic sequences with abnormal status, establishes an abnormal period index table, and generates abnormal period number groups.

[0033] The target calculation submodule calls the flow rate value in the auxiliary pump response injection data group corresponding to the abnormal cycle number group, matches it with the output reference interval in the regulation rule according to the current cycle number, calculates the regulation target value under the abnormal cycle, and obtains the output regulation target;

[0034] The signal generation submodule maps and encodes each flow rate regulation result in the output regulation target, combined with the abnormal cycle number and the preset control interface field definition, to construct a control instruction format that meets the output structure requirements and generate ventricular auxiliary output regulation control instructions.

[0035] As a further aspect of the present invention, the process of identifying the periodic status marked in the insufficient perfusion response label is specifically limited to comparing the period number recorded in the insufficient perfusion response label with the status mark field period by period, and only when the status mark field is a preset abnormal identifier value, the corresponding period number is included in the abnormal period number group.

[0036] The process of matching the current cycle number with the output reference interval in the adjustment rule is specifically defined as follows: the flow velocity value corresponding to the current cycle number is compared with the upper and lower limit parameters of the output reference interval for interval judgment; when the flow velocity value is outside the output reference interval, the corresponding adjustment amplitude parameter is selected according to the deviation direction, and the output adjustment target is calculated.

[0037] The process of constructing a control instruction format that meets the output structure requirements is specifically defined as follows: according to the preset field order in the control interface field, the abnormal cycle number, the flow rate regulation result, and the corresponding verification identifier are sequentially written into a data frame of fixed length to form the ventricular auxiliary output regulation control instruction.

[0038] On the other hand, a wearable ventricular assist control method for pediatric DCM, wherein the wearable ventricular assist control method for pediatric DCM is based on the aforementioned wearable ventricular assist device for pediatric DCM, includes the following steps:

[0039] S1: Obtain signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle, calculate the pressure stress amplitude and temperature fluctuation amplitude respectively, and compare them with the average level of the previous few collection values ​​in the current cycle to construct perfusion state response feature items and generate microcirculation perfusion sensing data set.

[0040] S2: Based on the perfusion state response feature item in the microcirculation perfusion sensing data group, extract the data tag that needs to be executed for the flow rate response in the current cycle, call the rotor speed and pump cavity parameters recorded in the auxiliary pump drive unit, calculate the output liquid volume of the cycle, and perform matching processing on the cycle corresponding to the sensing data to generate the auxiliary pump response perfusion data group.

[0041] S3: Based on the flow rate information recorded in the auxiliary pump response perfusion data group, combined with the perfusion response feature items corresponding to the sensing data group, the two data groups are mapped and integrated in the time dimension through the period number in the control chip to generate a wearable perfusion-pump speed association group.

[0042] S4: Extract the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, and make a two-way judgment against the preset perfusion state threshold. If both indicators of the current cycle are in the insufficient range at the same time, mark the cycle as perfusion abnormal and output the perfusion response insufficient label.

[0043] S5: Identify the cycle state marked by the insufficient perfusion response tag. If it is an abnormal state, use the current cycle flow rate value recorded in the auxiliary pump response perfusion data set, combine it with the adjustment rules to determine the current output target value, generate a pump speed control signal that can be issued, and generate a ventricular auxiliary output adjustment control command.

[0044] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0045] By periodically collecting signals from flexible chest pressure and temperature sensors, the system calculates pressure stress and temperature fluctuation amplitudes, and dynamically compares the data to conduct real-time assessments of the perfusion status. It links the sensing characteristics with auxiliary pump parameters to calculate the output volume and match the time cycle, forming perfusion-pump speed correlation data. Abnormal cycles are identified based on dual-index thresholds, and the output target is dynamically adjusted by combining flow rate data and regulation rules. This constructs a closed-loop control logic, integrating sensing, judgment, and control to improve the accuracy and response speed of perfusion anomaly identification, and enhances the device's intelligent adjustment capabilities and individual adaptability in dynamic cyclic management. Attached Figure Description

[0046] 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.

[0047] Figure 1 This is a schematic diagram of a wearable ventricular assist device for children with diabetic cerebral palsy (DCM) provided in an embodiment of the present invention;

[0048] Figure 2 This is a schematic diagram of the system framework of the present invention;

[0049] Figure 3 This is a flowchart of the sensing feedback module in this invention;

[0050] Figure 4 This is a flowchart of the flow velocity acquisition module in this invention;

[0051] Figure 5 This is a flowchart of the data alignment module in this invention;

[0052] Figure 6 This is a flowchart of the infusion determination module in this invention;

[0053] Figure 7 This is a flowchart of the output control module in this invention;

[0054] Figure 8 This is a flowchart of a wearable ventricular assist control method for children with diabetic pulmonary vascular disease (DCM) provided by an embodiment of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0056] 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.

[0057] This invention provides a wearable ventricular assist device for pediatric DCM, such as... Figure 1-2 The diagram shows a wearable ventricular assist device for pediatric DCM. The system includes:

[0058] The perception feedback module acquires signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle, calculates the pressure stress amplitude and temperature fluctuation amplitude respectively, and compares them with the average level of the collected values ​​in the previous N cycles (N is a preset positive integer) to construct perfusion state response feature items and generate microcirculation perfusion perception data set.

[0059] The flow rate acquisition module extracts the data tags that need to be executed in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data set, calls the rotor speed and pump chamber parameters recorded in the auxiliary pump drive unit, calculates the output liquid volume of the cycle, and performs matching processing on the cycle corresponding to the sensing data to generate the auxiliary pump response perfusion data set.

[0060] The data alignment module is based on the flow rate information recorded in the auxiliary pump response perfusion data group, combined with the corresponding perfusion response feature items in the sensing data group, and performs pairing processing on the flow rate information and feature items based on the preset period numbering rules. The two data groups are mapped and integrated in the time dimension to generate a wearable perfusion-pump speed association group.

[0061] The perfusion determination module extracts the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, and makes a two-way judgment against the preset perfusion state threshold. If both indicators of the current cycle are in the insufficient range at the same time, the cycle is marked as perfusion abnormal and a perfusion response insufficient label is output.

[0062] The output control module identifies the cycle status marked by the insufficient perfusion response tag. If it is an abnormal status, it uses the current cycle flow rate value recorded in the auxiliary pump response perfusion data set, combines it with the adjustment rules to determine the current output target value, and generates a pump speed control signal that can be issued, thus generating a ventricular auxiliary output adjustment control command.

[0063] The microcirculation perfusion sensing data set includes pressure stress amplitude, temperature fluctuation amplitude, and a comparison of the current cycle sensing signal change value with the average level. The auxiliary pump response perfusion data set includes execution cycle data tags, rotor speed, pump chamber parameters, and cycle output volume. The wearable perfusion-pump speed association set includes pressure stress amplitude, temperature fluctuation amplitude, cycle number, and time mapping results. The insufficient perfusion response tag includes cycle abnormality markers, low pressure stress value markers, and low temperature fluctuation value markers. The ventricular auxiliary output regulation control command includes the current cycle flow rate value, regulation rule, and target pump speed output value.

[0064] Specifically, such as Figure 2 , 3 As shown, the perception feedback module includes:

[0065] The signal acquisition submodule acquires signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle. Based on the two types of sensor signals acquired in the current cycle, it calls the time series and change amplitude parameters of the signal data, filters out signal segments that meet the change trend, and generates the signal change segment value of the current cycle.

[0066] A synchronous acquisition channel is connected to a flexible pressure sensor and a thermistor temperature sensor that fit the chest area. The start and end boundaries of the current cycle are locked. The cycle boundary is defined by the cycle number output by the control chip and the acquisition trigger edge. When the trigger edge arrives, the current cycle number is written and buffer writing is initiated. The acquisition frequency is set to 50 times per second. The pressure and temperature signals use the same clock source to align the sampling point numbers. The writing structure simultaneously writes a timestamp field and a sampling point number field, forming a double marker to eliminate alignment deviations caused by jitter in a single timestamp. After writing, null value removal is performed. The criterion is that if any sampling point has a missing field in either the pressure or temperature channel, that sampling point and one sampling point before and after it are removed, and a continuous valid sequence is reconstructed. Subsequently, denoising processing is performed. The pressure sequence undergoes 5-point moving median denoising, and the temperature sequence undergoes 5-point moving average denoising. The processing window scrolls incrementally according to the sampling point number within the current cycle. After denoising, the difference values ​​between adjacent sampling points are calculated for both sequences. The difference value is obtained by subtracting the previous value from the subsequent value of two adjacent points, resulting in the pressure change sequence and the temperature change sequence. Then, a trend integrity filter is performed on the change sequences. The filter action sequentially meets the following conditions: the length of a continuously increasing or decreasing segment is not less than 10 sampling points; the sign consistency rate of the difference values ​​within the segment is not less than 90%; and the average absolute value of the difference values ​​within the segment is not less than 0.02 kPa for pressure and not less than 0.02 degrees Celsius for temperature. Segments that meet these conditions are written into the current period signal change segment value structure. The structure fields sequentially include the segment start sampling point number, segment end sampling point number, segment pressure sequence, segment temperature sequence, segment pressure change sequence, and segment temperature change sequence, and a period number field is included as an index. In the example, a sampling result of 1.2 seconds per cycle was used. The pressure sequence fluctuated between 0.40 kPa and 0.62 kPa, and the temperature sequence fluctuated between 34.10 degrees Celsius and 34.60 degrees Celsius. After substituting the screening conditions, a continuous segment with the starting sampling point number 8 and the ending sampling point number 55 was obtained and written into the current cycle signal change segment value structure. The cycle number was written as 12, which served as the input data for the subsequent amplitude calculation submodule.

[0067] Table 1. Sensor Periodic Sampling Segment Data Table

[0068]

[0069] As shown in Table 1, the sequence in the table is the sampling point of the denoised segment. The sampling point number range corresponds to the sequence length. The subsequent amplitude calculation takes the complete sequence of the segment as input.

[0070] The amplitude calculation submodule calls the pressure change data and temperature change data in the current cycle signal change segment value, calculates the difference value of the fluctuation amplitude and change trend of the two types of signals based on the corresponding data sequence recorded in the previous few times of the current cycle, extracts the core parameters reflecting the amplitude state, and obtains the amplitude difference value of the current cycle.

[0071] The system reads the signal change segment structure of the current cycle, first retrieving the corresponding pressure and temperature segment sequences by cycle number, and then calling the corresponding data sequences recorded several times before the current cycle. The historical window length is set to the previous 20 cycles. Historical data is stored in the database as segment sequences from the same sensor, at the same location, and with the same sampling frequency. During storage, null values ​​identical to those of the current cycle have been removed and denoised. After reading the historical window, the system calculates the fluctuation amplitude and the difference in trend between the two types of signals. The fluctuation amplitude calculation is performed separately for two paths: the pressure fluctuation amplitude is the difference between the maximum and minimum values ​​of the pressure segment sequence, and the temperature fluctuation amplitude is the difference between the maximum and minimum values ​​of the temperature segment sequence. The median within the segment is also written for robustness control. The trend difference value calculation uses the pressure and temperature change sequences of the current cycle as input. First, the two change sequences are mapped to three-valued trend codes. A difference greater than 0.01 kPa is considered an increase, a difference less than -0.01 kPa is considered a decrease, and values ​​in between are considered stable. The temperature trend code is mapped using a 0.01 degree Celsius threshold. Next, consistency statistics are performed on the two trend codes at the same sampling point. The statistical action accumulates the number of consistent and inconsistent values ​​across the entire segment. The number of inconsistent values ​​is further subdivided into four categories: pressure increase and temperature decrease, pressure decrease and temperature increase, stable pressure and temperature change, and stable temperature and pressure change, forming a trend difference count vector. Then, the trend difference count vector is compared with the corresponding vectors of the 20 cycles within the historical window to calculate the current cycle's relative historical median difference. The difference is defined as the current cycle's number of inconsistent values ​​minus the historical median number of inconsistent values. The core parameter extraction action extracts five items from the above results: pressure fluctuation amplitude, temperature fluctuation amplitude, number of trend inconsistencies, percentage of trend inconsistency types, and relative historical median difference, and writes them into the current cycle amplitude difference value structure. In the example, the pressure fluctuation amplitude for cycle number 12 was 0.22 kPa, the temperature fluctuation amplitude was 0.50 degrees Celsius, the number of trend inconsistencies was 9, the median number of trend inconsistencies within the historical window was 6, and substituting these into the difference definition yielded a relative historical median difference of 3. The sampling segment length and historical window length fields are written to the end of this structure, forming the input for subsequent feature construction. The pressure sensor can cover a range of 0 to 60 kPa. Common variations in chest-fitting conditions fall within the low-pressure range, and the common range for skin and torso surface temperature falls within 33.5 degrees Celsius to 36.9 degrees Celsius. The amplitude threshold setting and anomaly judgment use this range as the upper bound reference to avoid judgment failure due to thresholds exceeding the measurable range.

[0072] The feature construction submodule determines the state classification standard to which the current signal feature belongs based on the changes in pressure and temperature amplitudes in the current cycle amplitude difference value, according to the set perfusion judgment parameters and state reference interval, establishes data feature items corresponding to the perfusion state, and generates microcirculation perfusion sensing data set.

[0073] The current cycle amplitude difference value structure is read first. Then, the perfusion judgment parameters and state reference intervals are loaded. The parameter loading is derived from the combined results of preoperative calibration experiments and in vivo continuous observation experiments. The calibration experiment used the same type of flexible pressure sensor and thermistor temperature sensor to continuously collect data at a fixed position on the chest for 30 minutes, with a sampling frequency of 50 times per second. The data was then sliced ​​according to cycle number to generate a baseline amplitude distribution. The in vivo continuous observation experiment used a wearable ventricular assist pump to collect data for 2 hours under stable support conditions. The data was then sliced ​​according to cycle to generate a support phase amplitude distribution. The parameter loading fields include the lower limit of pressure stress amplitude, the upper limit of pressure stress amplitude, the lower limit of temperature fluctuation amplitude, the upper limit of temperature fluctuation amplitude, the lower limit of trend difference, and the upper limit of trend difference. These are also written into a state classification label set, which uses three categories: normal response, mild insufficiency, and significant insufficiency. Interval settings are performed using experimental data quantile statistics. The lower limit of pressure stress amplitude is taken as the larger value of the 10th percentile of the baseline distribution and the 10th percentile of the support phase distribution. The upper limit of pressure stress amplitude is taken as the smaller value of the 90th percentile of the two distributions. The temperature fluctuation amplitude and trend difference are set in the same way. The example uses calibration and in vivo data statistics. The lower limit of pressure stress amplitude is set at 0.05 kPa, the upper limit at 0.35 kPa, the lower limit of temperature fluctuation amplitude is set at 0.10 degrees Celsius, the upper limit at 0.80 degrees Celsius, and the upper limit of trend difference is set at 12 times. A state classification action is then performed, prioritizing significant deficiency. If the pressure stress amplitude is less than 0.05 kPa and the temperature fluctuation amplitude is less than 0.10 degrees Celsius, and the number of trend inconsistencies is greater than 12 times or the relative historical median difference is greater than 5 times, then significant deficiency is recorded. If the pressure stress amplitude is between 0.05 kPa and 0.10 kPa or the temperature fluctuation amplitude is between 0.10 degrees Celsius and 0.18 degrees Celsius, and the number of trend inconsistencies is between 9 and 12 times, then mild deficiency is recorded. All other values ​​are recorded as normal response. In the example, cycle number 12 has a pressure stress amplitude of 0.22 kPa, a temperature fluctuation amplitude of 0.50 degrees Celsius, and 9 instances of trend inconsistency. Substituting these values ​​into the aforementioned discrimination rules, both the pressure stress amplitude and temperature fluctuation amplitude fall within the corresponding normal response range. Therefore, the current cycle's state is classified as a normal response. Subsequently, data feature items corresponding to the state are established. The feature item fields are sequentially written as cycle number, pressure stress amplitude, temperature fluctuation amplitude, number of trend inconsistencies, relative historical median difference, state classification label, and effective sampling segment length. These are then summarized and written into the microcirculation perfusion sensing data set. The data set is appended with the state classification label from the previous cycle for subsequent label extraction and association retrieval.

[0074] Specifically, such as Figure 2 , 4 As shown, the flow rate acquisition module includes:

[0075] The tag extraction submodule extracts data tags with the same timestamp in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data set. It then calls the time stamp and signal type parameters corresponding to the data tags, filters all tag information belonging to the current cycle range, and generates the current cycle response tag set.

[0076] The system reads the microcirculation perfusion sensing data set, locates the perfusion status response feature items according to the current cycle number, and first generates correlation data tags. The correlation rule uses a combination of status classification tags and amplitude fields for mapping. Normal response is mapped to the stable tag in the tag set, mild insufficiency is mapped to the attention tag, and significant insufficiency is mapped to the alarm tag. At the same time, under the attention and alarm tags, it is further subdivided into two categories: pressure-dominated and temperature-dominated. The subdivided action compares the distance of the pressure stress amplitude and temperature fluctuation amplitude relative to their respective lower limits, and the smaller distance corresponds to the dominant category. After generating tags, the system calls the time stamp and signal type parameters to perform tag filtering. The time stamp comes from the segment start and end sampling point sequence number and sampling timestamp written by the signal acquisition submodule. The signal type parameter takes three categories: pressure, temperature, and dual-channel. The filtering action first limits the scope of belonging, and the limiting condition is that the timestamp falls within the start and end boundary of the current cycle and the segment sampling point sequence number is continuous without gaps. Then, all tag information within the scope is written to the current cycle response tag set. The written fields record the cycle number, tag main class, tag subclass, segment start timestamp, segment end timestamp, segment duration, signal type, and segment effective length in sequence. In this example, period number 12 was identified as a tag of interest and was pressure-driven, with a segment duration of 0.96 seconds and a signal type of dual-channel. This tag was written into the current period's response tag group. Reserved index fields for the previous and next periods were appended to the end of this tag group. The next period field was backfilled when the next period was written, for use in subsequent data matching to perform cross-period boundary checks.

[0077] The liquid volume calculation submodule calls the corresponding time period range in the current cycle response tag group, combines the rotor speed and pump cavity structure parameters recorded in the auxiliary pump drive unit, and calculates the output liquid volume of the pump cavity within the cycle based on the correlation time mark and structural comparison relationship of the two types of parameters to obtain the cycle liquid output volume.

[0078] The system reads the time range from the current cycle response tag group. First, it calculates the pump drive record interval covered by the segment's start and end timestamps. The pump drive unit record fields include rotor speed, pump cavity structural parameter version number, impeller diameter, effective pump cavity volume, outlet cross-sectional area calibration value, and calibration coefficient. The calculation is time-marked and aligned. First, it retrieves the drive record closest to and no later than the segment's start timestamp as the interval's starting point. Then, it retrieves the drive record closest to and no earlier than the segment's end timestamp as the interval's ending point. Each drive record within the interval is weighted and summarized according to its duration to obtain the interval's representative rotor speed. The pump cavity output liquid volume is calculated using a structural comparison table. The structural comparison table uses the structural parameter version number as the key to look up the output calibration liquid volume per revolution and the speed interval correction coefficient. Subsequently, the interval output liquid volume derivation is performed. The interval representative rotor speed and segment duration are substituted into the derivation logic. The derivation logic uses the output calibration liquid volume per revolution and the speed correction coefficient together. First, the interval representative rotor speed and duration are converted into the interval cumulative revolutions. Then, the cumulative revolutions are multiplied by the output calibration liquid volume per revolution to obtain the preliminary output liquid volume. Finally, the preliminary output liquid volume is multiplied by the correction coefficient to obtain the periodic liquid output volume. The example data uses structural parameter version number 3, a calibration output liquid volume per revolution of 0.00032 liters, a speed correction coefficient of 0.98, an interval representative rotor speed of 3200 revolutions per minute for period number 12, and a segment duration of 0.96 seconds. Substituting these values ​​into the derivation logic yields a periodic liquid output volume of 0.016 liters. This result is written to the periodic liquid output volume field and appended to the interval representative rotor speed and structural parameter version number fields. The commonly used flow rate of wearable ventricular assist pumps for children can fall within the range of 0.3 liters per minute to 1.5 liters per minute or 0.5 liters per minute to 1.5 liters per minute. In the example, the minute output is approximately 1.0 liters per minute, which is within this range.

[0079] Table 2 Pump Drive Record and Cyclic Liquid Volume Calculation Data Table

[0080]

[0081] As shown in Table 2, the periodic liquid output is derived from the segment duration and the rotor speed represented by the interval. The structural parameter version number is associated with the calibrated output liquid volume per revolution and the correction coefficient. Subsequent data matching directly calls the periodic liquid output field.

[0082] The data matching submodule determines the correspondence between the period of the sensing data and the liquid volume data based on the periodic liquid output and the time tag recorded in the current periodic response tag group. It performs corresponding merging processing according to the period tag, establishes periodic matching data items between the auxiliary pump working output and the sensing tag, and generates the auxiliary pump response injection data group.

[0083] The system reads the periodic liquid output and the current periodic response tag group, first performing a period consistency check. This check compares the period number in the pump drive record with the period number in the response tag group. If they match, it proceeds to matching and writing; otherwise, it performs a time backtracking check. The backtracking check uses the segment start and end timestamps as boundaries, extending 0.20 seconds forward and backward in the pump drive record to search for a record segment with the same period number. If the search is successful, the period number is written back to the response tag group, and the check source is marked as backtracking check. After period consistency is achieved, a merging process is performed. The merging action establishes a matching data item between the auxiliary pump's working output and the sensing tag, using the period number as the primary key. The matching data item fields are written sequentially as follows: period number, periodic liquid output, range-representing rotor speed, tag main class, tag subclass, segment duration, segment start and end timestamps, signal type, and status classification tag. If multiple tag records exist within the same period, the merging process writes according to tag priority: alarms take precedence over concerns, and concerns take precedence over stability. Simultaneously, the segment durations of multiple tags are summed and written to the overwrite duration field. In the example, cycle number 12 has one record with a focus tag and a coverage time of 0.96 seconds. The corresponding cycle liquid output of 0.016 liters and the interval representative rotor speed of 3200 rpm are written into the same matching data item. This set of matching data items is sorted in ascending order by cycle number to form an auxiliary pump response injection data group. A cross-cycle integrity check field is appended to the end of the data group. This field records the number of missing consecutive cycles and a list of missing cycle numbers, which is used to avoid broken sequences from entering alignment during subsequent information extraction.

[0084] Specifically, such as Figure 2 , 5 As shown, the data alignment module includes:

[0085] The information extraction submodule uses the flow velocity information recorded in the auxiliary pump response injection data group to call the time tag and data identifier number parameters to extract the complete flow velocity content recorded within the current cycle range, and organizes it into a structural sequence according to the cycle order to generate a flow velocity cycle tag group.

[0086] The auxiliary pump response perfusion data set is read, and the recorded flow rate information is extracted and structured. The flow rate information is derived from the periodic fluid output and segment duration. The derivation process converts the periodic fluid output by the segment duration to the instantaneous output flow rate, with units of liters per minute (L / min). The conversion process converts the output per second to the output per minute. The extraction process first calls the timestamp and data identifier parameters. The timestamp is a combination of the segment start and end timestamps and the period number. The data identifier is an auto-incrementing number within the data set. Then, complete flow rate records within the current period are filtered. Completeness is determined using three constraints: the first constraint is that the periodic fluid output field is not empty and is greater than 0.0005 liters; the second constraint is that the segment duration field is greater than 0.20 seconds; and the third constraint is that the range representing the rotor speed falls within the range of 1000 rpm to 5000 rpm. This speed range comes from the common operating speed range of centrifugal blood pumps; data outside this range is marked as invalid and not included in subsequent alignment. After filtering, the data is organized into a structural sequence according to the cycle order. The structural sequence fields are: cycle number, segment start and end timestamps, segment duration, cycle liquid output, converted flow rate, range-representing rotor speed, main tag class, and sub-tag class. In the example, cycle number 12 has a cycle liquid output of 0.016 liters and a segment duration of 0.96 seconds. Substituting these values ​​into the conversion process yields a converted flow rate of approximately 1.00 liters per minute. The range-representing rotor speed of 3200 revolutions per minute meets the range constraint, and this entry is written into the structural sequence. The structural sequence is output as a flow rate cycle tag group in ascending order of cycle number, with the start and end cycle number fields appended to facilitate cycle pairing and boundary judgment.

[0087] The cycle pairing submodule calls the corresponding infusion response feature items in the flow rate cycle tag group and the sensing data group, performs tag comparison processing according to the cycle number in the control chip, filters data combinations with the same number, establishes data comparison items according to the number order, and obtains the cycle pairing mapping table.

[0088] The system reads perfusion response feature items from the flow rate cycle label group and the microcirculation perfusion sensing data group, and uses the cycle number in the control chip as the unique comparison key to perform label comparison processing. The comparison process first establishes cycle number index tables for both data groups, with fields recording the cycle number, row number position within the group, and timestamp range summary. Then, a consistency filter is performed. The filter retrieves cycle numbers one by one from the flow rate cycle label group and searches for entries with the same number in the perfusion sensing data group index table. If the search is successful, a data comparison item is created; otherwise, it is written to the missing mapping list and skipped. The fields written to the comparison item are, in order: cycle number, flow rate entry row number, perfusion entry row number, flow rate timestamp range, perfusion timestamp range, main label class, and state classification label. To avoid multiple matches due to multiple perfusion feature items in the same cycle, the pairing submodule performs a unique selection of perfusion items, prioritizing the effective fragment length. If the effective lengths are the same, priority is given to those with a larger number of inconsistent trends. In the example, cycle number 12 exists in both index tables, with flow rate entry row number 5 and injection entry row number 7. A comparison item is created and the main tag is set to "Attention" and the status category tag is "Slightly Insufficient". All comparison items are output in cycle number order as a cycle pairing mapping table. A mapping coverage field is written at the end of the mapping table. The coverage is obtained by dividing the number of successfully paired cycles by the number of cycles in the flow rate cycle tag group. The coverage in this example is 0.96. When the coverage is lower than 0.90, an alarm field will be triggered by the upper-layer interface, but the contents of the mapping table will not be changed.

[0089] The time mapping submodule extracts the time period information corresponding to the number based on the period number relationship marked in the period pairing mapping table, synchronously maps the flow rate data with the perfusion feature items according to the time dimension, integrates them into a unified structure output format, and generates wearable perfusion-pump rate association group.

[0090] The cycle pairing mapping table is read, and the time period information corresponding to the marked cycle numbers is extracted. The time period information comes from two types of fields: velocity timestamp range and injection timestamp range. The synchronization mapping action first determines a unified time axis, which uses the start and end points of the velocity timestamp range as the main axis boundary. If the injection timestamp range falls completely within the main axis boundary, it is directly aligned. If it is ahead or behind, boundary clipping is performed. The clipping action clips the injection timestamp range to within the main axis boundary and records the clipping duration field. Then, the velocity data and injection feature items are synchronously mapped according to the time dimension. The mapping action establishes a unified structure output entry for each cycle. The entry fields are written in the following order: cycle number, main axis start timestamp, main axis end timestamp, converted velocity, interval representative rotor speed, pressure stress amplitude, temperature fluctuation amplitude, number of trend inconsistencies, relative historical median difference, state classification label, label main class, label subclass, and clipping duration. In the example, the main axis time for cycle number 12 is from the start timestamp 10:00:12.120 to the end timestamp 10:00:13.080, and the infusion time range is from 10:00:12.160 to 10:00:13.040, falling within the main axis boundary. The trimming duration is 0.00 seconds. The entries include a converted flow rate of 1.00 L / min, a pressure stress amplitude of 0.22 kPa, and a temperature fluctuation amplitude of 0.50 degrees Celsius. All entries are output as wearable infusion and pump rate association groups in ascending order of cycle number, and an association group verification field is appended at the end. The verification field records the number of trimmings and the average trimming duration, serving as quality constraint inputs for subsequent indicator extraction.

[0091] Specifically, such as Figure 2 , 6 As shown, the infusion determination module includes:

[0092] The index extraction submodule extracts the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, calls the number information and corresponding amplitude value in each cycle, filters the cycle data of the signal records, combines and constructs a dual index data structure, and generates a perfusion fluctuation parameter group.

[0093] The wearable perfusion and pump rate correlation group was read, and the pressure stress amplitude and temperature fluctuation amplitude fields were extracted to construct a dual-indicator data structure. The extraction process first filtered the periodic data of the signal records by period number. The filtering criteria were that the average clipping time in the correlation group's verification field was no greater than 0.05 seconds, and the clipping time of the corresponding period entry was 0.00 seconds or no greater than 0.02 seconds. Entries exceeding the threshold were added to the exclusion list and not included in the inadequate perfusion determination. Subsequently, a perfusion fluctuation parameter group was constructed. The parameter group entry fields were sequentially written as period number, pressure stress amplitude, temperature fluctuation amplitude, converted flow rate, interval representative rotor speed, state classification label, and label main class. To support the reproducibility of the threshold setting, the indicator extraction submodule also wrote a threshold reference data index, which pointed to the baseline distribution statistics under the same body position and fitting pressure conditions in the experimental database. The threshold reference experimental design used two control groups: the first group was data from the stable support period, and the second group was data from the clinically recorded inadequate perfusion event window. The event window was labeled by the clinicians, and the labeling rule was that a decrease in skin temperature and a synchronous decrease in pressure amplitude occurred for three consecutive periods. The 10th percentile of the pressure stress amplitude and temperature fluctuation amplitude were calculated for both sets of data. The 10th percentile of the stable support period was used as a candidate lower limit, and the 90th percentile of the inadequate perfusion event window was used as the upper limit of the lower limit. The lower limit was chosen as the numerical point between the two that best distinguished the two sets. In the example, the lower limit of the pressure stress threshold was found to be 0.08 kPa, and the lower limit of the temperature fluctuation threshold was 0.15 degrees Celsius. The pressure stress amplitude of 0.22 kPa and the temperature fluctuation amplitude of 0.50 degrees Celsius in period number 12 were written into the perfusion fluctuation parameter group entry. The advantage of this calculation logic is that by having the pressure stress amplitude and temperature fluctuation amplitude enter the same entry and record them alongside the converted flow rate, a unified discriminant input that can be used for cross-period comparison is directly formed, reducing misjudgments caused by single-index fluctuations.

[0094] The status judgment submodule calls the pressure stress amplitude and temperature fluctuation amplitude in the perfusion fluctuation parameter group. Based on the set pressure stress threshold and temperature fluctuation threshold, it performs interval judgment on the values ​​of the two indicators in each cycle. When the pressure stress amplitude is less than the corresponding pressure stress threshold and the temperature fluctuation amplitude is less than the corresponding temperature fluctuation threshold, it records the corresponding cycle number and obtains the perfusion insufficient cycle number group.

[0095] The perfusion fluctuation parameter set is read, and pressure stress threshold and temperature fluctuation threshold are loaded to perform interval judgment. The interval judgment action checks whether the pressure stress amplitude and temperature fluctuation amplitude are within the lower limit interval for each cycle. The lower limit interval is defined as less than the corresponding lower threshold and not less than 0, avoiding negative values ​​from entering the judgment. The cycle numbers that are simultaneously within the lower limit interval are then recorded and written into the inadequate perfusion cycle number group. To meet the requirements for experimental verification of parameter values ​​and complete data disclosure, the state judgment submodule uses the confounding statistics results from the same batch of experiments for comparison. The experimental sample size is 1200 cycles, including 240 clinically labeled inadequate perfusion cycles and 960 non-inadequate cycles. After using dual threshold discrimination, 220 cycles were determined to be inadequate, of which 200 were consistent with clinical labeling, and 20 non-inadequate cycles were misjudged as inadequate. The comparison scheme uses only the skin temperature threshold for single threshold discrimination, resulting in 260 cycles determined to be inadequate, of which 180 were consistent and 80 were misjudged. Based on the above data, the correct identification rate of dual-threshold discrimination is 11.1% higher than that of single-threshold discrimination, while the false identification rate is 75.0% lower. In the example of cycle number 12, the pressure stress amplitude of 0.22 kPa and the temperature fluctuation amplitude of 0.50 degrees Celsius do not fall within the lower limit range and are not written into the inadequate perfusion cycle number group. The state judgment submodule performs consistency verification between the state classification labels of the inadequate perfusion cycle number group and the associated group at the end. If the inadequate perfusion judgment result is insufficient but the state classification label is normal response, it is written into the conflict flag field and output for the exception flag submodule to write to the end of the data structure. The experimental results show that dual-threshold discrimination improves the correct identification rate and reduces the false identification rate compared with single-threshold discrimination.

[0096] Table 3 Comparison of experimental results between dual-threshold and single-threshold discrimination

[0097]

[0098] Referring to Table 3, dual threshold discrimination outperforms single threshold discrimination in both the number of consistent cases and the number of false cases. The status judgment submodule writes the comparison result into the verification record field corresponding to the threshold version number for easy traceability later.

[0099] The anomaly marking submodule locates the corresponding cycle data in the wearable perfusion-pump speed association group based on the number content recorded in the insufficient perfusion cycle number group, establishes an anomaly marking field and writes it to the end of each cycle data item, outputs the marked data structure, and generates an insufficient perfusion response label.

[0100] First, a mapping index is established from cycle number to associated group row number. Then, the index is searched row by row according to the cycle number group with insufficient perfusion. If the search is successful, the target cycle entry is located. If the search fails, a missing list is written with a missing reason field. The reason field is marked with three categories: unpaired, pruned beyond limit, and data missing. The write action adds an anomaly flag field to the end of each cycle entry. The field value is limited to two categories: abnormal and non-abnormal, and fuzzy quantifiers are not used. An anomaly type field is added simultaneously with the write operation. The anomaly type is determined by the triggering condition. If both indicators reach their lower limits simultaneously, the lower limits of both indicators are written. If there is a conflict flag field, a conflict discrimination field is written. The output data structure of the labeled data after the write operation is a label for insufficient perfusion response. To ensure the reproducibility of the labeled data structure, the version numbers of the pressure stress threshold and temperature fluctuation threshold used in this operation are written to the label header and the threshold value field. The threshold value field is set to 0.08 kPa and 0.15 degrees Celsius. In the example, cycle number 12 was not included in the insufficient injection cycle number group. The anomaly flag field of the corresponding entry was written as "not abnormal," and the anomaly type field was written as "null value." The tag header still recorded the threshold version number and threshold value. This tag structure serves as the input to the output control module, and the order of its internal fields remains fixed to avoid ambiguity in interface parsing.

[0101] Specifically, such as Figure 2 , 7 As shown, the output control module includes:

[0102] The status recognition submodule identifies the periodic status marked in the insufficient infusion response tag, calls the period number and status mark fields, filters the periodic sequences with abnormal status, establishes an abnormal period index table, and generates abnormal period number groups.

[0103] The abnormal cycle index table fields are sequentially written with cycle number, abnormality type, spindle timestamp range, converted flow rate, range-represented rotor speed, pressure stress amplitude, temperature fluctuation amplitude, main tag category, and status classification tag. To avoid isolated abnormalities caused by sporadic noise, the status recognition submodule performs continuity verification. The verification action groups abnormal cycles by cycle number. Adjacent cycle numbers within a group with a difference of 1 are considered continuous. Groups with a continuous length of at least 3 are written into the abnormal cycle number group, while groups with a continuous length of less than 3 are written into the pending review list with a review mark field. In the example, if cycle numbers 20 to 22 are all marked as abnormal, then the abnormal cycle number group is written with 20, 21, and 22; if only cycle number 25 is marked as abnormal, then it is written into the pending review list. After the abnormal cycle number group is generated, the minimum cycle number, maximum cycle number, and percentage of abnormal types within the group are appended to it for the target calculation submodule to select control rules. The advantage of this operation logic is that by expanding the abnormal cycle from a single point to a sequence index through continuity verification, the output adjustment target can match the inertial response time recorded by the pump speed.

[0104] The target calculation submodule calls the flow rate value in the auxiliary pump response injection data group corresponding to the abnormal cycle number group, matches it with the output reference interval in the regulation rule according to the current cycle number, calculates the regulation target value under the abnormal cycle, and obtains the output regulation target;

[0105] The adjustment rules are derived from bench flow calibration and in vivo safety range verification. The output baseline range is set according to the child's weight stratification: a weight of 3 kg to 15 kg corresponds to a baseline flow rate range of 0.3 L / min to 1.5 L / min, and a weight of 15 kg to 30 kg corresponds to a baseline flow rate range of 1.0 L / min to 4.0 L / min. The range is based on the common design capabilities of pediatric ventricular assist pumps and is solidified in conjunction with clinical verification records. The matching action reads the child's weight field, which is obtained from the preoperative registration form and written into the control chip configuration area. The example weight is 12.0 kg, corresponding to a baseline flow rate range of 0.3 L / min to 1.5 L / min. The target calculation process is executed for each abnormal cycle. First, the converted flow rate, pressure stress amplitude, and temperature fluctuation amplitude of the abnormal cycle are read. The target increment is calculated using a tiered mapping. When both indicators have lower limits and the converted flow rate is below the lower limit of the baseline interval, the target flow rate is the lower limit of the baseline interval plus 0.10 liters per minute. When both indicators have lower limits but the converted flow rate is already within the baseline interval, the target flow rate is the current flow rate plus 0.05 liters per minute. If the abnormality type is a conflict, the target flow rate remains unchanged and is marked for verification. In the example, the current converted flow rate for abnormal cycle number 20 is 0.28 liters per minute. Substituting the lower limit of the baseline interval (0.3 liters per minute) and the superposition rule yields a target flow rate of 0.40 liters per minute. After obtaining the target flow rate, the target rotor speed is derived. The derivation logic calls the output calibration volume per revolution and correction coefficient corresponding to the structural parameter version number. First, the target flow rate is converted into the target output volume per unit time. Then, combined with the output calibration volume per revolution, the number of revolutions required per unit time is derived, and thus the target rotor speed is derived. Based on the positive correlation between the target flow rate and the rotor speed, and combined with the aforementioned pump cavity structural parameters and calibration relationship, the calculation example yields a target rotor speed of approximately 1280 revolutions per minute. The output adjustment target structural field is written with the cycle number, target flow rate, target rotor speed, rule number, baseline interval, and weight stratification number.

[0106] The signal generation submodule maps and encodes each flow rate regulation result in the output regulation target, combined with the abnormal cycle number and the preset control interface field definition, to construct a control instruction format that meets the output structure requirements and generate ventricular auxiliary output regulation control instructions.

[0107] The mapping and encoding process first loads the control interface field table, which records the control command header identifier, cycle number field position, target rotor speed field position, verification field position, execution window field position, and structural parameter version number field position. Then, command assembly is performed on each output adjustment target item. The assembly process writes the command header identifier to a fixed byte segment, the cycle number to the cycle number field position, the target rotor speed to the speed field position, the structural parameter version number to the version field position, and the execution window to the time window field position. The execution window is set to 0.10 seconds to 0.30 seconds after the current cycle end timestamp to avoid resource contention with sampling. Verification field generation uses a byte-by-byte XOR check, covering all bytes from the command header identifier to the version field position, and the generated check value is written to the verification field position. If the output adjustment target item carries a verification flag, the command assembly write maintains the speed flag bit and writes the target rotor speed to the current speed value to avoid unfounded changes. In the example, the target rotor speed of 3600 revolutions per minute for cycle number 20 is written into the instruction speed field, the execution time window is written from 10:00:20.300 to 10:00:20.500, and the XOR check value is written into the check field. After all instruction entries are sorted by the start timestamp of the execution time window, the output is a sequence of ventricular auxiliary output regulation control instructions. The sequence length and threshold version number are written at the beginning of the sequence, and the end identifier field is written at the end of the sequence. The end identifier field is fixed as a 1-byte termination code, which facilitates the control interface to parse and send the instructions to the auxiliary pump drive unit in one go.

[0108] Please see Figure 8 A wearable ventricular assist control method for pediatric DCM is implemented based on the aforementioned wearable ventricular assist device for pediatric DCM, and includes the following steps:

[0109] S1: Obtain signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle, calculate the pressure stress amplitude and temperature fluctuation amplitude respectively, and compare them with the average level of the previous few collection values ​​in the current cycle to construct perfusion state response feature items and generate microcirculation perfusion sensing data set.

[0110] S2: Based on the perfusion state response feature items in the microcirculation perfusion sensing data set, extract the data tags that need to be executed for the flow rate response in the current cycle, call the rotor speed and pump chamber parameters recorded in the auxiliary pump drive unit, calculate the output liquid volume of the cycle, and perform matching processing on the cycle corresponding to the sensing data to generate the auxiliary pump response perfusion data set.

[0111] S3: Based on the flow rate information recorded in the auxiliary pump response perfusion data set, combined with the perfusion response feature items corresponding to the sensing data set, the two data sets are mapped and integrated in the time dimension through the period number in the control chip to generate a wearable perfusion-pump speed association set.

[0112] S4: Extract the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, and make a two-way judgment against the preset perfusion state threshold. If both indicators of the current cycle are in the insufficient range at the same time, the cycle is marked as perfusion abnormal and a perfusion response insufficient label is output.

[0113] S5: Identify the cycle status marked by the insufficient perfusion response tag. If it is an abnormal status, use the current cycle flow rate value recorded in the auxiliary pump response perfusion data set, combine it with the adjustment rules to determine the current output target value, generate a pump speed control signal that can be issued, and generate a ventricular auxiliary output adjustment control command.

[0114] 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 wearable ventricular assist device for pediatric DCM, characterized in that, The device includes: The sensing feedback module collects signal changes from the flexible pressure sensor and the thermal temperature sensor that fit the chest area, calculates the pressure stress amplitude and temperature fluctuation amplitude respectively, and compares them with the average value of the previous N cycles to construct perfusion state response characteristic items and generate microcirculation perfusion sensing data set. The flow rate acquisition module extracts the response data tag of the current cycle based on the feature items in the microcirculation perfusion sensing data set, calls the rotor speed information output by the drive unit used to drive the operation of the ventricular auxiliary pump and the pump cavity parameters corresponding to the ventricular auxiliary pump, calculates the output fluid volume, matches the sensing data cycle, and generates the auxiliary pump response perfusion data set. The sensing data period is the data period formed by the sensing feedback module segmenting the sensor signal according to a preset sampling time window or period number. The data alignment module, based on the flow rate information in the auxiliary pump response perfusion data group and the feature items in the sensing data group, performs pairing processing on the flow rate information and the feature items according to the preset period numbering rules to form a wearable perfusion-pump speed association group. The perfusion determination module extracts the pressure stress amplitude and temperature fluctuation amplitude from the wearable perfusion-pump speed correlation group, and makes a bidirectional determination based on the preset perfusion threshold. If both indicators are lower than the threshold at the same time, it is marked as an abnormal perfusion cycle and a perfusion response insufficient label is output. The output control module identifies abnormal cycles based on the insufficient perfusion response label, reads the flow rate value of the corresponding cycle, determines the output target in combination with the adjustment rules, and generates ventricular assist output adjustment control commands. The infusion determination module includes: The index extraction submodule extracts the pressure stress amplitude and temperature fluctuation amplitude recorded in the wearable perfusion-pump speed correlation group, calls the number information and corresponding amplitude value in each cycle, filters the cycle data of the signal record, combines and constructs a dual index data structure, and generates a perfusion fluctuation parameter group. The status judgment submodule calls the pressure stress amplitude and temperature fluctuation amplitude in the perfusion fluctuation parameter group. Based on the set pressure stress threshold and temperature fluctuation threshold, it performs interval judgment on the values ​​of the two indicators in each cycle. When the pressure stress amplitude is less than the corresponding pressure stress threshold and the temperature fluctuation amplitude is less than the corresponding temperature fluctuation threshold, it records the corresponding cycle number and obtains the perfusion insufficient cycle number group. At the end, the status judgment submodule performs a consistency check between the status classification label of the insufficient injection cycle number group and the associated group. If the insufficient injection judgment result is insufficient but the status classification label is normal response, it writes to the conflict flag field and outputs it for the exception flag submodule to write to the end of the data structure. The anomaly marking submodule locates the corresponding cycle data in the wearable perfusion-pump speed association group based on the number content recorded in the insufficient perfusion cycle number group, establishes an anomaly marking field and writes it to the end of each cycle data item, outputs the marked data structure, and generates an insufficient perfusion response tag.

2. The wearable ventricular assist device for pediatric DCM according to claim 1, characterized in that: The microcirculation perfusion sensing data set includes pressure stress amplitude, temperature fluctuation amplitude, and a comparison of the current cycle sensing signal change value with the average level. The auxiliary pump response perfusion data set includes execution cycle data tags, rotor speed, pump chamber parameters, and cycle output volume. The wearable perfusion-pump speed association set includes pressure stress amplitude, temperature fluctuation amplitude, cycle number, and time mapping result. The insufficient perfusion response tag includes cycle abnormality markers, low pressure stress value markers, and low temperature fluctuation value markers. The ventricular auxiliary output regulation control command includes the current cycle flow rate value, regulation rule, and target pump speed output value.

3. A wearable ventricular assist device for pediatric DCM according to claim 1, characterized in that: The perception feedback module includes: The signal acquisition submodule acquires signal change data collected by the flexible pressure sensor and the thermal temperature sensor in the chest area of ​​the wearable ventricular assist device for children with DCM in the current cycle. Based on the two types of sensor signals acquired in the current cycle, it calls the time series and change amplitude parameters of the signal data, filters out signal segments that meet the change trend, and generates the signal change segment value of the current cycle. The amplitude calculation submodule calls the pressure change data and temperature change data in the current cycle signal change segment value, and calls the corresponding signal data recorded in the previous N cycles to calculate the average amplitude value of the previous N cycles. The pressure fluctuation amplitude and temperature fluctuation amplitude of the current cycle are compared with the average amplitude value respectively. Based on the corresponding data sequence recorded in the previous few times of the current cycle, the difference value of the fluctuation amplitude and change trend of the two types of signals is calculated, the core parameters reflecting the amplitude state are extracted, and the amplitude difference value of the current cycle is obtained. The feature construction submodule determines the state classification standard to which the current signal feature belongs based on the changes in pressure and temperature amplitudes in the current cycle amplitude difference value, according to the set perfusion judgment parameters and state reference interval, establishes data feature items corresponding to the perfusion state, and generates microcirculation perfusion sensing data set.

4. A wearable ventricular assist device for pediatric DCM according to claim 3, characterized in that: The flow rate acquisition module includes: The tag extraction submodule extracts data tags with the same timestamp in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data group, calls the time stamp and signal type parameters corresponding to the data tags, filters all tag information belonging to the current cycle range, and generates the current cycle response tag group. The liquid volume calculation submodule calls the corresponding time period range in the current cycle response tag group, combines the rotor speed and pump cavity structure parameters recorded in the auxiliary pump drive unit, and calculates the output liquid volume of the pump cavity within the cycle based on the correlation time marker and structure comparison relationship of the two types of parameters to obtain the cycle liquid output volume. The data matching submodule determines the correspondence between the period of the sensing data and the liquid volume data based on the periodic liquid output and the time tag recorded in the current periodic response tag group. It performs corresponding merging processing according to the periodic mark, establishes periodic matching data items between the auxiliary pump working output and the sensing tag, and generates the auxiliary pump response injection data group.

5. A wearable ventricular assist device for pediatric DCM according to claim 4, characterized in that: The process of extracting data tags with correlation in the current cycle based on the perfusion state response feature items in the microcirculation perfusion sensing data set is specifically defined as follows: the continuously collected perfusion state response feature items are time-aligned according to a fixed sampling time interval; when adjacent perfusion state response feature items are in the same time mark range and have the same signal type parameters, they are determined to have correlation and the corresponding data tags are extracted. The process of calculating the output liquid volume of the pump cavity within a cycle based on the correlation time markers and structural comparison relationships of the two types of parameters is specifically defined as follows: within the time period, the rotor speed recorded in the auxiliary pump drive unit is integrated at equal time intervals to obtain the cumulative rotor rotations within the corresponding time period, and based on the single-rotor output liquid volume or equivalent displacement parameters preset in the pump cavity structural parameters, the cumulative rotor rotations are converted into liquid volume to obtain the cycle liquid output volume. The specific limitation of the merging process based on the periodic marker is that a periodic matching data item between the auxiliary pump working output and the sensing tag is only established when the time tag of the periodic liquid output completely falls within the time period range corresponding to the current periodic response tag group.

6. A wearable ventricular assist device for pediatric DCM according to claim 5, characterized in that: The data alignment module includes: The information extraction submodule, based on the flow rate information recorded in the auxiliary pump response injection data group, calls the time tag and data identifier number parameters to extract the complete flow rate content recorded within the current cycle range, and organizes it into a structural sequence according to the cycle order to generate a flow rate cycle tag group. The cycle pairing submodule calls the corresponding infusion response feature items in the flow rate cycle tag group and the sensing data group, performs tag comparison processing according to the cycle number in the control chip, filters data combinations with the same number, establishes data comparison items according to the number order, and obtains the cycle pairing mapping table. The time mapping submodule extracts the time period information corresponding to the number according to the period numbering relationship marked in the period pairing mapping table, and synchronously maps the flow rate data with the perfusion feature items according to the time dimension, integrates them into a unified structure output format, and generates a wearable perfusion-pump rate association group.

7. A wearable ventricular assist device for pediatric DCM according to claim 1, characterized in that: The output control module includes: The status recognition submodule identifies the periodic status marked in the insufficient perfusion response label, calls the period number and status mark field, filters the periodic sequences with abnormal status, establishes an abnormal period index table, and generates abnormal period number groups. The target calculation submodule calls the flow rate value in the auxiliary pump response perfusion data group corresponding to the abnormal cycle number group, matches the current cycle number with the output reference interval preset in the adjustment rule to limit the output flow rate range of the ventricular auxiliary pump, calculates the control target value under the abnormal cycle, and obtains the output adjustment target. The signal generation submodule maps and encodes each flow rate regulation result in the output regulation target, combined with the abnormal cycle number and the preset control interface field definition, to construct a control instruction format that meets the output structure requirements and generate ventricular auxiliary output regulation control instructions.

8. A wearable ventricular assist device for pediatric DCM according to claim 7, characterized in that: The process of identifying the cycle status marked in the insufficient perfusion response label is specifically limited to comparing the cycle number recorded in the insufficient perfusion response label with the status mark field cycle by cycle, and only when the status mark field is a preset abnormal identifier value, the corresponding cycle number is included in the abnormal cycle number group. The process of matching the current cycle number with the output reference interval in the adjustment rule is specifically defined as follows: the flow velocity value corresponding to the current cycle number is compared with the upper and lower limit parameters of the output reference interval for interval judgment; when the flow velocity value is outside the output reference interval, the corresponding adjustment amplitude parameter is selected according to the deviation direction, and the output adjustment target is calculated. The process of constructing a control instruction format that meets the output structure requirements is specifically defined as follows: according to the preset field order in the control interface field, the abnormal cycle number, the flow rate regulation result, and the corresponding verification identifier are sequentially written into a data frame of fixed length to form the ventricular auxiliary output regulation control instruction.

Images