An injection device intelligent linkage control method, system, device and storage medium based on blood glucose monitoring data

Abstract

The application discloses an injection device intelligent linkage control method and system based on blood glucose monitoring data, equipment and a storage medium, relates to the technical field of data intelligent control, and comprises the following steps: acquiring blood glucose monitoring data and a current state, performing time alignment, abnormal jump identification and data confidence evaluation on the blood glucose monitoring data, generating an effective blood glucose sequence and data confidence; training a risk identification model, extracting multi-time window blood glucose evolution features through the effective blood glucose sequence; inputting the multi-time window blood glucose evolution features, the data confidence and the current state into the risk identification model, and outputting a hierarchical linkage control strategy; and executing linkage control through the hierarchical linkage control strategy. The method disclosed by the application achieves better effects in terms of data processing reliability, risk level output continuity and control process continuity.

Description

Technical Field

[0001] This invention relates to the field of data intelligent control technology, specifically to an intelligent linkage control method, system, device, and storage medium for an injection device based on blood glucose monitoring data. Background Technology

[0002] According to a 2024 report in *The Lancet Diabetes & Endocrinology*, the global prevalence of diabetes among adults aged 20-79 is 11.11%, with approximately 589 million patients. This number is projected to rise to 853 million by 2050. China has the largest number of diabetes patients globally, with approximately 148 million patients in 2024, accounting for nearly a quarter of the global total. Continuous monitoring and timely, stable control of blood glucose levels are crucial for preventing the progression of diabetes and its complications.

[0003] In the current technology, blood glucose monitoring devices and injection devices are mostly still operating in a relatively independent manner.

[0004] Blood glucose monitoring devices typically only output blood glucose values ​​and alarm information, while mobile terminals are mainly used for data display and historical data retrieval. Injection devices, on the other hand, independently execute injection control according to predetermined injection parameters. When abnormal changes occur in blood glucose monitoring data, especially a rapid drop in blood glucose or entry into the hypoglycemic risk zone, current solutions usually still rely on users or healthcare professionals to manually adjust the injection device's operation after receiving an alarm.

[0005] This type of injection treatment has a significant time lag, especially when the user is resting at night, the terminal is muted, there is communication interference, or the user fails to check the alarm information in time. There is no automatic connection mechanism between blood glucose monitoring data and injection device control, which results in blood glucose monitoring data failing to effectively participate in the subsequent injection control process.

[0006] Furthermore, existing control schemes based on blood glucose monitoring data typically have the following problems:

[0007] Firstly, the judgment is made directly based on a single sample value or a simple threshold, lacking the processing steps of time alignment, abnormal jump identification, and data confidence assessment of blood glucose monitoring data. This results in noisy data, distorted data, or communication anomaly data potentially entering the control link directly.

[0008] Secondly, existing solutions focus more on single-point blood glucose values ​​at the current moment, lacking multi-timescale analysis of continuous effective blood glucose sequences, and thus cannot fully reflect the rate of blood glucose change, stage fluctuations, and continuous trends.

[0009] Third, existing solutions often use fixed rules or single-layer judgment logic when outputting control actions, lacking a linkage analysis mechanism based on historical control records, current state, and continuous sampling cycle changes, making it difficult to form a hierarchical linkage control strategy that can dynamically adjust with the evolution of blood glucose status.

[0010] Fourth, after the injection device performs the control action, the existing solutions generally lack a mechanism for continuous verification and feedback update of the execution results, which makes it difficult for the output of the current control cycle to stably act on the input of the next control cycle, and the system as a whole lacks a closed-loop logic for continuous operation. Summary of the Invention

[0011] In view of the above-mentioned problems, the present invention is proposed.

[0012] Therefore, the technical problem solved by the present invention is that existing injection device control methods based on blood glucose monitoring data have problems such as insufficient processing of blood glucose monitoring data, low accuracy of risk identification, and poor linkage between blood glucose monitoring data and injection device control, as well as the problem of how to generate a hierarchical linkage control strategy based on blood glucose monitoring data and drive the injection device to perform continuous linkage control.

[0013] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an intelligent linkage control method for an injection device based on blood glucose monitoring data, comprising acquiring blood glucose monitoring data and current status, performing time alignment, abnormal jump identification and data confidence assessment on blood glucose monitoring data, and generating a valid blood glucose sequence and data confidence.

[0014] The risk identification model is trained by extracting multi-time-window blood glucose evolution features from effective blood glucose sequences. The multi-time-window blood glucose evolution features, data confidence, and current state are input into the risk identification model, which outputs a hierarchical linkage control strategy. The linkage control is then executed through the hierarchical linkage control strategy.

[0015] The output hierarchical linkage control strategy includes determining the hierarchical linkage control strategy by combining the risk level and control records corresponding to the previous control cycle.

[0016] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the generation of effective blood glucose sequences and data confidence includes the mobile terminal receiving raw blood glucose monitoring data sent by the blood glucose monitoring device.

[0017] Read the tagging information corresponding to each raw blood glucose monitoring data.

[0018] The tagging information includes sampling timestamps, monitoring status tags, and communication status tags.

[0019] Time alignment is performed on each sampling timestamp, and the time-aligned raw blood glucose monitoring data are grouped into a queue of data to be processed according to the sampling order.

[0020] Perform abnormal transition identification on adjacent sampling points in the data queue to be processed.

[0021] Abnormal jump identification includes calculating the blood glucose difference, time interval, and direction of difference change between adjacent sampling points, and comparing the calculation results with jump criteria.

[0022] When the direction of the difference between adjacent sampling points is inconsistent with the direction of change formed by multiple previous sampling periods, the corresponding sampling point is marked as an abnormal candidate point.

[0023] By combining monitoring status markers, communication status markers, and anomaly candidate point markers, a data confidence assessment is performed on each sampling point to obtain the data confidence value corresponding to each sampling point.

[0024] The queue of data to be processed is filtered and reorganized based on the data confidence value.

[0025] Sampling points that meet the preset reliability conditions are arranged in chronological order to form an effective blood glucose sequence.

[0026] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the training risk identification model includes acquiring historical blood glucose monitoring data, historical current state data, and control records corresponding to the historical blood glucose monitoring data.

[0027] Historical blood glucose monitoring data, historical current status data, and control record execution time are aligned to form a historical sample sequence.

[0028] The historical sample sequence is sliced ​​according to a preset time length to obtain multiple training samples.

[0029] For each training sample, extract historical blood glucose monitoring data, historical current status data, and control records within the corresponding time range.

[0030] Based on the blood glucose changes and control records of each training sample, the risk level of each training sample was labeled.

[0031] Each training sample with completed risk level labeling is input into the classification learning model for parameter training.

[0032] The iteration rounds are set to 37. Validation samples are read according to the iteration rounds. The current model output is compared with the risk level label corresponding to the validation sample. The model parameters are updated based on the comparison results to obtain the trained risk identification model.

[0033] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the extraction of multi-time-window blood glucose evolution features includes calling an effective blood glucose sequence and reading the data confidence level and current state corresponding to the effective blood glucose sequence.

[0034] Short time windows, medium time windows, and long time windows are constructed around the current sampling time.

[0035] The process involves constructing short, medium, and long time windows around the current sampling time, using the current sampling time as the end time of each time window, and selecting continuous sampling points by backtracking from the effective blood glucose sequence.

[0036] Continuous sampling points include the most recent 5 sampling points for short time windows, the most recent 15 sampling points for medium time windows, and the most recent 30 sampling points for long time windows.

[0037] Short time windows, medium time windows, and long time windows together characterize the blood glucose change process at different time scales before the current sampling time.

[0038] Segmented calculations are performed on the effective blood glucose sequence within short, medium, and long time windows. Data confidence and current status are appended to the calculation results to form the original feature sequence.

[0039] Based on the chronological order, sequential encoding is performed on the original feature sequences, and the original feature sequences after sequential encoding are determined as multi-time-window blood glucose evolution features.

[0040] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the output risk level includes calling the trained risk identification model.

[0041] The blood glucose evolution characteristics of multiple time windows, data confidence, and current status are input into the risk identification model.

[0042] The risk identification model performs hierarchical classification operations on the blood glucose evolution characteristics, data confidence, and current status across multiple time windows to obtain candidate risk levels and their corresponding level values.

[0043] Sort the candidate risk levels according to their corresponding level values, and select the current output risk level corresponding to the sorting result.

[0044] Read the risk level and control record corresponding to the previous control cycle.

[0045] Perform a consistency comparison between the current output risk level and the risk level corresponding to the previous control cycle.

[0046] The current output risk level is consistent with the risk level corresponding to the previous control cycle, so the current output risk level is determined to be a risk level.

[0047] The current output risk level is inconsistent with the risk level corresponding to the previous control cycle. By combining the control records, the direction of change of blood glucose evolution characteristics in multiple time windows is continuously judged.

[0048] When the direction of change of the multi-time window blood glucose evolution characteristics is consistent with the change process corresponding to the current output risk level, the current output risk level is determined as the risk level.

[0049] When the direction of change of blood glucose evolution characteristics in multiple time windows is inconsistent with the change process corresponding to the current output risk level, the risk level corresponding to the previous control cycle is maintained as the risk level.

[0050] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the output hierarchical linkage control strategy includes reading the control record corresponding to the risk level, the current state, and the previous control cycle.

[0051] Based on the risk level, query the action content corresponding to the risk level, and match the action content with the current state.

[0052] The current state corresponds to the action content, and a hierarchical linkage control strategy is generated based on the action content.

[0053] The current state does not correspond to the action content. Read the control record corresponding to the previous control cycle.

[0054] Adjust the action content based on the control record corresponding to the previous control cycle.

[0055] The adjusted action content is defined as a hierarchical linkage control strategy.

[0056] Write the hierarchical linkage control strategy into the control record corresponding to the current control cycle.

[0057] As a preferred embodiment of the intelligent linkage control method for injection devices based on blood glucose monitoring data described in this invention, the execution of linkage control includes the mobile terminal sending a hierarchical linkage control strategy to the injection device.

[0058] After receiving the hierarchical linkage control strategy, the injection device reads the target action, target parameters, action start time, action duration, re-evaluation content, and rollback content from the hierarchical linkage control strategy.

[0059] The corresponding linkage control is executed according to the hierarchical linkage control strategy, and the mobile terminal receives the execution result returned by the injection device.

[0060] The execution results include the actions performed, the current status, and the current parameters.

[0061] The execution results are compared for consistency with the hierarchical linkage control strategy.

[0062] When the executed action is consistent with the target action and the current parameters are consistent with the target parameters.

[0063] Update the current state in the execution result to the current state of the next control cycle, and write the hierarchical linkage control strategy and execution result into the control record.

[0064] If the executed action is inconsistent with the target action, the hierarchical linkage control strategy is resent and the resentment result is written to the control record.

[0065] If the current parameters are inconsistent with the target parameters, resend the hierarchical linkage control strategy and write the resend result into the control record.

[0066] At the start of the next control cycle, the updated current state and control records are invoked as inputs for outputting the risk level and the hierarchical linkage control strategy.

[0067] Another objective of this invention is to provide an intelligent linkage control system for injection devices based on blood glucose monitoring data. This system can effectively process blood glucose monitoring data and generate a hierarchical linkage control strategy based on the blood glucose evolution characteristics of multiple time windows. This solves the problems of insufficient data reliability, discontinuous risk level output, and poor linkage control in current injection device control technologies based on blood glucose monitoring data.

[0068] As a preferred embodiment of the intelligent linkage control system for injection devices based on blood glucose monitoring data described in this invention, it includes a blood glucose data processing module, a risk identification module, and a linkage control module.

[0069] The blood glucose data processing module is used to acquire blood glucose monitoring data and current status, perform time alignment, abnormal jump identification and data confidence assessment on the blood glucose monitoring data, and generate valid blood glucose sequences and data confidence scores.

[0070] The risk identification module is used to train the risk identification model and extracts multi-time-window blood glucose evolution features through effective blood glucose sequences.

[0071] The linkage control module is used to input the multi-time window blood glucose evolution characteristics, data confidence and current state into the risk identification model, output a hierarchical linkage control strategy, and execute linkage control through the hierarchical linkage control strategy.

[0072] Another objective of this invention is to provide an intelligent linkage control device for an injection device based on blood glucose monitoring data, comprising a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the intelligent linkage control method for the injection device based on blood glucose monitoring data.

[0073] Another object of the present invention is to provide an intelligent linkage control storage medium for an injection device based on blood glucose monitoring data, wherein a computer program is stored thereon, and when the computer program is executed by a processor, the steps of the intelligent linkage control method for the injection device based on blood glucose monitoring data are implemented.

[0074] The beneficial effects of this invention are as follows: The intelligent linkage control method for injection devices based on blood glucose monitoring data provided by this invention avoids abnormal sampling points from directly entering the subsequent control process by performing time alignment, abnormal jump identification, and data confidence assessment on blood glucose monitoring data; by extracting multi-time-window blood glucose evolution features from effective blood glucose sequences and inputting them into a risk identification model, a hierarchical linkage control strategy matching the current state is formed; by executing linkage control according to the hierarchical linkage control strategy and applying the execution results to the next control cycle, this invention achieves better results in terms of data processing reliability, continuity of risk level output, and control process connectivity. Attached Figure Description

[0075] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the 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.

[0076] Figure 1 The above is an overall flowchart of an intelligent linkage control method for an injection device based on blood glucose monitoring data, provided in Embodiment 1 of the present invention. Detailed Implementation

[0077] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0078] Example 1, referring to Figure 1 As an embodiment of the present invention, an intelligent linkage control method for an injection device based on blood glucose monitoring data is provided, comprising:

[0079] S1: Acquire blood glucose monitoring data and current status, perform time alignment, abnormal jump identification and data confidence assessment on blood glucose monitoring data, and generate valid blood glucose sequences and data confidence.

[0080] The mobile terminal receives raw blood glucose monitoring data sent by the blood glucose monitoring device.

[0081] Read the tagging information corresponding to each raw blood glucose monitoring data.

[0082] The tagging information includes sampling timestamps, monitoring status tags, and communication status tags.

[0083] Perform time alignment processing on each sampling timestamp.

[0084] The time-aligned raw blood glucose monitoring data are arranged into a queue of data to be processed according to the order of sampling.

[0085] Perform abnormal transition identification on adjacent sampling points in the data queue to be processed.

[0086] Abnormal jump identification includes calculating the blood glucose difference, time interval, and direction of difference change between adjacent sampling points, and comparing the calculation results with jump criteria.

[0087] When the direction of the difference between adjacent sampling points is inconsistent with the direction of change formed by multiple previous sampling periods, the corresponding sampling point is marked as an abnormal candidate point.

[0088] By combining monitoring status markers, communication status markers, and anomaly candidate point markers, a data confidence assessment is performed on each sampling point to obtain the data confidence value corresponding to each sampling point.

[0089] A preferred approach for performing data confidence assessment at each sampling point is as follows:

[0090]

[0091] in, This indicates the confidence level of the data corresponding to the current sampling point. This represents the time continuity evaluation value corresponding to the current sampling point. This represents the transition consistency evaluation value corresponding to the current sampling point. This represents the availability evaluation value corresponding to the current sampling point.

[0092] In this preferred scheme, the time continuity evaluation value is first determined based on the degree of continuity of the sampling timestamps after time alignment. Then, the jump consistency evaluation value is determined based on the degree of consistency between the direction of change of the difference between adjacent sampling points and the direction of change of the previous multiple sampling periods. Finally, the status availability evaluation value is determined based on both the monitoring status flags and the communication status flags. The average of the three values ​​is used to obtain the data confidence level of the current sampling point. When the confidence level of the data at the current sampling point When the retention conditions are met, the corresponding blood glucose monitoring data will be retained in the valid blood glucose sequence; when the confidence level of the data at the current sampling point is... If the retention criteria are not met, the corresponding blood glucose monitoring data will be removed from the valid blood glucose sequence.

[0093] The queue of data to be processed is filtered and reorganized based on the data confidence value.

[0094] Sampling points that meet the preset reliability conditions are arranged in chronological order to form an effective blood glucose sequence.

[0095] S2: Train the risk identification model and extract multi-time window blood glucose evolution features through effective blood glucose sequences.

[0096] Acquire historical blood glucose monitoring data, historical current status data, and control records corresponding to the historical blood glucose monitoring data.

[0097] Historical blood glucose monitoring data, historical current status data, and control record execution time are aligned to form a historical sample sequence.

[0098] The historical sample sequence is sliced ​​according to a preset time length to obtain multiple training samples.

[0099] For each training sample, extract historical blood glucose monitoring data, historical current status data, and control records within the corresponding time range.

[0100] Based on the blood glucose changes and control records of each training sample, the risk level of each training sample was labeled.

[0101] Each training sample with completed risk level labeling is input into the classification learning model for parameter training.

[0102] The iteration rounds are set to 37. Validation samples are read according to the iteration rounds. The current model output is compared with the risk level label corresponding to the validation sample. The model parameters are updated based on the comparison results to obtain the trained risk identification model.

[0103] Furthermore, the system calls upon the valid blood glucose sequence and reads the data confidence level and current status corresponding to the valid blood glucose sequence.

[0104] Short time windows, medium time windows, and long time windows are constructed around the current sampling time.

[0105] The process involves constructing short, medium, and long time windows around the current sampling time, using the current sampling time as the end time of each time window, and selecting continuous sampling points by backtracking from the effective blood glucose sequence.

[0106] Continuous sampling points include the most recent 5 sampling points for short time windows, the most recent 15 sampling points for medium time windows, and the most recent 30 sampling points for long time windows.

[0107] Short time windows, medium time windows, and long time windows together characterize the blood glucose change process at different time scales before the current sampling time.

[0108] Segmented calculations are performed on the effective blood glucose sequence within short, medium, and long time windows.

[0109] A preferred approach for performing segmented calculations on effective blood glucose sequences within short, medium, and long time windows is as follows:

[0110]

[0111] in, This indicates the multi-time-window blood glucose evolution characteristics corresponding to the current sampling time. This indicates the amount of blood glucose change within a short time window. This represents the average change in blood glucose levels within a given time window. This indicates the trend of blood glucose changes over a long period of time. Indicates the current state.

[0112] The data confidence level and the current state are appended to the calculation results to form the original feature sequence.

[0113] Based on the chronological order, sequential encoding is performed on the original feature sequence.

[0114] The original feature sequence after sequential encoding was determined as the multi-time-window blood glucose evolution feature.

[0115] S3: Input the multi-time window blood glucose evolution characteristics, data confidence and current state into the risk identification model, and output a hierarchical linkage control strategy.

[0116] The trained risk identification model is invoked, and the blood glucose evolution characteristics, data confidence, and current state are input into the risk identification model.

[0117] The risk identification model performs hierarchical classification operations on the blood glucose evolution characteristics, data confidence, and current status across multiple time windows to obtain candidate risk levels and their corresponding level values.

[0118] A preferred method for calculating risk levels is:

[0119]

[0120]

[0121] in, This represents the overall risk value corresponding to the current sampling time. This indicates the risk level corresponding to the current sampling time. Indicates the current state.

[0122] In this preferred scheme, the blood glucose change characteristics of the short time window, the medium time window, and the long time window, along with the data confidence level and the current state, are first substituted into the calculation to calculate the comprehensive risk value corresponding to the current sampling time.

[0123] Then, substitute the comprehensive risk value into the formula to determine the risk level corresponding to the current sampling time.

[0124] Among them, risk level Indicates low risk level, risk level Indicates the level of concern and the level of risk. Indicates restriction level, risk level Indicates the level of protection.

[0125] Sort the candidate risk levels according to their corresponding level values, and select the current output risk level corresponding to the sorting result.

[0126] Read the risk level and control record corresponding to the previous control cycle.

[0127] Perform a consistency comparison between the current output risk level and the risk level corresponding to the previous control cycle.

[0128] The current output risk level is consistent with the risk level corresponding to the previous control cycle, so the current output risk level is determined to be a risk level.

[0129] The current output risk level is inconsistent with the risk level corresponding to the previous control cycle. By combining the control records, the direction of change of blood glucose evolution characteristics in multiple time windows is continuously judged.

[0130] When the direction of change of the multi-time window blood glucose evolution characteristics is consistent with the change process corresponding to the current output risk level, the current output risk level is determined as the risk level.

[0131] When the direction of change of blood glucose evolution characteristics in multiple time windows is inconsistent with the change process corresponding to the current output risk level, the risk level corresponding to the previous control cycle is maintained as the risk level.

[0132] Furthermore, the risk level, current status, and control records corresponding to the previous control cycle are read.

[0133] Search for the action content corresponding to the risk level.

[0134] The action content includes the target action, target parameters, action start time, action duration, re-evaluation content, and rollback content.

[0135] Match the action content with the current state.

[0136] The current state corresponds to the action content, and a hierarchical linkage control strategy is generated based on the action content.

[0137] A preferred approach for generating a hierarchical linkage control strategy is as follows:

[0138]

[0139]

[0140]

[0141] in, This indicates the target action identifier corresponding to the current control cycle. This represents the target parameter corresponding to the current control cycle. This represents the target parameter corresponding to the previous control cycle. This indicates the duration of the action corresponding to the current control cycle. Indicates the duration of the base. This indicates the duration increment corresponding to the hierarchical adjustment. This indicates the parameter reduction factor at the level of interest. This represents the parameter reduction factor under the restriction level.

[0142] The current state does not correspond to the action content. Read the control record corresponding to the previous control cycle.

[0143] Adjust the action content based on the control record corresponding to the previous control cycle.

[0144] The adjusted action content is defined as a hierarchical linkage control strategy.

[0145] Write the hierarchical linkage control strategy into the control record corresponding to the current control cycle.

[0146] S4: Execute linkage control through a hierarchical linkage control strategy.

[0147] The mobile terminal sends the hierarchical linkage control strategy to the injection device.

[0148] After receiving the hierarchical linkage control strategy, the injection device reads the target action, target parameters, action start time, action duration, re-evaluation content, and rollback content from the hierarchical linkage control strategy.

[0149] Execute the corresponding linkage control according to the hierarchical linkage control strategy.

[0150] The mobile terminal receives the execution result returned by the injection device.

[0151] The execution results include the actions performed, the current status, and the current parameters.

[0152] The execution results are compared for consistency with the hierarchical linkage control strategy.

[0153] When the executed action is consistent with the target action and the current parameters are consistent with the target parameters.

[0154] Update the current state in the execution result to the current state of the next control cycle, and write the hierarchical linkage control strategy and execution result into the control record.

[0155] If the executed action is inconsistent with the target action, the hierarchical linkage control strategy is resent and the resentment result is written to the control record.

[0156] If the current parameters are inconsistent with the target parameters, resend the hierarchical linkage control strategy and write the resend result into the control record.

[0157] At the start of the next control cycle, the updated current state and control records are invoked as inputs for outputting the risk level and the hierarchical linkage control strategy.

[0158] Example 2 is an embodiment of the present invention, which provides an intelligent linkage control system for an injection device based on blood glucose monitoring data, including a blood glucose data processing module, a risk identification module, and a linkage control module.

[0159] The blood glucose data processing module is used to acquire blood glucose monitoring data and current status, perform time alignment, abnormal jump identification and data confidence assessment on the blood glucose monitoring data, and generate valid blood glucose sequences and data confidence.

[0160] The risk identification module is used to train the risk identification model and extracts multi-time-window blood glucose evolution features through effective blood glucose sequences.

[0161] The linkage control module is used to input the blood glucose evolution characteristics, data confidence and current status of multiple time windows into the risk identification model, output a hierarchical linkage control strategy, and execute linkage control through the hierarchical linkage control strategy.

[0162] This embodiment also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the intelligent linkage control method for injection devices based on blood glucose monitoring data as proposed in the above embodiment.

[0163] This embodiment also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the intelligent linkage control method for injection devices based on blood glucose monitoring data as proposed in the above embodiment.

[0164] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0165] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0166] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0167] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0168] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent linkage control of an injection device based on blood glucose monitoring data, characterized in that, include: Acquire blood glucose monitoring data and current status, perform time alignment, abnormal jump identification, and data confidence assessment on the blood glucose monitoring data, and generate valid blood glucose sequences and data confidence scores; Train a risk identification model and extract multi-time-window blood glucose evolution features using effective blood glucose sequences; The risk identification model is input into the multi-time window blood glucose evolution characteristics, data confidence level and current status, and outputs a hierarchical linkage control strategy. Interlocking control is executed through a hierarchical linkage control strategy; wherein... The output hierarchical linkage control strategy includes determining the hierarchical linkage control strategy by combining the risk level and control records corresponding to the previous control cycle.

2. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claim 1, characterized in that: The generation of effective blood glucose sequences and data confidence levels include, The mobile terminal receives raw blood glucose monitoring data sent by the blood glucose monitoring device; Read the tagging information corresponding to each raw blood glucose monitoring data; The tagging information includes sampling timestamps, monitoring status tags, and communication status tags; Perform time alignment processing on each sampling timestamp; The time-aligned raw blood glucose monitoring data are arranged into a data queue to be processed according to the sampling order. Perform abnormal transition identification on adjacent sampling points in the data queue to be processed; Abnormal jump identification includes calculating the blood glucose difference, time interval, and direction of difference change between adjacent sampling points, and comparing the calculation results with jump criteria; When the direction of the difference between adjacent sampling points is inconsistent with the direction of change formed by multiple previous sampling periods, the corresponding sampling point is marked as an abnormal candidate point. By combining monitoring status markers, communication status markers, and anomaly candidate point markers, a data confidence assessment is performed on each sampling point to obtain a data confidence value corresponding to each sampling point; Filtering and reorganizing the queue of data to be processed based on the data confidence value; Sampling points that meet the preset reliability conditions are arranged in chronological order to form an effective blood glucose sequence.

3. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claim 1 or 2, characterized in that: The trained risk identification model includes, Acquire historical blood glucose monitoring data, historical current status data, and control records corresponding to the historical blood glucose monitoring data; Historical blood glucose monitoring data, historical current status data, and control record execution time are aligned to form a historical sample sequence; The historical sample sequence is sliced ​​according to a preset time length to obtain multiple training samples; For each training sample, extract historical blood glucose monitoring data, historical current status data, and control records within the corresponding time range; Based on the blood glucose changes and control records of each training sample, the risk level of each training sample was labeled. Each training sample with completed risk level labeling is input into the classification learning model for parameter training; The iteration rounds are set to 37. Validation samples are read according to the iteration rounds. The current model output is compared with the risk level label corresponding to the validation sample. The model parameters are updated based on the comparison results to obtain the trained risk identification model.

4. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claim 3, characterized in that: The extracted multi-time-window blood glucose evolution features include: Call the valid blood glucose sequence and read the data confidence and current status corresponding to the valid blood glucose sequence; Short-time windows, medium-time windows, and long-time windows are constructed around the current sampling time. The process involves constructing short, medium, and long time windows around the current sampling time, using the current sampling time as the end time of each time window, and selecting continuous sampling points by backtracking from the effective blood glucose sequence. Continuous sampling points include the most recent 5 sampling points for short time windows, the most recent 15 sampling points for medium time windows, and the most recent 30 sampling points for long time windows; Short time windows, medium time windows, and long time windows together characterize the blood glucose change process at different time scales before the current sampling time; Segmented calculations are performed on the effective blood glucose sequence within short, medium, and long time windows; The data confidence level and current state are appended to the calculation results to form the original feature sequence; Based on the chronological order, sequential encoding is performed on the original feature sequence; The original feature sequence after sequential encoding was determined as the multi-time-window blood glucose evolution feature.

5. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claim 1, 2, or 4, characterized in that: Output risk levels include, Invoke the trained risk identification model; Input the multi-time-window blood glucose evolution characteristics, data confidence level, and current state into the risk identification model; The risk identification model is used to perform hierarchical classification operations on the blood glucose evolution characteristics, data confidence and current status in multiple time windows to obtain candidate risk levels and the corresponding level values ​​of each candidate risk level. Sort the candidate risk levels according to their corresponding level values, and select the current output risk level corresponding to the sorting result. Read the risk level and control record corresponding to the previous control cycle; Compare the current output risk level with the risk level corresponding to the previous control cycle; The current output risk level is consistent with the risk level corresponding to the previous control cycle, so the current output risk level is determined to be a risk level. The current output risk level is inconsistent with the risk level corresponding to the previous control cycle. The direction of change of blood glucose evolution characteristics in multiple time windows is continuously judged by combining the control records. When the direction of change of the multi-time window blood glucose evolution characteristics is consistent with the change process corresponding to the current output risk level, the current output risk level is determined as the risk level. When the direction of change of blood glucose evolution characteristics in multiple time windows is inconsistent with the change process corresponding to the current output risk level, the risk level corresponding to the previous control cycle is maintained as the risk level.

6. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claim 5, characterized in that: The output hierarchical linkage control strategy includes, Read the risk level, current status, and control records corresponding to the previous control cycle; Search for the action content corresponding to the risk level; Match the action content with the current state; The current state corresponds to the action content, and a layered linkage control strategy is generated based on the action content; The current state does not correspond to the action content. Read the control record corresponding to the previous control cycle. Adjust the action content based on the control record corresponding to the previous control cycle; The adjusted action content is defined as a hierarchical linkage control strategy; Write the hierarchical linkage control strategy into the control record corresponding to the current control cycle.

7. The intelligent linkage control method for injection devices based on blood glucose monitoring data as described in claims 1, 2, 4 or 6, characterized in that: The execution linkage control includes, The mobile terminal sends the hierarchical linkage control strategy to the injection device; After receiving the hierarchical linkage control strategy, the injection device reads the target action, target parameters, action start time, action duration, re-evaluation content, and rollback content from the hierarchical linkage control strategy. Execute the corresponding linkage control according to the hierarchical linkage control strategy; The mobile terminal receives the execution result returned by the injection device; The execution result includes the actions performed, the current status, and the current parameters; Compare the execution results with the hierarchical linkage control strategy for consistency. When the executed action is consistent with the target action, and the current parameters are consistent with the target parameters; Update the current state in the execution result to the current state of the next control cycle, and write the hierarchical linkage control strategy and execution result into the control record; If the executed action is inconsistent with the target action, resend the hierarchical linkage control strategy and write the resend result into the control record; If the current parameters are inconsistent with the target parameters, resend the hierarchical linkage control strategy and write the resend result into the control record. At the start of the next control cycle, the updated current state and control records are invoked as inputs for outputting the risk level and the hierarchical linkage control strategy.

8. An intelligent linkage control system for an injection device based on blood glucose monitoring data, employing the intelligent linkage control method for an injection device based on blood glucose monitoring data as described in any one of claims 1 to 7, characterized in that: It includes a blood glucose data processing module, a risk identification module, and a linkage control module; The blood glucose data processing module is used to acquire blood glucose monitoring data and current status, perform time alignment, abnormal jump identification and data confidence assessment on blood glucose monitoring data, and generate valid blood glucose sequences and data confidence. The risk identification module is used to train the risk identification model and extract multi-time-window blood glucose evolution features through effective blood glucose sequences. The linkage control module is used to input the multi-time window blood glucose evolution characteristics, data confidence and current state into the risk identification model, output a hierarchical linkage control strategy, and execute linkage control through the hierarchical linkage control strategy.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent linkage control method for injection devices based on blood glucose monitoring data as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent linkage control method for injection devices based on blood glucose monitoring data as described in any one of claims 1 to 7.

Images