Artificial pancreas system-oriented security resilience and physiological perception intelligent insulin pump control method and related equipment

Abstract

This invention discloses a safety-resilient and physiologically-aware intelligent insulin pump control method and related equipment for artificial pancreas systems. The method includes: constructing a state vector based on blood glucose concentration, distal insulin effect, and effective insulin quantity; constructing an initial artificial pancreas system model with interval type II fuzzy logic based on glucose dynamics and the state vector; designing an improved adaptive event triggering mechanism to generate an initial pump update command with bounded delay; constructing a target artificial pancreas system model based on the initial pump update command, attack signal, and the initial artificial pancreas system model; constructing a Lyapunov-Krasovsky functional based on the target artificial pancreas system model to obtain controller parameters; and controlling the insulin pump based on the target artificial pancreas system model and the controller parameters. This invention can improve the safety of blood glucose control and the robustness of system operation, and can be widely applied in the field of intelligent control technology for medical devices.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for medical devices, and in particular to a method and related equipment for controlling an intelligent insulin pump with safety, resilience, and physiological perception for artificial pancreas systems. Background Technology

[0002] Currently, automated blood glucose management integrates an artificial pancreas system (APS) that combines a continuous glucose monitoring sensor, an insulin pump, and a closed-loop control algorithm. In a closed-loop APS, the controller automatically calculates the insulin infusion rate based on real-time blood glucose information and sends it to the insulin pump, achieving closed-loop regulation of "measurement-decision-execution". The control algorithm needs to maintain blood glucose within the target safe range (e.g., 70 to 180 mg / dL) for as long as possible while ensuring clinical safety (avoiding severe hypoglycemia and prolonged hyperglycemia), and also considering patient comfort and medication economy. Existing research often uses mathematical models such as the Bergman minimum model to describe the blood glucose-insulin regulation process. These models are simple in structure, have clear parameter meanings, and are easy to analyze and design controllers. However, real blood glucose regulation systems have significant nonlinearity, strong uncertainty, and individual differences. Relying solely on single-point linearization or fixed-parameter models is difficult to accurately characterize actual dynamic characteristics, and the adaptability of closed-loop strategies to different patients and working conditions is limited. Current technologies introduce T-S fuzzy models to approximate the blood glucose regulation process in a piecewise linear manner, which to some extent balances model interpretability and computational complexity. However, in traditional type-I fuzzy models, the membership function is usually fixed in advance, making it difficult to explicitly describe the uncertainties caused by slow parameter drift, measurement noise, and modeling errors. This results in insufficient robustness for cross-individual and cross-scenario applications. Faced with multi-source uncertainties such as dietary interference, changes in insulin sensitivity, and sensor noise, existing APS modeling and control methods still fall short in terms of physiological consistency and safety margin.

[0003] With the development of wearable devices and telemedicine, closed-loop APS (Automatic Blood Pressure Monitoring System) increasingly uses wired or wireless networks to achieve information exchange between the controller and the insulin pump. If a fixed-period sampling and control update method is continuously used, on the one hand, it requires high communication frequency and hardware load, increasing system cost and energy consumption; on the other hand, frequent micro-dose adjustments during relatively stable blood glucose periods have limited clinical benefits and may introduce unnecessary blood glucose micro-oscillations. Existing methods mostly employ fixed thresholds or simple adaptive threshold strategies based on error amplitude. While these can suppress redundant updates to some extent, they often lack characterization of the physiological timescale of blood glucose, making it difficult to simultaneously consider the rapid postprandial dynamics and the slow steady-state evolution, easily leading to situations of "insufficient triggering during the postprandial disturbance period" or "excessive triggering during the steady-state period."

[0004] Furthermore, APS operating in open network environments also face security risks from cyberattacks. Typical attack forms include denial-of-service (DoS) attacks, which block channels with malicious traffic, causing control command packet loss or severe delays, as well as deceptive attacks such as replay attacks, fake data injection, and data tampering. Attackers can perturb critical data without significantly altering the communication link, thereby misleading the insulin pump to administer incorrect doses. Summary of the Invention

[0005] In view of this, the main objective of the embodiments of the present invention is to provide a safe, resilient, and physiologically sensitive intelligent insulin pump control method and related equipment for artificial pancreas systems, in order to solve at least one of the problems of the prior art. The present invention can improve the safety of blood glucose control and the robustness of system operation.

[0006] To achieve the above objectives, one aspect of the present invention provides a method for controlling a safe, resilient, and physiologically sensitive intelligent insulin pump for an artificial pancreas system, the method comprising:

[0007] Construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount; Based on blood glucose dynamics, an initial artificial pancreas system model with interval type II fuzzy logic is constructed according to the state vector. The design improves the adaptive event triggering mechanism to generate an initial pump update instruction with a bounded delay; Based on the initial pump update command, the attack signal, and the initial artificial pancreas system model, construct the target artificial pancreas system model; Based on the target artificial pancreas system model, a Lyapunov-Krasovsky functional was constructed to obtain controller parameters; Based on the target artificial pancreas system model and the controller parameters, the insulin pump is controlled.

[0008] In some embodiments, the step of constructing an initial artificial pancreas system model based on glycemic dynamics and the state vector includes the following steps: To obtain blood glucose deviation, insulin infusion rate, and postprandial exogenous glucose intake; The blood glucose deviation is used as the performance output, the insulin infusion rate is used as the control input, and the postprandial exogenous glucose intake is used as the disturbance input. Based on blood glucose dynamics, multiple fuzzy rules and local models of the fuzzy rules are constructed according to the state vector, the performance output, the control input, and the disturbance input. Obtain the interval type II fuzzy membership degree of the premise variables of the fuzzy rule; The interval type II fuzzy membership degree is weighted, reduced and normalized to obtain scalar weights; The initial artificial pancreas system model is constructed based on the scalar weights and the local model.

[0009] In some embodiments, the improved adaptive event triggering mechanism generates an initial pump update instruction with a bounded delay, including the following steps: Preset the monitoring cycle of the continuous glucose monitor; Obtain the state vector at the current monitoring moment to obtain the current physiological state; Obtain the state vector at the last pump update time to get the last stored state at the time of drug administration; Obtain the state difference between the current physiological state and the last stored state at the time of drug administration; The improved adaptive event triggering mechanism is constructed based on the state differences, the last stored state at the time of drug administration, the adaptive threshold factor, and the minimum update interval. Preset dosing delay time; The initial pump update instruction is constructed based on the drug administration delay time, the state difference, and the state feedback gain matrix.

[0010] In some embodiments, constructing a target artificial pancreas system model based on the initial pump update command, the attack signal, and the initial artificial pancreas system model includes the following steps: The attack signal is superimposed on the initial pump update command to obtain the target pump update command; Based on the initial artificial pancreas system model and the target pump update command, the target artificial pancreas system model is constructed.

[0011] In some embodiments, the safety, resilience, and physiologically sensitive intelligent insulin pump control method for artificial pancreas systems further includes the following steps: Anomaly detection and compensation are performed on the target pump update command based on a sparse error data injection detection and fixed window compensation algorithm.

[0012] In some embodiments, the step of constructing a Lyapunov-Krasovsky functional based on the target artificial pancreas system model and obtaining controller parameters includes the following steps: Based on the target artificial pancreas system model, the Lyapunov-Krasovsky functional is constructed. Based on the Lyapunov-Krasovsky functional, and combined with the semi-ring functional technique, the linear matrix inequality conditions are obtained. The controller parameters are obtained based on the solution of the linear matrix inequality condition.

[0013] To achieve the above objectives, another aspect of the present invention proposes a safety-resilient and physiologically-sensing intelligent insulin pump control device for artificial pancreas systems, the device comprising: The first module is used to construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount. The second module is used to construct an initial artificial pancreas system model based on glucose dynamics and the state vector, using a type II fuzzy interval model. The third module is used to design an improved adaptive event triggering mechanism to generate an initial pump update instruction with bounded delay. The fourth module is used to construct a target artificial pancreas system model based on the initial pump update command, the attack signal, and the initial artificial pancreas system model. The fifth module is used to construct a Lyapunov-Krasovsky functional based on the target artificial pancreas system model and obtain controller parameters; The sixth module is used to control the insulin pump based on the target artificial pancreas system model and the controller parameters.

[0014] To achieve the above objectives, another aspect of the present invention provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described above.

[0015] To achieve the above objectives, another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0016] To achieve the above objectives, another aspect of the present invention provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0017] The embodiments of the present invention include at least the following beneficial effects: The present invention provides a safety, resilience, and physiologically sensitive intelligent insulin pump control method and related equipment for artificial pancreas systems. This scheme constructs a state vector based on blood glucose concentration, distal insulin effect, and effective insulin quantity, integrating multiple key physiological parameters to lay a reliable foundation for subsequent precise modeling and control. Based on glucose dynamics, an initial artificial pancreas system model with interval type II fuzzy logic is constructed according to the state vector. The model's robustness and adaptability are enhanced by utilizing interval type II fuzzy logic to handle uncertainties such as individual differences and measurement noise. An improved adaptive event triggering mechanism is designed to generate bounded delays. The initial pump update command is dynamically and adaptively adjusted based on blood glucose levels, significantly reducing redundant pump actions and communication while ensuring control effectiveness, thus improving system resource efficiency and physiological consistency. Based on the initial pump update command, attack signals, and the initial artificial pancreas system model, a target artificial pancreas system model is constructed to enhance the safety of insulin pump control. Based on the target artificial pancreas system model, a Lyapunov-Krasovsky functional is constructed to obtain controller parameters, improving the robustness of system operation. Finally, based on the target artificial pancreas system model and controller parameters, the insulin pump is driven to perform safe, stable, and efficient infusion, achieving automatic regulation of blood glucose levels. Attached Figure Description

[0018] 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.

[0019] Figure 1 This is a flowchart of the intelligent insulin pump control method for safety, resilience, and physiological perception in artificial pancreas systems provided in this embodiment of the invention; Figure 2 This is a schematic diagram of the exogenous glucose occurrence rate curve under the three-meal scenario provided in the embodiments of the present invention; Figure 3a , Figure 3b , Figure 3c This is a schematic diagram of blood glucose change curves under different infusion rates and meal intervals provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the postprandial plasma glucose trajectory when the insulin infusion strategy is tampered with after being attacked by FDI, provided by an embodiment of the present invention. Figure 5 This is a partially enlarged schematic diagram of the tampered infusion interval provided in an embodiment of the present invention; Figure 6This is a comparison chart of the triggering times of three event triggering mechanisms, SETM, AETM, and IAETM, under the condition of no attack, provided by an embodiment of the present invention; Figure 7 This is a schematic diagram of the plasma glucose trajectory when the concurrent detection and recovery strategy is enabled in the event of infusion tampering, as provided in an embodiment of the present invention. Figure 8a , Figure 8b This is a schematic diagram of glucose concentration and corresponding insulin infusion rate curves under restorative infusion action provided in an embodiment of the present invention; Figure 9 This is a comparison chart of the triggering times of three event triggering mechanisms, SETM, AETM, and IAETM, when under attack, provided by an embodiment of the present invention. Figure 10 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this invention; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this invention as detailed in the appended claims.

[0021] It should be noted that although functional modules are divided in the system diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system or the order in the flowchart. The terms "first / S100" and "second / S200" in the specification, claims, and the foregoing drawings may be used herein to describe various concepts, but unless specifically stated otherwise, these concepts are not limited by these terms. These terms are used only to distinguish one concept from another. For example, first information may also be referred to as second information without departing from the scope of the embodiments of the invention, and similarly, second information may also be referred to as first information. Depending on the context, the words "if" or "when" as used herein may be interpreted as "when," "in response to a determination," or "in the event of a determination."

[0022] The terms “at least one,” “multiple,” “each,” “any,” etc., used in this invention, “at least one” includes one, two, or more than two; “multiple” includes two or more than two; “each” refers to each of the corresponding multiple; and “any” refers to any one of the multiple.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0024] Before providing a detailed description of the embodiments of the present invention, some of the nouns and terms involved in the embodiments of the present invention will be explained first. The nouns and terms involved in the embodiments of the present invention are subject to the following interpretations.

[0025] The Bergman Minimal Model is a simple and effective model for describing the glucose-insulin regulatory system. It is used to analyze the effects of insulin sensitivity on glucose tolerance and diabetes risk, as well as the role of insulin secretion.

[0026] An Artificial Pancreas System (APS) is a closed-loop blood glucose control system that integrates a continuous glucose monitoring sensor, control algorithm, and insulin pump to automatically regulate insulin infusion.

[0027] Interval Type-2 (IT-2) fuzzy modeling is a fuzzy modeling method that explicitly describes model uncertainty and individual differences through upper and lower membership functions and their envelopes.

[0028] Static Event-Triggering Mechanism (SETM) is an event triggering strategy that is pre-designed in a control system and whose trigger threshold or triggering conditions remain fixed throughout the entire operation.

[0029] The Adaptive Event-Triggering Mechanism (AETM) is an event triggering strategy that adjusts trigger thresholds or triggering conditions online based on real-time information (such as state errors, interference estimates, network load, or performance metrics) during operation.

[0030] The Improved Adaptive Event-Triggered Mechanism (IAETM) is a type of event triggering strategy that introduces additional correction terms or dynamic parameter update laws on the basis of the traditional Adaptive Event-Triggered Mechanism (AETM). In this invention, IAETM is an event triggering control strategy that adaptively adjusts the trigger threshold and minimum trigger interval according to signal changes, increases the update frequency when blood glucose fluctuates drastically, and reduces unnecessary updates when blood glucose is stable.

[0031] Lyapunov is a core tool for analyzing the stability of nonlinear systems. It determines whether a system converges by constructing an "energy function".

[0032] Among related technologies, existing APS systems are prone to problems such as inaccurate blood glucose regulation, increased risk of hypoglycemia / hyperglycemia, and excessive communication burden due to factors such as noise and delay in continuous blood glucose monitoring, significant individual differences among patients, strong postprandial exogenous disturbances, limited network communication, and even the risk of network attacks.

[0033] In view of this, this invention provides a safety-resilient and physiologically-aware intelligent insulin pump control method and related equipment for artificial pancreas systems. This solution reconstructs a glucose-insulin dynamics model with uncertainties, designs a state-aware event triggering strategy that takes into account the dynamic characteristics of glucose, and a network attack detection-compensation mechanism. This enables on-demand updates of insulin infusion commands and safe rollback under abnormal conditions. While improving the accuracy and safety of glucose control, it effectively reduces communication frequency and control redundancy, thereby improving the robustness and reliability of system operation.

[0034] Figure 1 This is an optional flowchart of an intelligent insulin pump control method for safety, resilience, and physiological awareness in an artificial pancreas system, provided by an embodiment of the present invention. Figure 1 The method may include, but is not limited to, steps S100 to S600: Step S100: Construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount; Step S200: Based on glucose dynamics and the state vector, construct an initial artificial pancreas system model with interval type II fuzzy mapping. Step S300: Design an improved adaptive event triggering mechanism to generate an initial pump update instruction with bounded delay; Step S400: Construct the target artificial pancreas system model based on the initial pump update command, attack signal, and initial artificial pancreas system model; Step S500: Based on the target artificial pancreas system model, construct the Lyapunov-Krasovsky functional and obtain the controller parameters; Step S600: Control the insulin pump based on the target artificial pancreas system model and controller parameters.

[0035] In step S100 of some embodiments, variables such as blood glucose concentration, distal insulin effect, and effective insulin amount are selected to form a state vector. .

[0036] In some embodiments, step S200 may include, but is not limited to, steps S210 to S260: Step S210: Obtain blood glucose deviation, insulin infusion rate, and postprandial exogenous glucose intake; Step S220: Use blood glucose deviation as performance output, insulin infusion rate as control input, and postprandial exogenous glucose intake as perturbation input. Step S230: Based on blood glucose dynamics, construct multiple fuzzy rules and local models of the fuzzy rules according to the state vector, performance output, control input, and disturbance input. Step S240: Obtain the interval type II fuzzy membership degree of the premise variables of the fuzzy rule; Step S250: Perform weighted reduction and normalization on the interval type II fuzzy membership degree to obtain scalar weights; Step S260: Construct an initial artificial pancreas system model based on scalar weights and the local model.

[0037] In some embodiments, steps S210 to S220 are used as performance outputs, with blood glucose deviation as the benchmark. With insulin infusion rate To control input, postprandial exogenous glucose intake was used as the perturbation input. .

[0038] In step S230 of some embodiments, based on the classical glycemic kinetic equation, a local model described by multiple fuzzy rules is constructed according to the state vector, performance output, control input, and disturbance input. For example, the rules... If the prerequisite variable It is the first A fuzzy set Given the described category, the expression for the local model is: (1) In the formula, Represents the state vector in relation to time The derivative; Indicates the first The system state matrix corresponding to the fuzzy rules describes the linear dynamic relationship between blood glucose and insulin states. Indicates the first The control input matrix corresponding to each fuzzy rule describes the effect of insulin infusion on the system state; Indicates the first The perturbation input matrix corresponding to the fuzzy rules describes the impact of perturbations such as postprandial exogenous glucose on the system state. Indicates the first The output matrix corresponding to each fuzzy rule is used to obtain the output quantity related to blood glucose deviation from the state vector.

[0039] In step S240 of some embodiments, the premise variables of the fuzzy rule are... This premise variable is relative to the first The membership degree of a fuzzy set is characterized by the IT-2 trigger strength. Therefore, the expression for obtaining the interval type-2 fuzzy membership of the premise variable is: (2) In the formula, Indicates the interval type II fuzzy membership degree; Indicates the lower bound membership function; Indicates the membership function of the upper bound.

[0040] In step S250 of some embodiments, the interval type II fuzzy membership degree is weighted and reduced by introducing a condition that satisfies... Weighting factor and We can obtain the following expression: (3) Then, the weighted reduced interval type II fuzzy membership degree is normalized to obtain the expression for the scalar weight: (4) In the formula, Indicates the first The weighted reduction of the interval type II fuzzy membership degree of the fuzzy rules; , Indicates the weighting factor; Indicates the lower bound of the trigger strength; Indicates the upper bound of the trigger strength; This represents scalar weights, also known as normalized weights (membership). Indicates the first The weighted reduction of the interval type II fuzzy membership degree of the fuzzy rules; Represents the total number of fuzzy rules; The index representing the fuzzy rule.

[0041] In step S260 of some embodiments, the local model is weighted and fused according to the scalar weights and the local model to obtain the expression of the initial artificial pancreas system model as follows: (5) In step S300 of some embodiments, an improved adaptive event triggering mechanism (IAETM) is designed to dynamically adjust the update rhythm of the insulin pump according to blood glucose. This enables on-demand updates during blood glucose fluctuations and extends the update interval and introduces a quiet period during stable blood glucose phases. This reduces redundant communication and pump execution burden while ensuring the safety of blood glucose control, thereby improving system resource efficiency and physiological consistency.

[0042] In some embodiments, step S300 may include, but is not limited to, steps S310 to S370: Step S310: Preset the monitoring cycle of the continuous glucose monitor; Step S320: Obtain the state vector at the current monitoring moment to obtain the current physiological state; Step S330: Obtain the state vector at the last pump update time to get the last stored state at the time of drug administration; Step S340: Obtain the state difference between the current physiological state and the last stored state at the time of drug administration; Step S350: Based on the state differences, the last stored state at the time of drug administration, the adaptive threshold factor, and the minimum update interval, construct an improved adaptive event triggering mechanism; Step S360: Preset the drug administration delay time; Step S370: Construct an initial pump update instruction based on the drug administration delay time, state difference, and state feedback gain matrix.

[0043] In step S310 of some embodiments, the monitoring period of the continuous glucose monitor is preset to be... Then the monitoring time is... , is used to represent the sequence number of the discrete sampling time. Where, , Represents the set of non-negative integers.

[0044] In some embodiments, in steps S320 to S340, it is set This is the most recent pump update time. Get the current monitoring time. The state vector yields the current physiological state. Get the most recent (previous) pump update time. The state vector is used to obtain the final stored state vector at the time of drug administration. The difference between the current physiological state and the last stored state at the time of drug administration is: In the formula, For state differences, Current physiological state This is the final storage state at the time of drug administration.

[0045] In step S350 of some embodiments, based on the state difference, the last stored state at the time of drug administration, the adaptive threshold factor, and the minimum update interval, the event triggering conditions for the improved adaptive event triggering mechanism can be designed as follows: (6) In the formula, Indicates the transpose operation; Represents the weight matrix. ; This represents the adaptive threshold factor.

[0046] Furthermore, a minimum update interval inspired by physiology is preset. To avoid frequent updates.

[0047] At each pump update, the adaptive threshold factor is updated as follows: (7) In the formula, Indicates the first The adaptive trigger threshold corresponding to the next monitoring (update) time; This indicates the discrete monitoring (or pump update) sequence number, and the corresponding time is recorded as follows: ; Indicates the lower bound of the threshold; Indicates the upper bound of the threshold; Represents the natural base; This indicates the sensitivity adjustment parameter to blood glucose status. .

[0048] In some embodiments, in steps S360 to S370, a segmented sensing-drug delivery delay is introduced. ,satisfy The initial pump update command is represented in a delayed form: (8) In the formula, This indicates the initial pump update command; Indicates the drug administration delay time; Indicates the upper limit of the drug administration delay time; Represents the state feedback gain matrix; Indicates the current moment.

[0049] By introducing segmented sensing-drug administration delay, event-triggered drug administration is modeled as a sensing-decision-execution pathway with bounded time delay, which facilitates subsequent stability analysis and controller design.

[0050] Transforming the delayed initial pump update instruction in formula (8) into a fuzzy rule-based initial pump update instruction yields: rule :if yes ,So: (9) In the formula, Indicates the first The initial pump update instruction for the fuzzy rule; Indicates the first The state feedback gain matrix of the fuzzy rule.

[0051] In some embodiments, step S400 may include, but is not limited to, steps S410 to S420: Step S410: Superimpose the attack signal onto the initial pump update command to obtain the target pump update command; Step S420: Construct the target artificial pancreas system model based on the initial artificial pancreas system model and the target pump update command.

[0052] In step S410 of some embodiments, during networked APS operation, the focus is on security threats to the command channel from the decision module to the insulin pump, particularly FDI attacks. By superimposing an attack signal onto the initial pump update command, the actual infusion dose and timing can be altered, resulting in the expression for the target pump update command: (10) In the formula, Indicates a target pump update command; Indicates the time of update Injected attack signals, and in the range Internal persistence. The attack signal satisfies the boundedness condition: ,in, This is a weight matrix that reflects the intensity of the attack.

[0053] In step S420 of some embodiments, by integrating the initial artificial pancreas system model and the target pump update instruction, the following target artificial pancreas system model can be constructed: (11) In the formula, Indicates the first The interval type II fuzzy rule at the sampling time Normalized weights (membership degrees).

[0054] In some embodiments, a sparse error data injection (FDI) detection and fixed window compensation algorithm is designed to significantly enhance the security and robustness of the artificial pancreas system in malicious network environments. This algorithm detects attacks by periodically comparing the infusion rate of the commanded data with the actual infusion rate, using a multi-evidence fusion strategy based on consistency of sign, amplitude ratio, and rate of change. Once the accumulated abnormal evidence exceeds a threshold and an FDI attack is confirmed, the system immediately initiates a fixed-duration compensation window. During this window, the system locks and maintains a safe reference infusion command to ensure stable and consistent insulin infusion, while simultaneously performing necessary compensation infusions based on the current blood glucose status, effectively preventing excessively high or low infusions due to malicious tampering.

[0055] By coupling sparse detection and fixed window compensation, the stability of pump-side drug delivery can be maintained even under attack conditions, enhancing the system's ability to protect against erroneous data injection attacks and ensuring the drug delivery safety and operational reliability of the artificial pancreas system.

[0056] In step S500 of some embodiments, stability and performance analysis based on a fixed trigger threshold is conducted for insulin infusion control of the artificial pancreas system under a traditional event-triggered mechanism. To ensure the safe and stable operation of the system under model uncertainty and network attacks, this embodiment proposes two key theorems: Theorem 1 proposes a set of linear matrix inequalities to guarantee the asymptotic stability of the closed-loop system and satisfy... Sufficient conditions for performance indicators; Theorem 2, based on this, gives a specific solution for the controller parameters (the controller's state feedback gain matrix), providing a feasible method for the design and implementation of controllers in clinical practice.

[0057] In some embodiments, step S500 may include, but is not limited to, steps S510 to S530: Step S510: Construct the Lyapunov-Krasovsky functional based on the target artificial pancreas system model; Step S520: Based on the Lyapunov-Krasovsky functional and combined with the semi-ring functional technique, obtain the linear matrix inequality conditions. Step S530: Obtain the controller parameters based on the solution of the linear matrix inequality conditions.

[0058] In steps S510 to S520 of some embodiments, for Theorem 1, based on the target artificial pancreas system model, a Lyapunov-Krasovsky functional is constructed, and combined with semi-ring functional techniques, a method is derived to ensure the asymptotic stability of the system and satisfy the following conditions: Performance indicators The linear matrix inequality condition. This stability condition is stated as: given a positive scalar and If there exists a set of positive definite symmetric matrices (as in equation (12)) and a set of real matrices of appropriate dimension (as in equation (13)) such that the set of matrix inequalities in equation (14) is simultaneously feasible.

[0059] (12) (13) , (14) In the formula, All of these are symmetric matrix decision variables introduced in stability analysis and performance design, used to construct the Lyapunov-Krasovsky functional and form the corresponding linear matrix inequality constraints. This is the connection matrix in the event triggering conditions, used to map state variables to quadratic indices in the triggering criteria; It is a set of positive definite symmetric matrices; The equivalence matrices are typical relaxation matrices introduced in the semi-cyclic functional, used to construct the semi-cyclic functional and form corresponding linear matrix inequality constraints in the stability analysis, so as to reduce the conservatism of the stability criterion. The free matrix introduced for scaling is used to scale and linearize related cross terms; This is a free matrix connected to the state equations, used to embed the coefficient matrix in the system state equations into the linear matrix inequality conditions, thereby explicitly reflecting the system dynamics in the stability criterion; Let be the set of real matrices; For the first A linear matrix inequality (LMI) condition matrix; This means for all From 1 to 4, matrix All are negative definite.

[0060] in: ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; .

[0061] To accurately characterize the system's behavior within the permissible region and facilitate subsequent stability analysis, the following Lyapunov-Krasovskii functionals (LKFs) are constructed: ; ; ; ; ; ; ; ; By differentiating LKFs, we can obtain the following expression: , (15) An inequality expression is given after differentiating the constructed Lyapunov–Krasovsky functional along the trajectory of the closed-loop system, which is used to derive stability and performance criteria in conjunction with the semi-loop functional method.

[0062] Therefore, according to the semi-cyclic functional method, when the LMIs condition in formula (14) holds, it can be guaranteed that... Thus, the closed-loop target artificial pancreas system model is asymptotically stable and satisfies the preset conditions. Performance metrics.

[0063] In step S530 of some embodiments, for Theorem 2, based on the target artificial pancreas system model, a Lyapunov-Krasovsky functional is constructed, and combined with semi-ring functional techniques, a method is derived to ensure the asymptotic stability of the system and satisfy the following conditions: Performance indicators The linear matrix inequality condition. This stability condition is stated as: given a positive scalar and If there exists a set of positive definite symmetric matrices (as in equation (16)) and a set of real matrices of appropriate dimension (as in equation (17)) such that the set of matrix inequalities in equation (18) is simultaneously feasible.

[0064] (16) (17) , (18) Among them, by scaling and variable substitution of each matrix in equations (12), (13) and (14) to introduce auxiliary design variables, we can obtain each matrix in equations (16), (17) and (18), and thus solve the controller feedback gain matrix through equations (16) to (18).

[0065] Then, the state feedback gain matrix of the controller can be derived. .

[0066] By definition As an auxiliary matrix variable related to the controller feedback gain matrix, the original LMIs criterion in formula (14) can be equivalently transformed into formula (18), which helps to improve the controller gain matrix. The calculation and implementation of system stability determination conditions.

[0067] In step S600 of some embodiments, the target artificial pancreas system model and the controller parameters obtained by solving linear matrix inequalities (including the state feedback gain matrix corresponding to each fuzzy rule) are used to... and event trigger weight matrix During system operation, real-time collected signals such as blood glucose concentration are processed via state vector. After characterization, the normalized weights are first calculated according to the interval type II fuzzy rule. and It also determines whether control commands need to be updated at the current moment based on the event triggering conditions; when an update is triggered, it is based on the bounded delay state. Differences from state Using the state feedback gain matrix Dynamically calculate the current control input It generates specific insulin infusion rate instructions; these instructions are sent to the insulin pump actuator via a secure communication protocol, driving it to complete a precise drug infusion, thus forming a closed loop of "sensing-decision-execution-feedback", ultimately achieving automatic, safe, stable and efficient control of the insulin pump.

[0068] In some embodiments, a closed-loop APS integrating a continuous glucose monitor, a control decision module, and an insulin pump was constructed based on the MATLAB / Simulink simulation platform to verify the performance and effectiveness of the proposed IT-2 fuzzy control method and ETM in blood glucose regulation. This model was established with reference to the Bergman minimal model widely used in clinical medical research. Its core physiological parameters were set according to the characteristics of typical type 1 diabetes patients, and the specific values ​​of glucose kinetics, insulin kinetics, and related control parameters are listed in Table 1.

[0069] Table 1: Physiological parameters used in the APS model

[0070] The coefficient matrix of the IT-2 fuzzy APS using two rules is shown below: ; ; ; ; The definitions of upper and lower bound weights are summarized in Table 2, while the membership functions of upper and lower bounds are listed in Table 3.

[0071] Table 2: Weights

[0072] Table 3: Membership Functions

[0073] here Both weighting functions and membership functions are uniformly expressed in a bounded form and applied to the current state in the system model. and the sampling state in the controller implementation .

[0074] By solving the derived LMI condition (18), and setting the maximum sampling interval to... And with the nominal input gain c fixed at 0.9, the calculated controller gain is: ; To comprehensively evaluate the performance of the proposed artificial pancreas system, Figure 2 The dynamics of exogenous glucose input induced by three standardized meals during the simulation were demonstrated. A rapid rise followed by a slow decline in glucose levels was observed after each meal, accurately simulating the physiological response of the human body after eating. These characteristic glucose input curves provide a realistic physiological context for subsequent evaluation of the control system's disturbance rejection performance.

[0075] Figure 3a The study demonstrated the dynamic glycemic response under a configuration with a fixed insulin rate of 16.6 mU / min and a meal interval of 250 minutes. Figure 3b The study demonstrated the dynamic glycemic response under a fixed meal interval of 300 minutes and an insulin rate of 16.6 mU / min. Figure 3cThis study demonstrates the dynamic glycemic response under a fixed meal interval of 300 minutes and an insulin rate of 40.0 mU / min, comparing the dynamic glycemic response under different basal insulin infusion rates and meal intervals. The results show that relying solely on basal insulin infusion is insufficient to effectively suppress postprandial glycemic fluctuations, and while an excessively high basal rate can reduce peak glycemic levels, it increases the risk of hypoglycemia. This finding physiologically validates the necessity of employing an intelligent closed-loop feedback control strategy, providing important evidence for the optimized design of control parameters.

[0076] Figure 4 This study demonstrates the system's postprandial blood glucose response when the insulin infusion strategy is subjected to a spoofed data injection attack. Compared to normal conditions, the blood glucose trajectory exhibits a significantly elevated peak and a delayed recovery process after the infusion strategy was maliciously altered, with the highest blood glucose concentration exceeding the clinically safe threshold of 180 mg / dL. This result vividly illustrates the serious threat that cyberattacks pose to treatment safety and highlights the importance of designing cybersecurity protection mechanisms.

[0077] Figure 5 By magnifying the attacked region and using linear fitting, the impact of abnormal infusion on the rate of glucose metabolism was quantitatively analyzed. The results showed that during the period of infusion tampering, the rate of glucose decline significantly slowed from the normal 0.57 mg / dL / min to 0.36 mg / dL / min, confirming the decrease in insulin efficacy caused by the attack. This finding provides a quantitative reference for adjusting parameters in subsequent compensation strategies.

[0078] Figure 6 The trigger timing distributions of SETM, AETM, and IAETM were compared under attack-free conditions. The results show that, compared to SETM and AETM, IAETM maintains equivalent glycemic control performance while exhibiting a sparser trigger sequence, particularly significantly reducing unnecessary control updates during the glycemic steady-state phase. This characteristic fully demonstrates the significant advantage of IAETM in balancing control performance and system resource consumption, validating its potential in reducing pump hardware load and conserving communication resources, and is of great value for the long-term clinical application of the system.

[0079] Figure 7 This demonstrates the system's 24-hour plasma glucose trajectory after suffering an FDI attack and activating detection and compensation mechanisms. Figures 3a to 3cCompared to the unprotected system response, after enabling the security mechanism, despite suffering the same attack, blood glucose concentrations were consistently maintained within the clinically safe range of 70-180 mg / dL, without significant abnormal increases or dangerous fluctuations. This result directly demonstrates that the aforementioned detection and compensation mechanism can effectively resist cyberattacks, ensuring the integrity of insulin infusion and the safety of treatment. The key lies in the fact that the compensation mechanism, by maintaining an infusion baseline based on security logic, offsets the impact of attack signals, ensuring the continuous and reliable operation of the closed-loop system.

[0080] Figure 8a The study demonstrates the trajectory of blood glucose changes after insulin infusion was restored during the period when the attack occurred and the recovery mechanism was activated. Figure 8b The changes in insulin infusion rate after infusion were demonstrated during the period when the attack occurred and the recovery mechanism was activated. Figure 8a and Figure 8b Together, they depicted the synergistic changes in glucose concentration and corresponding insulin infusion rate during the attack and recovery mechanism activation, clearly revealing the working sequence and effect of the safety mechanism: In the early stage of the attack (time t1), blood glucose began to rise abnormally, at which point the detection mechanism identified the inconsistency; at time t2, the compensation window was activated, the insulin infusion rate was corrected and increased to a reasonable physiological profile; subsequently, the upward trend in blood glucose was curbed and began to decline. This process vividly demonstrates how this invention integrates the "detection-decision-compensation" safety closed loop into conventional control, achieving proactive defense.

[0081] Figure 9 The triggering intervals and effective thresholds of three event triggering mechanisms—SETM, AETM, and IAETM—were compared under the same severe scenario of an FDI attack. The results show that IAETM maintains excellent resource efficiency while ensuring system stability, and its triggering sequence is significantly sparser than the other two mechanisms. This demonstrates that the adaptability of IAETM extends beyond normal operating conditions. When security events occur and the system requires additional resources for compensation and adjustment, IAETM can intelligently maintain necessary control updates through its dynamic thresholds, while filtering out a large number of unnecessary redundant triggers, ensuring the system's high efficiency and stability under stress.

[0082] In this embodiment of the invention, firstly, a closed-loop system structure integrating a continuous glucose monitor, a control decision module, and an insulin pump is established, and the IT-2 fuzzy modeling method is used to accurately describe the uncertainty of glucose-insulin physiological dynamics. Secondly, an IAETM is proposed, which adaptively adjusts the update rhythm of the insulin pump according to blood glucose dynamics, significantly reducing communication and hardware load while ensuring control performance. To address security threats in networked environments, an attack detection algorithm based on multi-evidence fusion and a fixed window compensation strategy are designed to effectively ensure the integrity and safety of insulin infusion. By constructing linear matrix inequality conditions based on LKFs, a system is designed to ensure asymptotic stability and satisfy the following conditions. A robust controller with high performance metrics. Simulation results verify that the proposed strategy exhibits excellent glycemic control accuracy, rapid dynamic response, significantly reduced communication overhead, and strong security resilience under dietary disturbances and network attacks, demonstrating good clinical applicability and engineering application value. This invention provides solid theoretical support and technical implementation path for constructing a safe, reliable, and resource-efficient artificial pancreas system for practical clinical applications.

[0083] This invention also provides a safety- and resilience-enhancing intelligent insulin pump control device for artificial pancreas systems, which can realize the above-mentioned safety- and resilience-enhancing intelligent insulin pump control method for artificial pancreas systems. The device includes: The first module is used to construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount. The second module is used to construct an initial artificial pancreas system model based on glucose dynamics and state vectors, using a type II fuzzy interval model. The third module is used to design an improved adaptive event triggering mechanism to generate an initial pump update instruction with bounded delay. The fourth module is used to construct the target artificial pancreas system model based on the initial pump update command, attack signal, and initial artificial pancreas system model. The fifth module is used to construct the Lyapunov-Krasovsky functional based on the target artificial pancreas system model and obtain the controller parameters; The sixth module is used to control the insulin pump based on the target artificial pancreas system model and controller parameters.

[0084] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0085] This invention also provides an electronic device, which includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including a tablet computer, an in-vehicle computer, or similar device.

[0086] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0087] refer to Figure 10 , Figure 10 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention. The memory 1002 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and is called and executed by the processor 1001. Input / output interface 1003 is used to implement information input and output; The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004); The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0088] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0089] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0090] This invention also provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions to cause the computer device to perform the aforementioned method.

[0091] In summary, the safety, resilience, and physiologically sensitive intelligent insulin pump control method and related equipment for artificial pancreas systems according to embodiments of the present invention have the following advantages: 1. A physiological perception APS model based on interval type II fuzzy logic was constructed. By introducing footprint uncertainty, the nonlinearity and inter-individual variability of glucose-insulin dynamics are effectively characterized, thereby improving the physiological consistency of the model and the clinical rationality of the control strategy.

[0092] 2. An IAETM is proposed, which can dynamically and adaptively adjust the insulin pump update threshold and minimum update interval based on blood glucose levels. While ensuring control performance, it significantly reduces redundant instructions and pump hardware load, achieving resource-efficient closed-loop control.

[0093] 3. A false data injection attack detection and compensation mechanism for pump command channels was designed. Through multi-evidence consistency verification and fixed window compensation strategy, the threat of network attacks to drug administration integrity is effectively identified and mitigated, and the security and reliability of the system in malicious environments are enhanced.

[0094] 4. Based on LKFs and semi-ring functional techniques, a stability analysis framework suitable for event-triggered sampling and network attack scenarios is constructed. A stability criterion expressed in linear matrix inequalities is derived to ensure the asymptotic stability of the system under various uncertainties. performance.

[0095] 5. Through numerical simulation verification under typical dietary scenarios and network attacks, the proposed method has shown excellent performance in maintaining a safe blood glucose range, reducing pump update frequency, and resisting network threats, demonstrating good clinical applicability and engineering promotion potential.

[0096] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0097] 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.

[0098] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0099] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A method for controlling a safe, resilient, and physiologically sensitive intelligent insulin pump for an artificial pancreas system, characterized in that, Includes the following steps: Construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount; Based on blood glucose dynamics, an initial artificial pancreas system model with interval type II fuzzy logic is constructed according to the state vector. The design improves the adaptive event triggering mechanism to generate an initial pump update instruction with a bounded delay; Based on the initial pump update command, the attack signal, and the initial artificial pancreas system model, construct the target artificial pancreas system model; Based on the target artificial pancreas system model, a Lyapunov-Krasovsky functional was constructed to obtain controller parameters; Based on the target artificial pancreas system model and the controller parameters, the insulin pump is controlled.

2. The method according to claim 1, characterized in that, The method for constructing an initial artificial pancreas system model based on glycemic dynamics and the state vector includes the following steps: To obtain blood glucose deviation, insulin infusion rate, and postprandial exogenous glucose intake; The blood glucose deviation is used as the performance output, the insulin infusion rate is used as the control input, and the postprandial exogenous glucose intake is used as the perturbation input. Based on blood glucose dynamics, multiple fuzzy rules and local models of the fuzzy rules are constructed according to the state vector, the performance output, the control input, and the disturbance input. Obtain the interval type II fuzzy membership degree of the premise variables of the fuzzy rule; The interval type II fuzzy membership degree is weighted, reduced and normalized to obtain scalar weights; The initial artificial pancreas system model is constructed based on the scalar weights and the local model.

3. The method according to claim 1, characterized in that, The improved adaptive event triggering mechanism generates an initial pump update instruction with a bounded delay, including the following steps: Preset the monitoring cycle of the continuous glucose monitor; Obtain the state vector at the current monitoring moment to obtain the current physiological state; Obtain the state vector at the last pump update time to get the last stored state at the time of drug administration; Obtain the state difference between the current physiological state and the last stored state at the time of drug administration; The improved adaptive event triggering mechanism is constructed based on the state differences, the last stored state at the time of drug administration, the adaptive threshold factor, and the minimum update interval. Preset dosing delay time; The initial pump update instruction is constructed based on the drug administration delay time, the state difference, and the state feedback gain matrix.

4. The method according to claim 1, characterized in that, The step of constructing the target artificial pancreas system model based on the initial pump update command, attack signal, and initial artificial pancreas system model includes the following steps: The attack signal is superimposed on the initial pump update command to obtain the target pump update command; Based on the initial artificial pancreas system model and the target pump update command, the target artificial pancreas system model is constructed.

5. The method according to claim 4, characterized in that, The method further includes the following steps: Anomaly detection and compensation are performed on the target pump update command based on a sparse error data injection detection and fixed window compensation algorithm.

6. The method according to claim 1, characterized in that, The process of constructing a Lyapunov-Krasovsky functional based on the target artificial pancreas system model and obtaining controller parameters includes the following steps: Based on the target artificial pancreas system model, the Lyapunov-Krasovsky functional is constructed. Based on the Lyapunov-Krasovsky functional, and combined with the semi-ring functional technique, the linear matrix inequality conditions are obtained. The controller parameters are obtained based on the solution of the linear matrix inequality condition.

7. A safe, resilient, and physiologically sensitive intelligent insulin pump control device for artificial pancreas systems, characterized in that, include: The first module is used to construct a state vector based on blood glucose concentration, distal insulin effect, and effective insulin amount. The second module is used to construct an initial artificial pancreas system model based on glucose dynamics and the state vector, using a type II fuzzy interval model. The third module is used to design an improved adaptive event triggering mechanism to generate an initial pump update instruction with bounded delay. The fourth module is used to construct a target artificial pancreas system model based on the initial pump update command, the attack signal, and the initial artificial pancreas system model. The fifth module is used to construct a Lyapunov-Krasovsky functional based on the target artificial pancreas system model and obtain controller parameters; The sixth module is used to control the insulin pump based on the target artificial pancreas system model and the controller parameters.

8. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The storage medium stores a program that is executed by a processor to implement the method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images