Closed-loop artificial pancreas insulin infusion control system

Abstract

The application discloses a closed-loop artificial pancreas insulin infusion control system, which comprises a detection module, a program module connected with the detection module, an rPID algorithm and a target blood glucose value G B preset in the program module for converting blood glucose asymmetry in an original physical space to blood glucose risk approximate symmetry in a risk space, the rPID algorithm calculates an insulin infusion indication through the blood glucose risk, and an infusion module connected with the program module, wherein the program module controls the infusion module to infuse insulin according to the insulin infusion indication calculated by the rPID algorithm. The rPID algorithm in the system converts blood glucose asymmetry in an original physical space to blood glucose risk approximate symmetry in a risk space, which can not only retain the simple and robust characteristics of the PID algorithm, but also has the advantages of precision and flexibility, and realizes the precise control of the closed-loop artificial pancreas insulin infusion system.

Description

Technical Field

[0001] This invention relates primarily to the field of medical devices, and in particular to a closed-loop artificial pancreas insulin infusion control system. Background Technology

[0002] In a healthy person, the pancreas automatically secretes the necessary insulin / glucagon based on blood glucose levels, thus maintaining a reasonable range of blood sugar fluctuations. However, in diabetic patients, pancreatic function is abnormal, and the pancreas is unable to secrete the required insulin normally. Diabetes is a metabolic disease and is a lifelong condition. Current medical technology cannot cure diabetes; it can only control the occurrence and development of diabetes and its complications by stabilizing blood sugar levels.

[0003] Diabetic patients need to have their blood glucose levels checked before insulin injection. Current methods allow for continuous glucose monitoring, sending the data in real-time to a display device for user viewing. This method is called Continuous Glucose Monitoring (CGM). The device is attached to the skin, with its probe inserted into the subcutaneous fluid to perform the measurement. Based on the CGM reading, the infusion device delivers the required amount of insulin subcutaneously, creating a closed-loop or semi-closed-loop artificial pancreas.

[0004] Currently, to achieve closed-loop or semi-closed-loop control of insulin infusion, proportional-integral-derivative (PID) controllers and mathematical models of the metabolism and interaction between glucose and insulin in the human body are widely studied and used. However, due to the simple structure and few parameters of PID algorithms, the optimization results are generally conservative to achieve robustness and prevent overshoot oscillations, resulting in a long stabilization time. This makes them suitable for scenarios where target adjustments are infrequent, external disturbances are minimal, or the mechanism of action is stable. However, the use scenarios of artificial pancreas include large-amplitude, high-frequency, and rapid-acting events such as meals and exercise. Furthermore, there is a significant time delay between food absorption and insulin action, making it impossible for simple PID algorithms to achieve ideal control results.

[0005] Therefore, there is an urgent need in the current technology for a closed-loop artificial pancreas insulin infusion control system containing an optimized PID algorithm. Summary of the Invention

[0006] This invention discloses a closed-loop artificial pancreas insulin infusion control system. The rPID algorithm in this system transforms the asymmetric blood glucose in the original physical space into an approximately symmetric blood glucose risk in the risk space. It retains the simplicity and robustness of the PID algorithm while also having the advantages of precision and flexibility, thus achieving precise control of the closed-loop artificial pancreas insulin infusion system.

[0007] This invention discloses a closed-loop artificial pancreas insulin infusion control system, comprising: a detection module for continuously detecting the current blood glucose value G; and a program module connected to the detection module, wherein the program module is pre-programmed with an rPID algorithm to convert blood glucose levels that are asymmetrical in the original physical space to blood glucose levels that are approximately symmetrical in the risk space, and a target blood glucose value G. B The rPID algorithm calculates insulin infusion instructions based on blood glucose risk; and the infusion module is connected to the program module. Based on the insulin infusion instructions calculated by the rPID algorithm, the program module controls the infusion module to infuse insulin.

[0008] According to one aspect of the present invention, the rPID algorithm is as follows:

[0009]

[0010] in:

[0011] rPID(t) is the insulin infusion instruction sent to the infusion module after risk conversion;

[0012] K P This is the gain coefficient for the proportional portion;

[0013] K I This is the gain coefficient for the integral part;

[0014] K D The gain coefficient is the differential part.

[0015] C is a constant;

[0016] r represents blood glucose risk.

[0017] According to one aspect of the invention, the blood glucose risk r is related to the current blood glucose value G and the target blood glucose value G. B The blood glucose deviation Ge is related to the blood glucose deviation, and the blood glucose deviation Ge = GG B .

[0018] According to one aspect of the present invention, the relationship between the blood glucose risk r and the blood glucose deviation Ge is as follows:

[0019]

[0020] According to one aspect of the present invention, the relationship between the blood glucose risk r and the blood glucose deviation Ge is as follows:

[0021]

[0022] According to one aspect of the present invention, the blood glucose risk r is calculated as follows:

[0023]

[0024] in,

[0025] r(G) = 10 * f(G) 2

[0026] Where f(G) is the transformation function, and its calculation formula is:

[0027] f(G) = 1.509 * [(ln(G)) 1.084 -5.381].

[0028] According to one aspect of the present invention, the blood glucose risk r is calculated as follows:

[0029]

[0030] According to one aspect of the invention, the maximum value of the blood glucose risk r is limited to: |r| = min(|r|, n).

[0031] According to one aspect of the invention, the maximum value n is limited to 0 to 80 mg / dL.

[0032] According to one aspect of the invention, the maximum value n is limited to 60 mg / dL.

[0033] According to one aspect of the invention, when the current blood glucose value G is greater than the target blood glucose value G... B At that time, the blood glucose risk r is calculated as follows:

[0034] r = r(G), if G > G B

[0035] in,

[0036] r(G) = 10 * f(G) 2

[0037] Where f(G) is the transformation function, and its calculation formula is:

[0038] f(G) = 1.509 * [(ln(G)) 1.084 -5.381];

[0039] When the current blood glucose level G is not greater than the target blood glucose level G B At that time, the blood glucose risk r is calculated as follows:

[0040] r = GG B if G≤G B .

[0041] According to one aspect of the invention, the maximum value of the blood glucose risk r is limited to: |r| = min(|r|, n).

[0042] According to one aspect of the invention, the maximum value n is limited to 0 to 80 mg / dL.

[0043] According to one aspect of the invention, the maximum value n is limited to 60 mg / dL.

[0044] According to one aspect of the invention, when the current blood glucose value G is not greater than the target blood glucose value G... B At that time, the blood glucose risk r is calculated as follows:

[0045] r = -r(G), if G ≤ G B

[0046] in:

[0047] r(G) = 10 * f(G) 2

[0048] Under the transformation function f(G):

[0049] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0050] When the current blood glucose level G is greater than the target blood glucose level G B At that time, the blood glucose risk r is calculated as follows:

[0051] r = --4.8265 * 10 4 -4*G 2 +0.45563*G-44.855, if G>G B .

[0052] According to one aspect of the invention, when the current blood glucose value G is not greater than the target blood glucose value G... B At that time, the blood glucose risk r is calculated as follows:

[0053] r = -r(G), if G ≤ G B

[0054] in:

[0055] r(G) = 10 * f(G) 2

[0056] Under the transformation function f(G):

[0057] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0058] When the current blood glucose level G is greater than the target blood glucose level G B At that time, the blood glucose risk r is calculated as follows:

[0059]

[0060] According to one aspect of the invention, when the current blood glucose value G is not greater than the target blood glucose value G... B At that time, the blood glucose risk r is calculated as follows:

[0061] r = -r(G), if G ≤ G B

[0062] in:

[0063] r(G) = 10 * f(G) 2

[0064] Under the transformation function f(G):

[0065] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0066] When the current blood glucose level G is greater than the target blood glucose level G B At that time, the blood glucose risk r is calculated as follows:

[0067] r = 100 * (GG B ) / G, if G>G B .

[0068] According to one aspect of the invention, the target blood glucose value G B The concentration is 80–140 mg / dL.

[0069] According to one aspect of the invention, the target blood glucose value G B The concentration is 110–120 mg / dL.

[0070] According to one aspect of the present invention, the rPID algorithm further includes one or more of the following processing methods:

[0071] ① Based on the insulin absorption delay in the artificial pancreas control system, deduct the amount of unabsorbed plasma insulin in the body.

[0072]

[0073] in,

[0074] γ is the compensation coefficient of the estimated plasma insulin concentration to the algorithm output;

[0075] For estimation of plasma insulin concentration;

[0076] rPIDc(t) represents the compensated infusion instruction sent to the insulin infusion system after risk conversion;

[0077] rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion;

[0078] ②Based on the delayed onset of insulin in the artificial pancreas control system, subtract the amount of insulin (IOB(t)) that has not yet taken effect in the body:

[0079] rPID′(t)=rPID(t)-IOB(t)

[0080] in:

[0081] rPID′(t) represents the infusion instruction sent to the insulin infusion system after risk conversion, excluding the amount of insulin that has not yet taken effect in the body;

[0082] rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion;

[0083] IOB(t) represents the amount of insulin in the body that has not yet taken effect at time t.

[0084] ③ An autoregressive method was used to compensate for the sensing delay of blood glucose and interstitial fluid glucose concentrations.

[0085] According to one aspect of the invention, plasma insulin concentration is estimated by an autoregressive method.

[0086] According to one aspect of the invention, the compensation coefficient γ is 0.4-0.6.

[0087] According to one aspect of the invention, the compensation coefficient γ is 0.5.

[0088] According to one aspect of the invention, the amount of insulin IOB(t) that has not yet taken effect in the body is obtained by an IOB curve.

[0089] According to one aspect of the invention, when calculating the amount of insulin IOB(t) that has not yet taken effect in the body, the amount of insulin is...

[0090] Management of insulin before and after meals:

[0091] IOB(t) = IOB m，t +IOB o，t

[0092]

[0093] in,

[0094] IOB m，t This represents the amount of insulin consumed after a meal that has not yet taken effect in the body at time t.

[0095] IOB o，t This represents the amount of non-meal insulin that has not yet taken effect in the body at time t;

[0096] Di (i = 2 - 8) represent the corresponding coefficients of the IOB curves for insulin action time i;

[0097] I m，t Indicates the amount of insulin consumed after a meal;

[0098] I 0，t This indicates the amount of insulin not consumed during meals;

[0099] IOB(t) represents the amount of insulin in the body at time t that has not yet taken effect.

[0100] According to one aspect of the present invention, two of the detection module, program module and infusion module are interconnected to form an integral structure, and are respectively attached to different positions on the skin along with the third module.

[0101] According to one aspect of the present invention, the detection module, the program module, and the infusion module are connected to form an integral structure and are attached to the same location on the skin.

[0102] Compared with the prior art, the technical solution of the present invention has the following advantages:

[0103] In the closed-loop artificial pancreas insulin infusion control system disclosed in this invention, the rPID algorithm transforms the asymmetric blood glucose in the original physical space into an approximately symmetric blood glucose risk in the risk space. It retains the simplicity and robustness of the PID algorithm while also possessing the advantages of precision and flexibility, thus achieving precise control of the closed-loop artificial pancreas insulin infusion system.

[0104] Furthermore, the rPID algorithm can be processed individually or in combination with segmented weighted transformation, relative value transformation, BRGI method, and improved CVGA method. It can flexibly select target blood glucose concentration, zero-risk point blood glucose concentration, or equal-risk point data pairs according to the actual situation, making the rPID algorithm more robust and still having a slow adjustment capability in relatively flat ranges. This allows the closed-loop artificial pancreas to face more complex usage scenarios, thereby achieving more precise blood glucose control.

[0105] Furthermore, the rPID algorithm also compensates for delays in insulin absorption, insulin onset, and the sensing delays in blood glucose and interstitial fluid glucose concentrations, making the output calculated by the rPID algorithm more reliable.

[0106] Furthermore, to compensate for the delayed onset of insulin, the rPID algorithm processes the IOB (Insulin Expected to Bite) for both meal-related insulin and other insulins. This allows for faster insulin clearance during meals and periods of high blood glucose, resulting in greater insulin output and faster blood glucose regulation. When approaching the target level, a longer insulin action time curve is used for more conservative blood glucose regulation.

[0107] Furthermore, the detection module, program module, and infusion module are connected to form a single integrated structure and attached to the same location on the skin. By connecting the three modules into a unified whole and attaching them to the same spot, the number of devices attached to the user's skin is reduced, thus lessening the interference with the user's movements caused by having too many devices. Simultaneously, it effectively solves the problem of poor wireless communication between separate devices, further enhancing the user experience. Attached Figure Description

[0108] Figure 1 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to an embodiment of the present invention;

[0109] Figure 2 This is a comparison diagram of the relationship between blood glucose levels in the risk space and the original physical space, obtained by segmented weighting and relative value conversion in one embodiment of the present invention.

[0110] Figure 3 This is a comparison chart of the relationship between blood glucose levels in the risk space obtained by conversion using the BGRI and CVGA methods according to an embodiment of the present invention and the original physical space.

[0111] Figure 4 An insulin IOB curve according to an embodiment of the present invention;

[0112] Figure 5 This is a schematic diagram illustrating four mainstream clinically optimal baseline rate settings cited in one embodiment of the present invention;

[0113] Figure 6 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to another embodiment of the present invention;

[0114] Figure 7 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to yet another embodiment of the present invention;

[0115] Figure 8 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas multi-drug infusion control system according to another embodiment of the present invention;

[0116] Figure 9 This is a schematic diagram of dual-drug switching according to an embodiment of the present invention;

[0117] Figure 10 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to another embodiment of the present invention. Detailed Implementation

[0118] As mentioned earlier, in existing artificial pancreas technologies, the PID algorithm has a simple structure and few parameters. In order to achieve robustness and prevent overshoot oscillation, the optimization results are generally conservative and the stabilization time is relatively long. It is not suitable for actual situations with large amplitude, high frequency and fast action, such as eating and exercising. At the same time, there is a long time delay in the absorption of food and the action of insulin. The simple PID algorithm cannot achieve the ideal control effect.

[0119] To address this problem, the present invention provides a closed-loop artificial pancreas insulin infusion control system. The system is pre-programmed with an rPID algorithm, which transforms the asymmetric blood glucose in the original physical space into a near-symmetric blood glucose risk in the risk space. This retains the simplicity and robustness of the PID algorithm while also possessing its advantages of precision and flexibility, thus achieving precise control of the closed-loop artificial pancreas insulin infusion system.

[0120] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments should not be construed as limiting the scope of the invention.

[0121] Furthermore, it should be understood that, for ease of description, the dimensions of the various components shown in the accompanying drawings are not necessarily drawn to actual scale; for example, the thickness, width, length, or distance of some units may be enlarged relative to other structures.

[0122] The following description of exemplary embodiments is merely illustrative and is not intended to limit the invention or its application or use in any way. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail herein, but where applicable, such techniques, methods, and apparatus should be considered part of this specification.

[0123] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined or described in a figure, it will not need to be discussed further in the subsequent description of the figures.

[0124] Figure 1 This is a schematic diagram of the module relationships of the closed-loop artificial pancreas insulin infusion control system according to an embodiment of the present invention.

[0125] The closed-loop artificial pancreas insulin infusion control system disclosed in this embodiment of the invention mainly includes a detection module 100, a program module 101, and an infusion module 102.

[0126] The detection module 100 is used to continuously detect the user's current blood glucose level. Generally, the detection module 100 is a continuous glucose monitoring (CGM) device, which can detect the user's current blood glucose level in real time, monitor blood glucose changes, and send the current blood glucose level to the program module 101.

[0127] Program module 101 is used to control the operation of detection module 100 and infusion module 102. Therefore, program module 101 is connected to both detection module 100 and infusion module 102. Here, connection includes conventional electrical connection or wireless connection.

[0128] The infusion module 102 contains the necessary mechanical structures for insulin infusion and is controlled by the program module 101. Based on the current insulin infusion volume data issued by the program module 101, the infusion module 102 infuses the required amount of insulin into the user's body. Simultaneously, the infusion status of the infusion module 102 is also fed back to the program module 101 in real time.

[0129] The embodiments of the present invention do not limit the specific locations and connection relationships of the detection module 100, the program module 101 and the infusion module 102, as long as the aforementioned functional conditions are met.

[0130] In one embodiment of the invention, the three components are electrically connected to form a single integrated structure. Therefore, all three are adhered to the same location on the user's skin. By connecting the three modules into a single unit and adhering them to the same location, the number of devices that can be attached to the user's skin is reduced, thereby lessening the interference to the user's activities caused by having too many devices. Simultaneously, it effectively solves the problem of reliable wireless communication between separate devices, further enhancing the user experience.

[0131] In another embodiment of the invention, program module 101 and infusion module 102 are interconnected to form an integral structure, while detection module 100 is separately disposed in another structure. In this case, detection module 100 and program module 101 transmit wireless signals to each other to achieve interconnection. Therefore, program module 101 and infusion module 102 are attached to one location on the user's skin, while detection module 100 is attached to other locations on the user's skin.

[0132] In another embodiment of the invention, program module 101 and detection module 100 are interconnected to form a single device, while infusion module 102 is disposed separately in another structure. Infusion module 102 and program module 101 transmit wireless signals to each other to achieve interconnection. Therefore, program module 101 and detection module 100 can be attached to one location on the user's skin, while infusion module 102 can be attached to other locations on the user's skin.

[0133] In another embodiment of the invention, the three components are disposed in different structures. Therefore, they are respectively attached to different locations on the user's skin. At this time, the program module 101 transmits wireless signals to the detection module 100 and the infusion module 102 to achieve interconnection.

[0134] It should be noted that the program module 101 in this embodiment of the invention also has functions such as storage, recording, and database access; therefore, the program module 101 can be reused. This not only allows for the storage of user health data but also saves production costs and user operating costs. As mentioned above, when the detection module 100 or the infusion module 102 reaches the end of its lifespan, the program module 101 can be separated from the detection module 100, the infusion module 102, or simultaneously from both.

[0135] Generally, the detection module 100, program module 101, and infusion module 102 have different lifespans. Therefore, when the three are electrically connected to form a single device, they can also be separated into pairs. If one module reaches the end of its lifespan first, the user can replace only that module and retain the other two modules for continued use.

[0136] It should be noted that the program module 101 in this embodiment of the invention may also include multiple sub-modules. Depending on the function of each sub-module, different sub-modules may be set in different structures; no specific limitations are imposed here, as long as the control conditions of the program module 101 are met.

[0137] Specifically, program module 101 is pre-set with an rPID (risk-proportional-integral-derivative) algorithm that converts the asymmetric blood glucose in the original physical space to the approximately symmetric blood glucose risk in the risk space. The rPID algorithm is obtained by conversion processing based on the classic PID (proportional-integral-derivative) algorithm. The specific processing method will be detailed below. According to the corresponding infusion instruction calculated by the rPID algorithm, program module 101 controls infusion module 102 to infuse insulin.

[0138] The classic PID algorithm can be expressed by the following formula:

[0139]

[0140] in:

[0141] K P It is the gain coefficient of the proportional part;

[0142] K I It is the gain coefficient of the integral part;

[0143] K D It is the gain coefficient of the differential part;

[0144] G represents the current blood glucose level;

[0145] G B Indicates the target blood glucose level;

[0146] C represents a constant;

[0147] PID(t) represents the infusion instruction sent to the insulin infusion system.

[0148] Considering the actual distribution characteristics of glucose concentration in diabetic patients, such as the normal blood glucose range of 80-140 mg / dL, which can be relaxed to 70-180 mg / dL, hypoglycemia can reach 20-40 mg / dL, while hyperglycemia can reach 400-600 mg / dL.

[0149] The distribution of high / low blood glucose exhibits significant asymmetry in the original physical space. In clinical practice, the risks of high and low blood glucose corresponding to the same degree of deviation from the normal range can be significantly different. For example, a decrease of 70 mg / dL from 120 mg / dL to 50 mg / dL is considered severe hypoglycemia, carrying a high clinical risk and requiring emergency measures such as carbohydrate supplementation. On the other hand, an increase of 70 mg / dL from 120 mg / dL to 190 mg / dL is only slightly above the normal range. For diabetic patients, this degree of hyperglycemia is not serious and is frequently reached in daily life, generally requiring no intervention.

[0150] To address the asymmetric nature of clinical risk associated with glucose concentration, the algorithm transforms the asymmetric blood glucose levels in the original physical space into approximately symmetrical blood glucose risk in the risk space, making the PID algorithm more robust.

[0151] Accordingly, the rPID algorithm formula is transformed into the following form:

[0152]

[0153] in:

[0154] rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion;

[0155] r indicates blood sugar risk;

[0156] The meanings of the other symbols are as described above.

[0157] To maintain the stability of the PID integral, and considering the physiological effect of insulin in lowering blood glucose, in one embodiment of the present invention, the input parameter of the PID—the blood glucose deviation Ge = GG B Processing is required, such as setting Ge=GG B Perform segmented weighted processing as follows:

[0158]

[0159] In another embodiment of the invention, for blood glucose levels greater than the target G... B The deviation is converted using relative values, as follows:

[0160]

[0161] Figure 2 This is a comparison diagram of the blood glucose risk space obtained through segmented weighting and relative value transformation with the blood glucose relationship in the original physical space.

[0162] In the original PID algorithm, the blood glucose risk (Ge) on both sides of the target blood glucose value exhibits a severe asymmetry consistent with the original physical space. After being transformed into the blood glucose risk space, the blood glucose risk on both sides of the target blood glucose value is approximately symmetrical, so the integral term can remain stable, making the rPID algorithm more robust.

[0163] In another embodiment of the invention, a fixed zero-risk point exists during risk transformation, and data deviating from the zero-risk point are processed. Original parameters corresponding to values ​​greater than the zero-risk point are positive when transformed into the risk space, while original parameters corresponding to values ​​less than the zero-risk point are negative. Specifically, the classic Blood Glucose Risk Index (BGRI) method can be used as a reference. This method, based on clinical practice, considers hypoglycemia at 20 mg / dL and hyperglycemia at 600 mg / dL to have comparable clinical risk, and uses logarithmic transformation to comprehensively process blood glucose levels within the 20-600 mg / dL range. The blood glucose value corresponding to the zero-risk point of this method is set as the target blood glucose value G. B The risk space transformation formula is as follows:

[0164]

[0165] in:

[0166] r(G) = 10 * f(G) 2

[0167] The transformation function f(G) is as follows:

[0168] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0169] In the classic glycemic risk index method, the blood glucose value corresponding to the zero-risk point is 112 mg / dL. In other embodiments of the present invention, the blood glucose value at the zero-risk point can also be adjusted based on the risk and data trends in clinical practice, and is not specifically limited here. A risk space for blood glucose values ​​greater than the zero-risk point is fitted, and the specific fitting method is also not specifically limited.

[0170] In another embodiment of the present invention, an improved Control Variability Grid Analysis (CVGA) method is used. The original CVGA defined the zero-risk point blood glucose value as 110 mg / dL and assumed the following equally risky blood glucose value pairs (90 mg / dL, 180 mg / dL; 70 mg / dL, 300 mg / dL; 50 mg / dL, 400 mg / dL). In this embodiment of the present invention, considering the actual risk in clinical practice and the trend characteristics of the data, adjustments are made. The equally risky data pair (70 mg / dL, 300 mg / dL) is corrected to (70 mg / dL, 250 mg / dL), and the zero-risk point blood glucose value is set as the target blood glucose value G. B A multinomial model was then fitted to it, yielding the following risk functions for processing both sides of the zero-risk point:

[0171]

[0172] And its maximum value was limited:

[0173] |r| = min(|r|, n)

[0174] The maximum value n is limited to a range of 0 to 80 mg / dL, and preferably, the value of n is 60 mg / dL.

[0175] In other embodiments of the present invention, the zero-risk point blood glucose value and the equivalent risk data can also be adjusted in combination with the actual risk and data trend in clinical practice, without being specifically limited here. Then, the equivalent risk point is fitted, and the specific fitting method is not specifically limited; the specific value used to limit the maximum value is also not specifically limited.

[0176] Figure 3 This is a comparison chart showing the relationship between blood glucose risk converted to the risk space using the BGRI and CVGA methods and blood glucose levels in the original physical space.

[0177] Similar to Zone-MPC, within the normal blood glucose range, the blood glucose risk converted by BGRI and CVGA methods is quite mild, especially in the 80-140 mg / dL range. Unlike Zone-MPC, which has a risk of zero in this range and loses the ability to be further optimized, rPID, while having a mild risk in this range, still has a stable and slow adjustment ability, allowing blood glucose to be further adjusted towards the target value, achieving more precise blood glucose control.

[0178] In another embodiment of the present invention, data deviating from the zero-risk point can be processed using a uniform method, as in the aforementioned embodiments, where data deviating from the zero-risk point can be processed using either the BGRI or CVGA method; alternatively, different processing methods can be used, such as combining the BGRI and CVGA methods simultaneously. In this case, the same zero-risk point blood glucose value, such as the target blood glucose value G, can be used. B When the blood glucose level is lower than the target blood glucose level (G) B When using the BGRI method, the blood glucose level is greater than the target blood glucose level (G). B When using the CVGA method, at this time:

[0179] r = -r(G), if G ≤ G B

[0180] in:

[0181] r(G) = 10 * f(G) 2

[0182] The transformation function f(G) is as follows:

[0183] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0184] r = -4.8265 * 10 4 -4*G 2 +0.45563*G-44.855, if G>G B .

[0185] Similarly, this can also be done when the blood glucose level is lower than the target blood glucose level (G). B When using the CVGA method, the blood glucose level is greater than the target blood glucose level (G). B When using the BGRI method, at this time:

[0186] r = r(G), if G > G B

[0187] in:

[0188] r(G) = 10 * f(G) 2

[0189] Under the transformation function f(G):

[0190] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0191] r = GG B if G≤G B .

[0192] It also allows setting a limit on the maximum value:

[0193] |r| = min(|r|, n)

[0194] The maximum value n is limited to a range of 0 to 80 mg / dL, and preferably, the value of n is 60 mg / dL.

[0195] In other embodiments of the present invention, the blood glucose value at the zero-risk point can also be set as the target blood glucose value G. B For blood glucose levels less than or equal to the target blood glucose value (G) B The data were obtained using the BGRI method, and for values ​​greater than the target blood glucose level (G... B The data is processed using deviation methods, such as segmented weighting or relative value processing.

[0196] When segmented weighted processing is used, then:

[0197] r = -r(G), if G ≤ G B

[0198] in:

[0199] r(G) = 10 * f(G) 2

[0200] Under the transformation function f(G):

[0201] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0202]

[0203] When using relative values:

[0204] r = -r(G), if G ≤ G B

[0205] in:

[0206] r(G) = 10 * f(G) 2

[0207] Under the fitted symmetric transformation function f(G):

[0208] f(G) = 1.509 * [(ln(G)) 1.084 -5.381]

[0209] r = 100 * (GG B ) / G, if G>G B

[0210] When the blood glucose value corresponding to the zero risk point is the target blood glucose value (G) BWhen, for blood glucose levels less than or equal to the target blood glucose value G B The processing function remains consistent when using segmented weighting, relative value processing, and CVGA methods for data. Therefore, when processing data less than or equal to the target blood glucose value G... B The data is processed using segmented weighting or relative value processing. When the BGRI method is used for data with blood glucose values ​​greater than the zero risk point, the processing result is equivalent to the aforementioned processing when the blood glucose value is less than or equal to the target blood glucose value G. B When using the CVGA method, the blood glucose level is greater than the target blood glucose level (G). B The BGRI method is used, and its calculation formula will not be repeated here.

[0211] It should be noted that, in the various embodiments of the present invention, the target blood glucose value G... B The target blood glucose level is 80–140 mg / dL, preferably G. B The concentration is 110–120 mg / dL.

[0212] The above processing methods can transform the asymmetric blood glucose in the original physical space into an approximately symmetric blood glucose risk in the risk space. This retains the simplicity and robustness of the PID algorithm while also providing targeted and clinically valuable blood glucose risk control, thus achieving precise control of the closed-loop artificial pancreatic insulin infusion system.

[0213] Three major delays exist in closed-loop artificial pancreas control systems: delayed insulin absorption (approximately 20 minutes from subcutaneous tissue to the bloodstream, and approximately 100 minutes to the liver), delayed insulin onset of action (approximately 30-100 minutes), and delayed sensing of interstitial fluid glucose concentration and blood glucose (approximately 5-15 minutes). Any attempt to accelerate closed-loop responsiveness may lead to unstable system behavior and system oscillations. To compensate for the delayed insulin absorption in the closed-loop artificial pancreas control system, an insulin feedback compensation mechanism is introduced in one embodiment of the invention. The amount of insulin not yet absorbed in the body is subtracted from the output, resulting in a component proportional to the estimated plasma insulin concentration. (In reality, human insulin secretion also uses the insulin concentration in the blood plasma as a negative feedback regulatory signal). The formula is as follows:

[0214]

[0215] in:

[0216] PID(t) represents the infusion instruction sent to the insulin infusion system;

[0217] PID c (t) indicates a compensated infusion instruction sent to the insulin infusion system;

[0218] γ represents the compensation coefficient of the estimated plasma insulin concentration to the algorithm output. A larger coefficient will make the algorithm more conservative, while a smaller coefficient will make it more aggressive. Therefore, in the embodiments of the present invention, the range of γ is 0.4-0.6, and preferably, γ is 0.5.

[0219] Estimates of plasma insulin concentration can be obtained using various conventional prediction algorithms, such as direct calculation from the infused insulin based on the insulin pharmacokinetic curve, or by using conventional autoregressive methods.

[0220]

[0221] in:

[0222] This represents an estimate of the current plasma insulin concentration.

[0223] PID c (n-1) represents the compensated output of the previous time step;

[0224] This represents an estimate of the plasma insulin concentration at the previous moment.

[0225] This represents an estimate of the plasma insulin concentration at the previous time step.

[0226] K0 represents the coefficient of the compensated output part in the previous time step;

[0227] K1 represents the coefficient of the estimated portion of the plasma insulin concentration at the previous moment;

[0228] K2 represents the coefficient of the estimated plasma insulin concentration at the previous time step;

[0229] Wherein, initial value The time intervals can be selected according to actual needs.

[0230] Accordingly, the compensation output formula after risk conversion using the aforementioned method is as follows:

[0231]

[0232] in:

[0233] rPIDc(t) represents the compensated infusion instruction sent to the insulin infusion system after risk conversion;

[0234] rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion;

[0235] The meanings of the other characters are as described above.

[0236] To compensate for the delayed onset of insulin in the closed-loop artificial pancreas control system, in one embodiment of the present invention, insulin IOB (insulin on board) that has not yet taken effect in the body is introduced. IOB is deducted from the insulin output to prevent insulin accumulation and overdose, which could lead to risks such as postprandial hypoglycemia.

[0237] Figure 4 This is the insulin IOB curve according to an embodiment of the present invention.

[0238] according to Figure 4 The IOB curve shown can be used to calculate the cumulative residual amount of previously infused insulin. The specific curve can be selected based on the user's actual insulin action time.

[0239] PID′(t) = PID(t) - IOB(t)

[0240] in:

[0241] PID'(t) represents the infusion instruction sent to the insulin infusion system after deducting IOB;

[0242] PID(t) represents the infusion instruction sent to the insulin infusion system;

[0243] IOB(t) represents the amount of insulin in the body that has not yet taken effect at time t.

[0244] Accordingly, the output formula after risk conversion using the aforementioned method, minus the amount of insulin that has not yet taken effect in the body, is as follows:

[0245] rPID′(t)=rPID(t)-IOB(t)

[0246] in:

[0247] rPID′(t) represents the infusion instruction sent to the insulin infusion system after risk conversion, excluding the amount of insulin that has not yet taken effect in the body;

[0248] rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion;

[0249] The meanings of the other characters are as described above.

[0250] To achieve a more ideal control effect, the calculation of IOB was processed as follows: IOB m IOB o These correspond to the IOBs for mealtime insulin and other insulins (excluding mealtime insulin), respectively. The formulas are as follows:

[0251] IOB(t) = IOB m，t +IOB o，t

[0252] in:

[0253]

[0254] in:

[0255] IOB m，t This represents the amount of insulin consumed after a meal that has not yet taken effect in the body at time t.

[0256] IOB o，t This represents the amount of non-meal insulin that has not yet taken effect in the body at time t;

[0257] D i (i = 2 - 8) represent the corresponding coefficients of the IOB curves for insulin action time i;

[0258] I m，t Indicates the amount of insulin consumed after a meal;

[0259] I 0，t This indicates the amount of insulin not consumed during meals;

[0260] IOB(t) represents the amount of insulin in the body at time t that has not yet taken effect.

[0261] Differentiating between mealtime and non-mealtime insulin administration for IOB allows for faster insulin clearance during meals and periods of high blood glucose, resulting in greater insulin output and quicker blood glucose regulation. Conversely, when approaching the target level, a longer insulin action time curve is used, leading to slower insulin clearance and more conservative and stable blood glucose regulation.

[0262] When PID'(t) > 0 or rPID'(t) > 0, the final amount of insulin infused is PID'(t) or rPID'(t);

[0263] When PID'(t) < 0 or rPID'(t) < 0, the final amount of insulin infused is 0.

[0264] To compensate for the sensing delay of tissue fluid glucose concentration and blood glucose in the closed-loop artificial pancreas control system, an autoregressive method is used in one embodiment of the present invention, as shown in the following formula:

[0265]

[0266] in,

[0267] G SC (n) represents the glucose concentration in the interstitial fluid at the current moment, i.e., the measurement value of the sensing system;

[0268] This indicates the estimated blood glucose concentration at the previous moment;

[0269] G SC (n-1) and G SC (n-2) represent the glucose concentrations in the interstitial fluid at the previous and previous time points, respectively;

[0270] K0 represents the coefficient of the estimated blood glucose concentration at the previous moment;

[0271] K1 and K2 represent the coefficients of interstitial fluid glucose concentration at the previous and previous time points, respectively.

[0272] At the initial moment,

[0273] By estimating blood glucose concentration using interstitial fluid glucose concentration, the sensing delay between interstitial fluid glucose concentration and blood glucose is compensated, making the PID algorithm more accurate. Correspondingly, the rPID algorithm can also more accurately calculate the actual insulin requirement of the human body.

[0274] In this embodiment of the invention, the delays in insulin absorption, insulin onset, tissue fluid glucose concentration, and blood glucose sensing can be partially or fully compensated. Preferably, all delay factors are considered and fully compensated to make the rPID algorithm more accurate.

[0275] In another embodiment of the present invention, the program module 101 is pre-set with an rMPC (Risk-Model-Prediction-Control) algorithm that converts blood glucose that is asymmetric in the original physical space into blood glucose risk that is approximately symmetric in the risk space. The rMPC algorithm is obtained by conversion processing based on the classic MPC (Model-Prediction-Control) algorithm. According to the corresponding infusion instruction calculated by the rMPC algorithm, the program module 101 controls the infusion module 102 to infuse insulin.

[0276] The classic MPC algorithm consists of three elements: a prediction model, a value function, and constraints. The classic MPC prediction model is as follows:

[0277] x t+1 =Ax t +BI t

[0278] G t =Cx t

[0279] in:

[0280] x t+1 Indicates the state parameters at the next moment.

[0281] xt This represents the state parameters at the current moment.

[0282] I t This indicates the current insulin infusion volume;

[0283] G t This indicates the blood glucose concentration at the current moment.

[0284] The parameter matrix is ​​as follows:

[0285]

[0286]

[0287] C = [1 0 0]

[0288] b1, b2, b3, and K are prior values.

[0289] The value function of MPC consists of the sum of squares of the deviations in the output G (blood glucose level) and the sum of squares of the changes in the input I (insulin level). MPC aims to find the minimum solution to the value function.

[0290]

[0291] in:

[0292] I' t+j This indicates the change in insulin infusion volume after step j;

[0293] This represents the difference between the predicted blood glucose concentration and the target blood glucose value after step j;

[0294] t represents the current time;

[0295] N and P are the number of steps within the control time window and the prediction time window, respectively;

[0296] R is the weighting coefficient for the insulin component.

[0297] The insulin infusion volume in step j is I. t +I′ t+j .

[0298] In this embodiment of the invention, the control time window T c =30min, prediction time window T p =60min, the weighting coefficient R of insulin amount is 11000. It should be noted that although the control time window used in the calculation is 30min, only the first step calculation result of insulin output is used in actual operation. After running, the minimum solution of the above value function is recalculated based on the latest blood glucose value.

[0299] In this embodiment of the invention, the infusion time step within the control time window is j. n j n The value range is 0–30 min, preferably 2 min. Number of steps N = T c / j n The range of j is from 0 to N.

[0300] In other embodiments of the present invention, the weighting coefficients for the control time window, prediction time window, and insulin amount can be selected as other values, which are not specifically limited here.

[0301] As mentioned earlier, due to the significant asymmetry in the distribution of high / low blood glucose (original physical space), the risks of high and low blood glucose corresponding to the same degree of deviation from the normal range in clinical practice will be significantly different. To address the asymmetry in clinical risk related to glucose concentration, the asymmetric blood glucose risk in the original physical space is transformed into a nearly symmetrical blood glucose risk in the risk space, making the MPC algorithm more accurate and flexible. The value function of the rMPC algorithm after risk transformation is as follows:

[0302]

[0303] in,

[0304] r t+j This represents the blood glucose risk value after step j;

[0305] I' t+j This indicates the change in insulin infusion volume after step j.

[0306] The deviation of blood glucose values ​​is converted into corresponding blood glucose risk, and the specific conversion method is consistent with that in the aforementioned rPID algorithm, such as segmented weighting and relative value processing. It also includes setting a fixed zero-risk point in the risk space, where the blood glucose concentration can be set as the target blood glucose value. Data deviating from the zero-risk point on both sides are processed, such as using BGRI and improved CVGA methods; it also includes using different methods to process data deviating from the target blood glucose value on both sides.

[0307] Specifically, when using segmented weighted processing:

[0308]

[0309] When using relative values:

[0310]

[0311] When using the classic glycemic risk index method:

[0312]

[0313] in:

[0314] r(G t+j )=10*f(G t+j ) 2

[0315] Transformation function f(G) t+j )as follows:

[0316] f(G t+j )=1.509*[(ln(G t+j )) 1.084 -5.381]

[0317] When using a controlled volatile grid analysis method:

[0318]

[0319] At the same time, its maximum value was also limited:

[0320] |r t+j |=min(|r t+j |,n)

[0321] The maximum value n is limited to a range of 0 to 80 mg / dL, and preferably, the value of n is 60 mg / d.

[0322] When blood glucose level is lower than the target blood glucose level (G) B When using the BGRI method, the blood glucose level is greater than the target blood glucose level (G). B When using the CVGA method:

[0323] r t+j =-r(G t+j ), if G t+j ≤G B

[0324] in:

[0325] r(G t+j )=10*f(G t+j ) 2

[0326] Transformation function f(G) t+j )as follows:

[0327] f(G t+j )=1.509*[(ln(G t+j )) 1.084 -5.381]

[0328] r t+j = -4.8265 * 10 4 -4*G t+j2 +0.45563*G t+j -44.855, if G t+j >G B

[0329] When blood glucose levels are lower than the target blood glucose level (G) B When using the CVGA method, the blood glucose level is greater than the target blood glucose level (G). B When using the BGRI method:

[0330] r t+j =r(G t+j ), if G t+j >G B ,

[0331] in:

[0332] r(G t+j )=10*f(G t+j ) 2

[0333] Transformation function f(G) t+j )as follows:

[0334] f(G t+j )=1.509*[(ln(G t+j )) 1.084 -5.381]

[0335] r t+j =G t+j -G B , if G t+j ≤G B .

[0336] It also allows setting a limit on the maximum value:

[0337] |r t+j |=min(|r t+j |,n)

[0338] The maximum value n is limited to a range of 0 to 80 mg / dL, and preferably, the value of n is 60 mg / dL.

[0339] When blood glucose level is lower than the target blood glucose level (G) B When using the BGRI method, the blood glucose level is greater than the target blood glucose level (G). B When using a segmented weighted method:

[0340] r t+j =-r(G t+j ), if G t+j ≤G B

[0341] in:

[0342] r(G t+j )=10*f(G t+j ) 2

[0343] Transformation function f(G) t+j )as follows:

[0344] f(G t+j )=1.509*[(ln(G t+j )) 1.084 -5.381]

[0345]

[0346] When blood glucose level is lower than the target blood glucose level (G) B When using the BGRI method, the blood glucose level is greater than the target blood glucose level (G). B When using relative value conversion:

[0347] r t+j =-r(G t+j ), if G t+j ≤G B

[0348] in:

[0349] r(G t+j )=10*f(G t+j ) 2

[0350] Transformation function f(G) t+j )as follows:

[0351] f(G t+j )=1.509*[(ln(G t+j )) 1.084 -5.381]

[0352]

[0353] When the target blood glucose level is less than or equal to G B The data is processed using segmented weighting or relative value processing. When the BGRI method is used for data with blood glucose values ​​greater than the zero risk point, the processing result is equivalent to the aforementioned processing when the blood glucose value is less than or equal to the target blood glucose value G. B When using the CVGA method, the blood glucose level is greater than the target blood glucose level (G). B The BGRI method is used, and its calculation formula will not be repeated here.

[0354] It should be noted that in the above conversion formulas:

[0355] r t+jThis represents the blood glucose risk value at step j.

[0356] G t+j This represents the blood glucose level detected at step j.

[0357] Target blood glucose level G B The target blood glucose level is 80–140 mg / dL, preferably G. B The concentration is 110–120 mg / dL.

[0358] The beneficial effects after risk conversion and the comparison of the relationship between blood glucose and blood glucose risk are consistent with those in the rPID algorithm, and will not be repeated here.

[0359] Similarly, insulin feedback compensation can be used to compensate for insulin uptake delay; IOB compensation can be used to compensate for insulin onset delay; and autoregressive compensation can be used to compensate for the sensing delay of tissue fluid glucose concentration and blood glucose concentration. The specific compensation methods are consistent with those in the rPID algorithm.

[0360] The compensation formula for delayed insulin uptake is as follows:

[0361]

[0362] in:

[0363] I t+j This indicates the infusion instruction sent to the insulin infusion system at step j;

[0364] rI c(t+j) This indicates the infusion instruction sent to the insulin infusion system at step j after risk conversion;

[0365] γ represents the compensation coefficient of the estimated plasma insulin concentration to the algorithm output. A larger coefficient will make the algorithm more conservative, while a smaller coefficient will make it more aggressive. Therefore, in the embodiments of the present invention, the range of γ is 0.4-0.6, and preferably, γ is 0.5. This represents the estimated plasma insulin concentration at step j.

[0366] For delayed insulin onset, the compensation formula is as follows:

[0367] rI′ t+j =rI t+j -IOB(t+j)

[0368] in:

[0369] rI′ t+j This indicates the infusion instruction sent to the insulin infusion system after deducting IOB at step j following risk conversion;

[0370] rI t+jThis indicates the infusion instruction sent to the insulin infusion system at step j after risk conversion;

[0371] IOB(t+j) represents the amount of insulin in the body that has not yet taken effect at time t+j.

[0372] Similarly, IOB(t+j) can be distinguished between being in a meal and not being in a meal, in which case:

[0373] IOB(t+j)=IOB m，t+j +IOB o，t+j

[0374] in:

[0375]

[0376] in:

[0377] IOB m，t+j This represents the amount of insulin consumed after a meal that has not yet taken effect in the body at time t+j.

[0378] IOB o，t+j This represents the amount of non-meal insulin that has not yet taken effect in the body at time t+j;

[0379] D i (i = 2 - 8) represent the corresponding coefficients of the IOB curves for insulin action time i;

[0380] I m，t+j This represents the insulin level at time t+j after the meal.

[0381] I 0，t+j This represents the non-meal insulin amount at time t+j;

[0382] IOB(t+j) represents the amount of insulin in the body at time t+j that has not yet taken effect.

[0383] When rI′ t+j When >0, the final infused insulin volume is rI′. t+j ;

[0384] When rI′ t+j When <0, the final amount of insulin infused is 0.

[0385] For the sensing delay of tissue fluid glucose concentration and blood glucose concentration, autoregressive compensation can also be used, as shown in the following formula:

[0386]

[0387] in,

[0388] G SC(t+j) represents the glucose concentration in the interstitial fluid at time t+j, which is the measured value of the sensing system;

[0389] This represents the estimated blood glucose concentration at time t+j-1;

[0390] G SC (t+j-1) and G SC (t+j-2) represent the glucose concentrations in the interstitial fluid at times t+j-1 and t+j-2, respectively;

[0391] K0 represents the coefficient of the estimated blood glucose concentration at time t+j-1;

[0392] K1 and K2 represent the coefficients of interstitial fluid glucose concentration at time t+j-1 and t+j-2, respectively.

[0393] At the initial moment,

[0394] The beneficial effects of various compensation methods are consistent with those of the rPID algorithm, and will not be repeated here.

[0395] It should be noted that, in the rMPC algorithm, it is preferable to compensate for the delay in insulin onset and the sensing delay in tissue fluid glucose concentration and blood glucose concentration.

[0396] In another embodiment of the present invention, the program module 101 is pre-set with a composite artificial pancreas algorithm, which includes a first algorithm and a second algorithm. When the detection module 100 detects the current blood glucose value and sends the current blood glucose value to the program module 101, the first algorithm calculates the first insulin infusion volume I1, the second algorithm calculates the second insulin infusion volume I2, the composite artificial pancreas algorithm optimizes the calculation of the first insulin infusion volume I1 and the second insulin infusion volume I2 to obtain the final insulin infusion volume I3, and sends the final insulin infusion volume I3 to the infusion module 102. The infusion module 102 performs insulin infusion according to the final infusion volume I3.

[0397] The first and second algorithms are one of the classic PID algorithm, the classic MPC algorithm, the rMPC algorithm, or the rPID algorithm. The rMPC algorithm or the rPID algorithm is an algorithm that transforms blood glucose levels that are asymmetric in the original physical space into blood glucose risk levels that are approximately symmetric in the risk space. The transformation method of blood glucose risk in the rMPC algorithm and the rPID algorithm is as described above.

[0398] When I1 = I2, I3 = I1 = I2;

[0399] When I1 ≠ I2, the arithmetic mean of I1 and I2 can be substituted into the first and second algorithms respectively to re-optimize the algorithm parameters. After parameter optimization, the required insulin infusion volume at the current time is calculated again using the first and second algorithms respectively. If I1 and I2 are still different, the arithmetic mean of I1 and I2 is taken again and the above process is repeated until I1 and I2 are the same, that is:

[0400] ① Calculate the average value of the first insulin infusion volume I1 and the second insulin infusion volume I2.

[0401] ② Average value Substitute these parameters into the first and second algorithms respectively, and adjust the algorithm parameters accordingly;

[0402] ③ Based on the current blood glucose level, the first algorithm and the second algorithm after parameter adjustment, recalculate the first insulin infusion volume I1 and the second insulin infusion volume I2;

[0403] ④ Repeat steps ① to ③ until I1 = I2, and the final insulin infusion volume I3 = I1 = I2.

[0404] At this point, when the first or second algorithm is a PID or rPID algorithm, the algorithm parameter is KP, and K D =T D / K P T D You can take 60-90 minutes, K I =T I *K P T I The timeframe can be 150min-450min. When the first or second algorithm is an MPC or rPMC algorithm, the algorithm parameter is K.

[0405] When I1 ≠ I2, I1 and I2 can be weighted, and the weighted values ​​can be substituted into the first and second algorithms to re-optimize the algorithm parameters. After parameter optimization, the required insulin infusion volume at the current time can be calculated again using the first and second algorithms. If I1 and I2 are still different, I1 and I2 can be weighted again, the weighting coefficients adjusted, and the above process repeated until I1 and I2 are the same.

[0406] ① Calculate the weighted average of the first insulin infusion volume I1 and the second insulin infusion volume I2. Where α and β are the weighting coefficients of the first insulin infusion volume I1 and the second insulin infusion volume I2, respectively;

[0407] ②The weighted average Substitute these parameters into the first and second algorithms and adjust the algorithm parameters accordingly.

[0408] ③ Based on the current blood glucose level, the first algorithm and the second algorithm after parameter adjustment, recalculate the first insulin infusion volume I1 and the second insulin infusion volume I2;

[0409] ④ Repeat steps ① to ③ until I1 = I2, and the final insulin infusion volume I3 = I1 = I2.

[0410] Similarly, when the first or second algorithm is a PID or RPID algorithm, the algorithm parameter is K. P And K D =T D / K P T D You can take 60-90 minutes, K I =T I *K P T I The timeframe can be 150min-450min. When the first or second algorithm is an MPC or rPMC algorithm, the algorithm parameter is K.

[0411] In this embodiment of the invention, α and β can be adjusted according to the magnitude of the first insulin infusion volume I1 and the second insulin infusion volume I2. When I1≥I2, α≤β; when I1≤I2, α≥β; preferably, α+β=1. In other embodiments of the invention, α and β can also be other ranges of values, which are not specifically limited here.

[0412] When the calculation results of the two algorithms are the same, i.e., I3 = I1 = I2, it can be considered that the insulin infusion volume at the current moment can bring the blood glucose level to the ideal level. Through the above processing, the algorithms refer to each other. Preferably, the first algorithm and the second algorithm are the rMPC algorithm and the rPID algorithm, respectively. The two refer to each other to further improve the accuracy of the output results and make the results more feasible and reliable.

[0413] In another embodiment of the present invention, the program module 101 is further provided with a memory for storing information such as the user's historical physical state, blood glucose level, and insulin infusion volume. Statistical analysis can be performed based on the information in the memory to obtain the current statistical analysis result I4. When I1 ≠ I2, I1, I2, and I4 are compared respectively to calculate the final insulin infusion volume I3. The one between I1 and I2 that is closer to the statistical analysis result I4 is selected as the final calculation result of the composite artificial pancreas algorithm, i.e., the final insulin infusion volume I3. The program module 101 sends the final insulin infusion volume I3 to the infusion device 102 for infusion; that is:

[0414]

[0415] By comparing with historical data, the reliability of insulin infusion volume is ensured from another perspective.

[0416] In another embodiment of the present invention, when I1 and I2 are inconsistent and have a large difference, they can be adjusted to be similar by changing the blood glucose risk space transformation method and / or the compensation method for the delay effect in the rMPC algorithm and / or rPID algorithm. Then, the output result of the composite artificial pancreas algorithm is finally determined by the above arithmetic mean, weighted processing, or comparison with the statistical analysis results.

[0417] In another embodiment of the present invention, the closed-loop artificial pancreas control system further includes a meal recognition module and an exercise recognition module. These modules are used to identify whether the user is eating or exercising. A common method for meal recognition is based on the rate of change in blood glucose levels, determined by a specific threshold. The rate of change in blood glucose levels can be calculated from two consecutive moments or obtained through linear regression over a period of time. Specifically, when using the rate of change from two consecutive moments, the calculation formula is:

[0418] dG t / dt=(G t -G t-1 ) / Δt

[0419] in:

[0420] G t This indicates the current blood glucose level;

[0421] G t-1 This indicates the blood glucose level at the previous moment;

[0422] Δt represents the time interval between the current moment and the previous moment.

[0423] When using the formula for calculating the rate of change at three points, the formula is:

[0424] dG t / dt=(3G t -4G t-1 +G t-2 ) / 2Δt

[0425] in:

[0426] G t This indicates the current blood glucose level;

[0427] G t-1 This indicates the blood glucose level at the previous moment;

[0428] G t-2 This indicates the blood glucose level at the time two hours ago;

[0429] Δt represents the time interval between the current moment and the previous moment.

[0430] Before calculating the rate of change in blood glucose, the raw continuous glucose data can be filtered or smoothed. The threshold can be set between 1.8 mg / mL and 3 mg / mL, or it can be customized.

[0431] Similar to meal recognition, since exercise causes a rapid drop in blood glucose, exercise recognition can also be based on the rate of blood glucose change and determined using a specific threshold. The calculation of the rate of blood glucose change can also be performed as described previously, and the threshold can be customized. To more quickly determine the occurrence of exercise, the closed-loop artificial pancreas insulin infusion control system also includes a motion sensor (not shown). The motion sensor is used to automatically detect the user's physical activity, and the program module 101 can receive information about the user's physical activity status. The motion sensor can automatically and accurately sense the user's physical activity state and send the activity status parameters to the program module 101, improving the reliability of the composite artificial pancreas algorithm's output in exercise scenarios.

[0432] The motion sensor can be disposed in the detection module 100, the program module 101, or the infusion module 102. Preferably, in this embodiment of the invention, the motion sensor is disposed in the program module 101.

[0433] It should be noted that the embodiments of the present invention do not limit the number of motion sensors or the placement of multiple motion sensors, as long as the conditions for motion sensors to sense user activity are met.

[0434] The motion sensor includes a three-axis accelerometer or a gyroscope. A three-axis accelerometer or gyroscope can more accurately sense the intensity, pattern, or posture of body activity. Preferably, in this embodiment of the invention, the motion sensor is a combination of a three-axis accelerometer and a gyroscope.

[0435] It should be noted that during the calculation process, the blood glucose risk conversion methods used by the rMPC algorithm and the rPID algorithm can be the same or different, and the compensation methods for the delay effect can also be the same or different. Adjustments can also be made during the calculation process according to the actual situation.

[0436] In another embodiment of the invention, the program module 101 is further provided with an adaptive unit that adjusts the algorithm gain coefficient according to the user's weight. In some embodiments of the invention, the infusion module 102 or the program module 101 can indicate the user's total daily insulin requirement (DIR). In another embodiment of the invention, DIR can be calculated from body weight (BW), specifically, DIR is proportional to BW, i.e., DIR = e * BW, where e is the weight adjustment coefficient.

[0437] For patients with type 1 diabetes, the weight adjustment factor (e) can be chosen from the population mean of 0.53 U / kg. It can also be personalized based on their exercise habits; for example, a lower weight adjustment factor, such as 0.4 U / kg, can be chosen for professional-level exercisers, while a higher weight adjustment factor, such as 0.6 U / kg, can be chosen for those who exercise less. For patients with type 2 diabetes, a personalized weight adjustment factor can be chosen within a wider range, such as 0.1-1.5 U / kg, with 0.6-1.1 U / kg being the most commonly used range, considering their pancreatic secretory function and insulin resistance.

[0438] In one embodiment of the present invention, the algorithm preset in the program module 101 is a classic PID algorithm or an rPID algorithm, and the gain coefficient of its proportional part is Kp = DIR / (BW*m), where m is the user weight compensation coefficient, and the value is 50 to 500. Preferably, m is 135.

[0439] The gain coefficient KI of the integral part and the gain coefficient K of the derivative part in the PID algorithm or rPID algorithm. D All can be converted into coefficients related to Kp, such as K D =T D / K P T D You can take 60-90 minutes, K I =T I *K P T I The time can be 150-450 minutes. (T) D T I Larger values ​​result in a more aggressive algorithm, while smaller values ​​result in a more conservative one. Different coefficient settings can be used for daytime and nighttime sleep; for example, a smaller time parameter can be selected for nighttime.

[0440] In another embodiment of the present invention, the algorithm preset in program module 101 is a classic MPC algorithm or an rMPC algorithm, and its gain coefficient K is related to BW:

[0441]

[0442] in:

[0443] c is the safety factor;

[0444] s is the clinical experience coefficient;

[0445] e is the weight adjustment factor, U / Kg.

[0446] Based on the risk of nocturnal hypoglycemia, the safety factor c can be selected between 1.25 and 3; the clinical experience factor s can be 1500, 1700, 1800, 2000, 2200, 2500, etc., and can be adjusted according to clinical results, without specific limitations. In a preferred embodiment of the present invention, the clinical experience factor s is 1700. The range of the weight adjustment factor e is as described above.

[0447] In the two embodiments described above, the gain coefficient Kp of the PID algorithm or rPID algorithm and the gain coefficient K of the MPC algorithm or rMPC algorithm can also be adjusted by introducing a coefficient Sb(t) related to the basal insulin requirement. Correspondingly, K′ P =K P *Sb(t), K′=K*Sb(t).

[0448] The coefficient Sb(t) related to basal insulin requirements is the ratio of the basal insulin requirement B(t) at time t to the average daily basal insulin intake Ba, i.e., Sb(t) = B(t) / Ba. Here, Ba = y * DIR / 24, where y is the basal insulin compensation coefficient, ranging from 0.1 to 5. The average value of this coefficient for the general population is 0.47, and it is slightly lower for children, for example, 0.3-0.4.

[0449] The average daily basal insulin dose, Ba, can be calculated based on the user's actual basal rate setting. The basal insulin requirement at time t, B(t), can be set according to one of the four mainstream clinically optimal basal rate settings. Figure 5 There are four types of optimal basal insulin rates for mainstream clinical settings, which are derived from the reference [Holterhus, PM, J. Bokelmann, et al. (2013). "Predicting the Optimal Basal Insulin Infusion Pattern in Children and Adolescents on Insulin Pumps." Diabetes Care 36(6): 1507-1511.]. The horizontal axis represents time, 24 hours a day, and the vertical axis represents the relative deviation between the basal insulin requirement at the corresponding time and the average basal insulin amount Ba throughout the day, mostly between 0.5 and 1.5.

[0450] B(t) can also refer to the clinically commonly used baseline rate segmentation settings, such as a three-segment setting, as follows:

[0451] ① When time t is from 0:00 to 4:00 AM, B(t) = 0.5DIR / 48;

[0452] ②When time t is from 4 a.m. to 10 a.m., B(t) = 1.5DIR / 48;

[0453] ③ When time t is from 10:00 AM to 0:00 AM, B(t) = DIR / 48.

[0454] In other embodiments of the invention, B(t) can also be calculated based on a base rate setting that is known and suitable to the user.

[0455] In this embodiment of the invention, Sb(t) ranges from 0.2 to 2, preferably from 0.5 to 1.5. By introducing a coefficient Sb(t) related to the basal insulin requirement at different time periods, the gain coefficient is adjusted over time to meet the user's insulin requirements at different times, further improving the accuracy of closed-loop control.

[0456] In this embodiment of the invention, the conversion methods of rPID and rMPC algorithms for transforming blood glucose levels that are asymmetric in the original physical space to approximately symmetric blood glucose risk in the risk space, the compensation methods for various delays, and the beneficial effects are as described above and will not be repeated here. Furthermore, the calculation results of each algorithm can be further processed; the further processing methods and beneficial effects are as described above and will not be repeated here.

[0457] Figure 6 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to another embodiment of the present invention.

[0458] In this embodiment of the invention, the closed-loop artificial pancreas insulin infusion control system mainly includes a detection module 100, an infusion module 102, and an electronic module 103.

[0459] The detection module 100 is used to continuously detect the user's real-time blood glucose level. Generally, the detection module 100 is a continuous glucose monitoring (CGM) device, which can detect blood glucose levels in real time, monitor blood glucose changes, and send the current blood glucose level to the infusion module 102 and the electronic module 103.

[0460] The infusion module 102 includes the mechanical structure necessary for insulin infusion, as well as components such as the infusion processor 1021 capable of executing the first algorithm, and is controlled by the electronic module 103. After receiving the current blood glucose value sent by the detection module 100, the infusion module 102 calculates the required first insulin infusion volume I1 using the first algorithm, and sends the calculated first insulin infusion volume I1 to the electronic module 103.

[0461] Electronic module 103 is used to control the operation of detection module 100 and infusion module 102. Therefore, electronic module 103 is connected to both detection module 100 and infusion module 102. Here, electronic module 103 is an external electronic device such as a mobile phone or handheld device; therefore, the connection refers to a wireless connection. Electronic module 103 includes a second processor. In this embodiment of the invention, the second processor is an electronic processor 1031 or similar element capable of executing a second algorithm and a third algorithm. After receiving the current blood glucose value sent by detection module 100, electronic module 103 calculates the required second insulin infusion amount I2 using the second algorithm. Here, the first and second algorithms used by electronic module 103 and infusion module 102 to calculate the required insulin amount are different.

[0462] After receiving the first insulin infusion volume I1 from the infusion module 102, the electronic module 103 further optimizes and calculates the first insulin infusion volume I1 and the second insulin infusion volume I2 using a third algorithm to obtain the final insulin infusion volume I3. This final insulin infusion volume I3 is then sent to the infusion module 102, which infuses the required insulin I3 into the user's body. Simultaneously, the infusion status of the infusion module 102 is also fed back to the electronic module 103 in real time. The specific optimization method is as described above. That is:

[0463] When I1 = I2, I3 = I1 = I2;

[0464] When I1 ≠ I2, the electronic module 103 further substitutes the arithmetic mean or weighted value of the two into the algorithm to recalculate the current insulin infusion volume I1 and I2. If the data are still different, the above process is repeated until I3 = I1 = I2, that is:

[0465] ① Calculate the average value of the first insulin infusion volume I1 and the second insulin infusion volume I2.

[0466] ② Average value Substitute these parameters into the first and second algorithms and adjust the algorithm parameters accordingly.

[0467] ③ Based on the current blood glucose level, the first algorithm and the second algorithm after parameter adjustment, recalculate the first insulin infusion volume I1 and the second insulin infusion volume I2;

[0468] ④ Repeat steps ① to ③ until I1 = I2, and finally the insulin infusion volume I3 = I1 = I2.

[0469] or:

[0470] ① Calculate the weighted average of the first insulin infusion volume I1 and the second insulin infusion volume I2. Where α and β are the weighting coefficients of the first insulin infusion volume I1 and the second insulin infusion volume I2, respectively;

[0471] ②The weighted average Substitute these parameters into the first and second algorithms and adjust the algorithm parameters accordingly.

[0472] ③ Based on the current blood glucose level, the first algorithm after parameter adjustment, and the second algorithm, recalculate the first insulin infusion volume I1 and the second insulin infusion volume I2;

[0473] ④ Repeat steps ① to ③ until I1 = I2, and the final insulin infusion volume I3 = I1 = I2.

[0474] When the two differ, the electronic module 103 can also perform statistical analysis on both I1 and I2 with historical information such as the user's past physical state, blood glucose level, and insulin infusion volume to obtain the current statistical analysis result I4. The electronic module 103 then selects the one closer to the statistical analysis result I4 as the final insulin infusion volume I3 and sends the final insulin infusion volume I3 to the infusion device 102 for infusion; that is:

[0475]

[0476] In this embodiment of the invention, the user's historical information can be stored in the electronic module 103 or in the cloud management system (not shown). The cloud management system and the electronic module 103 are wirelessly connected.

[0477] Figure 7 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to another embodiment of the present invention.

[0478] In this embodiment of the invention, the closed-loop artificial pancreas insulin infusion control system mainly includes a detection module 100, an infusion module 102, and an electronic module 103.

[0479] The detection module 100 is used to continuously detect the user's real-time blood glucose level. Generally, the detection module 100 is a continuous glucose monitoring (CGM) device, which can detect blood glucose levels in real time and monitor blood glucose changes. The current blood glucose value is only sent to the infusion module 102. The detection module 100 also includes a second processor. In this embodiment of the invention, the second processor is a component such as the detection processor 1001 that can execute a second algorithm. After detecting the real-time blood glucose level, the detection module 100 directly calculates the second insulin infusion volume I2 using the second algorithm and sends the calculated second insulin infusion volume I2 to the electronic module 103.

[0480] As described above, the infusion module 102 receives the current blood glucose value sent by the detection module 100, calculates the first insulin infusion volume I1 using a first algorithm, and sends the first insulin infusion volume I1 to the electronic module 103. Here, the first algorithm and the second algorithm used by the detection module 103 and the infusion module 102 to calculate the insulin volume are different.

[0481] After receiving the first insulin infusion volume I1 and the second insulin infusion volume I2 from the detection module 100 and the infusion module 102 respectively, the electronic module 103 further optimizes the first insulin infusion volume I1 and the second insulin infusion volume I2 using a third algorithm to obtain the final insulin infusion volume I3, and sends the final insulin infusion volume I3 to the infusion module 102, which then infuses the required insulin I3 into the user's body. Simultaneously, the infusion status of the infusion module 102 can also be fed back to the electronic module 103 in real time. The specific optimization method is as described above.

[0482] In the two embodiments of the present invention described above, after the detection module 100 detects the current blood glucose value, the infusion processor 1021 initially calculates the first insulin infusion volume I1, and the second processor (such as the electronic processor 1031 and the detection processor 1001) initially calculates the second insulin infusion volume I2, and sends I1 and I2 to the electronic module 103. The electronic module 103 performs further optimization, and then sends the optimized final insulin infusion volume I3 to the infusion module 102 for insulin infusion, thereby improving the accuracy of the infusion command.

[0483] In the two embodiments of the present invention described above, the first algorithm and the second algorithm are one of the classic PID algorithm, the classic MPC algorithm, the rMPC algorithm, or the rPID algorithm. The advantages of using the rPID or rMPC algorithm for calculation are as described above, and the beneficial effects of further optimization methods are also as described above, and will not be repeated here.

[0484] The embodiments of the present invention do not limit the specific location and connection relationship of the detection module 100 and the infusion module 102, as long as the aforementioned functional conditions are met.

[0485] In one embodiment of the invention, the two modules are electrically connected to form a single integrated structure and are adhered to the same location on the user's skin. By connecting the two modules into a single unit and adhering them to the same location, the number of devices that can be attached to the user's skin is reduced, thereby lessening the interference with the user's movement caused by having too many devices attached. Simultaneously, this effectively solves the problem of poor wireless communication between separate devices, further enhancing the user experience.

[0486] In another embodiment of the invention, the two modules are disposed in different structures and attached to different locations on the user's skin. In this case, the detection module 100 and the infusion module 102 transmit wireless signals to each other to achieve interconnection.

[0487] Figure 8 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas multi-drug infusion control system according to another embodiment of the present invention.

[0488] As described above, the closed-loop artificial pancreas insulin infusion control system in this embodiment of the invention mainly includes a detection module 100, a program module 101, and an infusion module 102. The infusion module 102 has a multi-drug infusion function. The drugs can be a combination of drugs for regulating blood sugar control in diabetic patients, whose metabolite is glucose. The main drugs are hypoglycemic drugs such as insulin and its analogues. Other combination drugs are hypoglycemic drugs with opposite effects, such as glucagon and its analogues, cortisol and its analogues, growth hormone and its analogues, epinephrine and its analogues, glucose, etc., and dextrin analogues with similar effects (such as pramlintide), etc.

[0489] The infusion module 102 can infuse hypoglycemic drugs and / or hypoglycemic drugs into the user's body according to the infusion instructions for hypoglycemic drugs and / or hypoglycemic drugs issued by the program module 101. Hypoglycemic drugs and hypoglycemic drugs can be infused through different drug tubing, or they can be infused through the same drug tubing but not at the same time. The specific design of the drug tubing is not limited here.

[0490] Figure 9 This is a schematic diagram of dual-drug infusion switching according to two embodiments of the present invention.

[0491] In one embodiment of the present invention, the infusion command for a hypoglycemic drug and / or the current infusion command for a hyperglycemic drug are estimated by comparing blood glucose concentration. P With target blood glucose level G B The resulting blood glucose concentration estimate G P Blood glucose levels can be estimated using the rMPC prediction model or other suitable algorithms; blood glucose-lowering drug infusion data and / or blood glucose-raising drug infusion data can be calculated using the aforementioned rMPC algorithm, rPID algorithm, or combined artificial pancreas algorithm. Specifically:

[0492] When G P ≥G BAt that time, the infusion module 102 begins to calculate the hypoglycemic drug infusion data I based on the rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. t Administer hypoglycemic drugs via infusion;

[0493] When G P <G B At that time, the infusion module 102 begins to calculate the blood glucose-raising drug infusion data D based on the rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. t Administer blood glucose-raising medication via infusion;

[0494] It should be noted that, in the embodiments of the present invention, I b This means controlling blood glucose at the target blood glucose level (G) without interference. B The amount of hypoglycemic drugs that need to be infused at that time, when G P =G B At that time, I t =I b When G P >G B At that time, with the infusion of hypoglycemic drugs, G P Further reduction, I t It will also decrease. When the infusion module 102 has only one set of drug infusion tubing, when G P <G B At that time, i.e., I t <I b At that time, the infusion module 102 begins infusion of blood glucose-raising drugs, and the blood glucose-raising drug infusion data D... t The infusion of hypoglycemic drugs can be calculated using the rMPC algorithm, rPID algorithm, or a combined artificial pancreas algorithm, while simultaneously stopping the infusion of hypoglycemic drugs to prevent them from interfering with each other due to antagonistic effects. When the infusion module 102 has at least two sets of drug infusion lines, when 0 ≤ I t <I b At the same time, while infusing blood glucose-raising drugs, blood glucose-lowering drugs can also be continued, which can effectively prevent hypoglycemia; when I t When the blood glucose level is <0, the infusion of hypoglycemic drugs should be stopped and only hypoglycemic drugs should be infused.

[0495] In another embodiment of the present invention, the hypoglycemic drug infusion command and / or the current hypoglycemic drug infusion command can be directly compared by the required amount of hypoglycemic drug I. t and target blood glucose lowering drug dose I b To determine the required amount of blood sugar-lowering medication I t and target blood glucose lowering drug dose I b The calculation can be performed using the aforementioned rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. Specifically: when the infusion module 102 has at least two sets of drug infusion tubing:

[0496] When I t ≥I b At that time, the infusion module 102 begins to calculate the hypoglycemic drug infusion data I based on the rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. t Administer hypoglycemic drugs via infusion;

[0497] When 0≤I t <I b At the same time, while infusing blood glucose-raising drugs, blood glucose-lowering drugs can also be infused, which can effectively prevent hypoglycemia. Blood glucose-lowering drug infusion data I t Blood glucose-raising drug infusion data D t All of these can be calculated using the aforementioned rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm.

[0498] When I t When <0, the infusion of hypoglycemic drugs is stopped and only hypoglycemic drugs are infused. Hypoglycemic drug infusion data D t It can be calculated using the rMPC algorithm, the rPID algorithm, or a composite artificial pancreas algorithm.

[0499] When the infusion module 102 has only one set of drug infusion tubing:

[0500] When I t When the blood glucose level is ≥0, the infusion module 102 starts calculating the blood glucose-lowering drug infusion data I based on the rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. t Administer hypoglycemic drugs via infusion;

[0501] When I t When the blood glucose level is <0, the infusion of hypoglycemic drugs should be stopped and only hypoglycemic drugs should be infused.

[0502] Preferably, in this embodiment of the invention, the hypoglycemic drug is insulin, and the hypoglycemic drug is glucagon.

[0503] It should be noted that in the above embodiments, the calculation methods for the hypoglycemic drug infusion data and glucagon infusion data at each stage can be the same or different. Preferably, the same algorithm architecture is used for calculation to ensure the consistency of the basic conditions during calculation and to make the calculation results more accurate. More preferably, a composite artificial pancreas algorithm is used for calculation to fully utilize the advantages of the rPID algorithm and the rMPC algorithm to deal with complex scenarios and to achieve a more ideal blood glucose control level.

[0504] Figure 10 This is a schematic diagram of the module relationships of a closed-loop artificial pancreas insulin infusion control system according to another embodiment of the present invention.

[0505] In this embodiment of the invention, the closed-loop artificial pancreas insulin infusion control system disclosed in this embodiment mainly includes a detection module 200 and an infusion module 202. The detection module 200 is used to continuously detect the user's current blood glucose level. Generally, the detection module 200 is a continuous glucose monitoring (CGM) device, which can detect the user's current blood glucose level in real time and monitor blood glucose changes. The detection module 200 also includes a detection processing unit 2001, which has a preset algorithm for calculating the insulin infusion volume. When the detection module 200 detects the user's current blood glucose level, the detection processing unit 2001 calculates the amount of insulin required by the user through the preset algorithm and sends the required amount of insulin to the infusion module 202.

[0506] The infusion module 202 includes the necessary mechanical structure for insulin infusion and an electronic transceiver that receives user insulin dosage information from the detection module 200. Based on the current insulin infusion volume data sent by the detection module 200, the infusion module 202 infuses the required amount of insulin into the user's body. Simultaneously, the infusion status of the infusion module 102 is also fed back to the detection module 200 in real time.

[0507] In this embodiment of the invention, the algorithm preset in the detection and processing unit 2001 for calculating the insulin infusion volume is one of the classic PID algorithm, classic MPC algorithm, rMPC algorithm, rPID algorithm, or composite artificial pancreas algorithm. The calculation method and beneficial effects of using rPID, rMPC algorithm, or composite artificial pancreas algorithm are as described above and will not be repeated here.

[0508] The embodiments of the present invention do not limit the specific location and connection relationship of the detection module 2100 and the infusion module 202, as long as the aforementioned functional conditions can be met.

[0509] In one embodiment of the invention, the two modules are electrically connected to form a single integrated structure and are adhered to the same location on the user's skin. By connecting the two modules into a single unit and adhering them to the same location, the number of devices that can be attached to the user's skin is reduced, thereby lessening the interference with the user's movement caused by having too many devices attached. Simultaneously, this effectively solves the problem of poor wireless communication between separate devices, further enhancing the user experience.

[0510] In another embodiment of the invention, the two modules are disposed in different structures and attached to different locations on the user's skin. In this case, the detection module 200 and the infusion module 202 transmit wireless signals to each other to achieve interconnection.

[0511] In summary, this invention discloses a closed-loop artificial pancreas insulin infusion control system. The system is pre-programmed with an rPID algorithm, which transforms the asymmetric blood glucose in the original physical space into an approximately symmetric blood glucose risk in the risk space. It retains the simplicity and robustness of the PID algorithm while also possessing the advantages of PID's precision and flexibility, thus achieving precise control of the closed-loop artificial pancreas insulin infusion system.

[0512] While specific embodiments of the invention have been described in detail by way of examples, those skilled in the art should understand that the above examples are for illustrative purposes only and are not intended to limit the scope of the invention. Those skilled in the art should understand that modifications can be made to the above embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.

Claims

1. A closed-loop artificial pancreas insulin infusion control system, characterized in that, include: The detection module is used to continuously detect the current blood glucose value G; A program module connected to the detection module includes a pre-defined rPID algorithm that converts asymmetric blood glucose levels in the original physical space into approximately symmetric blood glucose risk in the risk space, and a target blood glucose value G. B The rPID algorithm calculates insulin infusion indications based on blood glucose risk. and An infusion module, which is connected to the program module, controls the infusion module to infuse insulin according to the insulin infusion instruction calculated by the rPID algorithm; The rPID algorithm is as follows: in: rPID(t) is the insulin infusion instruction sent to the infusion module after risk conversion; K P This is the gain coefficient for the proportional portion; K I This is the gain coefficient for the integral part; K D The gain coefficient is the differential part. C is a constant; r represents blood sugar risk; The blood glucose risk space transformation method of the rPID algorithm is an improved control variability grid analysis transformation: Where r t+j G represents the blood glucose risk value after step j. B Indicates the target blood glucose level; G t+j This represents the blood glucose value at step j; At the same time, its maximum value is limited: |r t+j |=min(|r t+j |,n) The maximum value n is limited to a range of 0–80 mg / dL.

2. The closed-loop artificial pancreas insulin infusion control system according to claim 1, characterized in that, The blood glucose risk r is related to the current blood glucose level G and the target blood glucose level G. B The blood glucose deviation Ge is related to the amount of blood glucose deviation, where Ge = GG B .

3. The closed-loop artificial pancreas insulin infusion control system according to claim 2, characterized in that, The relationship between the blood glucose risk r and the blood glucose deviation Ge is as follows:

4. The closed-loop artificial pancreas insulin infusion control system according to claim 2, characterized in that, The relationship between the blood glucose risk r and the blood glucose deviation Ge is as follows:

5. The closed-loop artificial pancreas insulin infusion control system according to claim 1, characterized in that, The maximum value n is set to 60 mg / dL.

6. The closed-loop artificial pancreas insulin infusion control system according to any one of claims 1-5, characterized in that, The target blood glucose value G B The concentration is 80–140 mg / dL.

7. The closed-loop artificial pancreas insulin infusion control system according to any one of claims 6, characterized in that, The target blood glucose value G B The concentration is 110–120 mg / dL.

8. The closed-loop artificial pancreas insulin infusion control system according to any one of claims 1-5, characterized in that, The rPID algorithm also includes one or more of the following processing methods: ① Based on the insulin absorption delay in the artificial pancreas control system, deduct the amount of unabsorbed plasma insulin in the body. in, γ is the compensation coefficient for the estimated plasma insulin concentration on the algorithm output; For estimation of plasma insulin concentration; rPIDc(t) represents the compensated infusion instruction sent to the insulin infusion system after risk conversion; rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion; ②Based on the delayed onset of insulin in the artificial pancreas control system, subtract the amount of insulin (IOB(t)) that has not yet taken effect in the body: rPID′(t)=rPID(t)-IOB(t) in: rPID′(t) represents the infusion instruction sent to the insulin infusion system after risk conversion, excluding the amount of insulin that has not yet taken effect in the body; rPID(t) represents the infusion instruction sent to the insulin infusion system after risk conversion; IOB(t) represents the amount of insulin in the body that has not yet taken effect at time t. ③ An autoregressive method was used to compensate for the sensing delay of blood glucose and interstitial fluid glucose concentrations.

9. The closed-loop artificial pancreas insulin infusion control system according to claim 8, characterized in that, The plasma insulin concentration was estimated using an autoregressive method.

10. The closed-loop artificial pancreas insulin infusion control system according to claim 8, characterized in that, The compensation coefficient γ is 0.4-0.

6.

11. The closed-loop artificial pancreas insulin infusion control system according to claim 10, characterized in that, The compensation coefficient γ is 0.

5.

12. The closed-loop artificial pancreas insulin infusion control system according to claim 8, characterized in that, The amount of insulin IOB(t) that has not yet taken effect in the body was obtained through an IOB curve.

13. The closed-loop artificial pancreas insulin infusion control system according to claim 8, characterized in that, When calculating the amount of insulin IOB(t) that has not yet taken effect in the body, the insulin amount is treated with meal insulin and non-meal insulin: IOB(t)<IOB m,t +IOB o,t ; in, IOB m,t This represents the amount of insulin consumed after a meal that has not yet taken effect in the body at time t. IOB o,t This represents the amount of non-meal insulin that has not yet taken effect in the body at time t; D i (i = 2 - 8) represent the corresponding coefficients of the IOB curves for insulin action time i; I m,t Indicates the amount of insulin consumed after a meal; I 0,t This indicates the amount of insulin not consumed during meals; IOB(t) represents the amount of insulin in the body at time t that has not yet taken effect.

14. The closed-loop artificial pancreas insulin infusion control system according to claim 1, characterized in that, Two of the detection module, the program module, and the infusion module are interconnected to form an integrated structure, and are respectively attached to different locations on the skin along with the third module.

15. The closed-loop artificial pancreas insulin infusion control system according to claim 1, characterized in that, The detection module, the program module, and the infusion module are connected to form an integrated structure and are attached to the same location on the skin.

Images