Artificial pancreas control method and device including insulin patch

Abstract

To provide an insulin patch control method and a device therefor that ensure high accuracy and stability.SOLUTION: An insulin patch control method and a device therefor are provided. The insulin patch control method according to one embodiment of the present invention comprises the steps of: obtaining, from an insulin patch, blood glucose information measured by a blood glucose sensor; inputting the obtained blood glucose information into a first neural network; estimating a pump parameter of the insulin patch on the basis of output data of the first neural network; and simulating the insulin patch by inputting the blood sugar information and the pump parameter into a second neural network.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a method and apparatus for controlling an insulin patch, and more particularly to a method and apparatus for controlling an insulin patch using pump parameters of the insulin patch obtained using an artificial neural network. [Background technology]

[0002] There has been active development of insulin injectors for use by diabetes patients. Diabetic patients inject insulin into their bodies via insulin pens or insulin pumps, but this has the disadvantage of greatly reducing their activity in daily life.

[0003] As a result, smart insulin patches have been developed, but they have limitations such as excessive cost, difficulty in continuous data management, and difficulty in real-time blood glucose management. Furthermore, there are risks that may occur when adjusting blood glucose levels using insulin patches alone, so there is a need to provide technology that can solve these issues.

[0004] The above-mentioned background art is technical information that the inventor possessed in order to derive the present invention, or that he acquired in the process of deriving the present invention, and is not necessarily publicly known art that was disclosed to the general public prior to the filing of the application for the present invention. Summary of the Invention [Problem to be solved by the invention]

[0005] SUMMARY OF THE DISCLOSURE The present invention provides an insulin patch control method and device that ensures high accuracy and stability. [Means for solving the problem]

[0006] A method for controlling an insulin patch including an insulin pump and a blood glucose sensor according to one embodiment of the present invention can include the steps of acquiring blood glucose information from a blood glucose sensor, estimating pump parameters of the insulin patch based on the acquired blood glucose information, and transmitting a control signal to the insulin pump based on the estimated pump parameters.

[0007] An insulin patch control method performed by a computing device according to another embodiment of the present invention may include the steps of acquiring blood glucose information measured by a blood glucose sensor from an insulin patch, inputting the acquired blood glucose information into a first neural network, estimating pump parameters of the insulin patch based on output data of the first neural network, and simulating the insulin patch using the blood glucose information and pump parameters.

[0008] In one embodiment, the step of inputting the acquired blood glucose information into a first neural network may include a step of transmitting the acquired blood glucose information to an external server and receiving a feedback signal regarding the validity of the blood glucose information from the external server, and the step of simulating the insulin patch using the blood glucose information and pump parameters may further include a step of transmitting a simulation result of the insulin patch to an external server and receiving a feedback signal regarding the simulation result of the insulin patch from the external server.

[0009] Other aspects, features and advantages in addition to those described above will become apparent from the following drawings, claims and detailed description of the invention. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of an insulin patch control system according to one embodiment of the present invention. [Diagram 2]FIG. 2 is a block diagram for explaining the internal configuration of the insulin patch control device according to one embodiment of the present invention. [Diagram 3] FIG. 3 is a flowchart of an insulin patch control method according to one embodiment of the present invention. [Figure 4] FIG. 4 is an exemplary diagram illustrating an embodiment of an insulin patch control system according to an embodiment of the present invention. [Diagram 5] FIG. 5 is a diagram for explaining in detail a part of the operation of FIG. [Figure 6] FIG. 6 is a diagram for explaining in detail a part of the operation of FIG. [Figure 7] FIG. 7 is a diagram for explaining in detail a part of the operation of FIG. BEST MODE FOR CARRYING OUT THEINVENTION

[0011] An insulin patch control method and device are provided. The insulin patch control method according to one embodiment of the present invention includes the steps of acquiring blood glucose information measured by a blood glucose sensor from an insulin patch, inputting the acquired blood glucose information into a first neural network, estimating pump parameters of the insulin patch based on output data of the first neural network, and inputting the blood glucose information and pump parameters into a second neural network to simulate the insulin patch. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The detailed description of the present invention below refers to the accompanying drawings, which show, by way of example, specific embodiments in which the present invention can be practiced. Such embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different from one another but are not necessarily mutually exclusive. For example, specific shapes, structures and characteristics described herein can be modified from one embodiment to another without departing from the spirit and scope of the present invention. It should also be understood that the location or arrangement of individual components in each embodiment can be modified without departing from the spirit and scope of the present invention. Therefore, the detailed description below is not to be taken in a limiting sense, and the scope of the present invention should be accepted as encompassing the scope of the claims and all equivalents thereto. In the drawings, like reference numerals represent the same or similar components throughout the various aspects.

[0013] Artificial intelligence (AI) disclosed in the present invention may refer to a field that studies artificial intelligence or a methodology that can create it, and machine learning, as a field of artificial intelligence technology, may be an algorithm that allows a computing device to learn through data and understand a specific object or condition, or a computer to analyze data as a technical method of finding and classifying patterns in data. Machine learning disclosed in the present invention may be understood to include a method of operation for learning an artificial intelligence model.

[0014] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following, various embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those skilled in the art to easily carry out the present invention.

[0015] Hereinafter, an insulin patch control system according to one embodiment of the present invention will be described in detail with reference to FIG.

[0016] In one embodiment, the insulin patch control system may be composed of only the insulin patch 10, or may be composed of the insulin patch 10 and the insulin patch control device 100. Alternatively, in other embodiments, the insulin patch system may include the insulin patch 10, the insulin patch control device 100, and the medical service provider server 200.

[0017] In one embodiment, the insulin patch 10 can include an insulin pump for injecting insulin into a person's body and a blood glucose sensor for measuring the person's blood glucose. As an example, the insulin patch can be an automatic insulin injector in the form of a wearable device, which is a kind of wearable artificial pancreas device. Also, the insulin patch 10 can determine the insulin injection amount independently without communication with an external device according to the method described below in FIG. 5.

[0018] In one embodiment, the insulin patch 10 can include a blood glucose measuring sensor to measure blood glucose. In this embodiment, the insulin patch 10 can obtain pump parameters through a predetermined algorithm based on blood glucose information obtained through the blood glucose measuring sensor. Then, the insulin patch 10 can independently perform insulin patch control using the above-mentioned pump parameters.

[0019] In one embodiment, the insulin patch control device 100 can communicate with the insulin patch 10 via a network. The communication method between the insulin patch control device 100 and the insulin patch 10 described above is not limited. In one embodiment, the network described above may be a short-range wireless communication network. For example, the network may be a Bluetooth (registered trademark), BLE (Bluetooth Low Energy), or Wifi communication network.

[0020] In one embodiment, the insulin patch control device 100 can obtain blood glucose information from the insulin patch 10 and can use the obtained blood glucose information to estimate pump parameters that determine the amount of insulin infused by the insulin patch 10. In this case, the insulin patch control device 100 can estimate the pump parameters through a neural network using the above-mentioned blood glucose information.

[0021] Then, the insulin patch control device 100 can transmit the estimated pump parameters to the insulin patch 10, and can obtain the control result data of the insulin patch from the insulin patch 10. The form and type of the control result data of the insulin patch according to some embodiments of the present invention are not limited, and in one embodiment, the control result data of the insulin patch can include the above-mentioned pump parameters and blood glucose information corresponding to the pump parameters. In addition, the above-mentioned control result data can include time-series data including blood glucose information over time. Then, the insulin patch control device 100 according to one embodiment can estimate the pathological diagnosis result data of the insulin patch user using the acquired blood glucose information, the estimated pump parameters, and the acquired control result data of the insulin patch.

[0022] In addition, in one embodiment, the insulin patch control device 100 can estimate pump parameters of the insulin patch 10 using a neural network, and after clustering the output data of the above-mentioned neural network, can select the final pump parameters according to a specified criterion. In this case, the insulin patch control device 100 can simulate the insulin patch 10 based on the selected pump parameters.

[0023] That is, in this embodiment, the insulin patch control device 100 performs a simulation through an insulin injection algorithm using pump parameters to improve stability, and then transmits the above-mentioned pump parameters to the insulin patch 10 only if the validity of the pump parameters is verified.

[0024] In another embodiment, the insulin patch control device 100 can estimate the above-mentioned pump parameters by further using the user information of the insulin patch. In this embodiment, the insulin patch control device 100 can store the past insulin patch usage information of the user who uses the insulin patch, and can obtain the user's previous pathological diagnosis result data from the medical service providing server 200. This allows the insulin patch control device 100 to estimate the pump parameters of the personalized insulin patch based on the user information.

[0025] Furthermore, a medical service providing server 200 connected to the insulin patch control device 100 of one embodiment via a network can perform validation of the algorithm for estimating the pump parameters of the insulin patch, and can also perform validation of the simulation results described above.

[0026] In one embodiment, the medical service providing server 200 can also estimate the pathological diagnosis result data of the insulin patch user instead of the above-mentioned insulin patch control device 100. In one embodiment, the medical service providing server 200 can estimate the pathological diagnosis result data of the insulin patch user using blood glucose information acquired from the insulin patch 10, pump parameters of the insulin patch 10, and control result data of the insulin patch 10. As an example, the medical service providing server 200 can input the blood glucose information, pump parameters, and control result data of the insulin patch into an artificial neural network, and can estimate the pathological diagnosis result data using the artificial neural network. The type and form of the pathological diagnosis result data are not limited. For example, the pathological diagnosis result data according to one embodiment can include a diabetes diagnosis result.

[0027] Thereafter, the medical service providing server 200 according to an embodiment of the present invention may receive a user's feedback signal regarding the estimated pathological diagnosis result data. In this case, the user may be a medical staff member providing medical services. Thereafter, the medical service providing server 200 may input the pathological diagnosis result data corrected according to the above-mentioned user's feedback signal into the above-mentioned neural network to perform re-learning and / or estimation. By repeating the above-mentioned steps, the accuracy of the user's pathological diagnosis result data estimated by the medical service providing server 200 according to an embodiment of the present invention may be improved.

[0028] Hereinafter, the configuration and operation of the hardware that constitutes the insulin patch control system will be described in detail with reference to FIG.

[0029] As described above, the insulin patch control system of one embodiment includes an insulin patch 10, and in other embodiments may further include an insulin patch control device 100, and in some other embodiments may further include a medical service provider server 200.

[0030] In one embodiment, the insulin patch 10, the insulin patch control device 100, and the medical service providing server 200 may include an input / output interface 11, 101, 201, a memory 12, 102, 202, a processor 13, 103, 203, and a communication module 14, 104, 204. The memory 12, 102, 202 is a computer-readable recording medium and may include a non-permanent mass storage device such as a random access memory (RAM), a read only memory (ROM), and a disk drive. As an example, the memory 12 of the insulin patch 10 may temporarily or permanently store blood glucose information and / or pump parameter information, and the memory 102 of the insulin patch control device 100 may store user-specific blood glucose information and / or pump parameter information of the insulin patch. In addition, in one embodiment, the memory 202 of the medical service providing server 200 may store pathological diagnosis result data of a user.

[0031] The processor 13, 103, 203 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input / output operations. The instructions may be provided to the processor 13, 103, 203 by the memory 12, 102, 202 or the communication module 14, 104, 204. For example, the processor 13, 103, 203 may be configured to execute instructions received by a program code stored in a recording device such as the memory 12, 102, 202. In one embodiment, the processor 103 of the insulin patch control device 100 may obtain blood glucose information from the insulin patch 10, input the obtained blood glucose information into a first neural network, estimate pump parameters of the insulin patch based on the output data of the first neural network, send a control signal to the insulin patch 10 based on the estimated pump parameters, obtain control result data from the insulin patch 10, input the blood glucose information, the pump parameters, and the control result data into a second neural network, and estimate user pathology diagnosis result data of the insulin patch based on the output data of the second neural network.

[0032] The communication modules 14, 104, and 204 can provide a function for communication via a network. As an example, a request generated by the processor 103 of the insulin patch control device 100 according to a program code stored in a recording device such as the memory 102 can be transmitted to the insulin patch 10 via a network under the control of the communication module 104. Also, control signals, instructions, content, files, etc. provided under the control of the processor 203 of the medical service providing server 200 can be transmitted to the insulin patch control device 100 via the communication module 104 via the network. Furthermore, the control signals, instructions, etc. of the medical service providing server 200 received via the communication module 104 of the insulin patch control device 100 can be transmitted to the processor 103 or memory 102 of the insulin patch control device 100, and blood glucose information and / or control result data from the insulin patch 10 can be stored in a storage medium that the insulin patch control device 100 can further include. Furthermore, the communication method of the communication modules 14, 104, and 204 is not limited, and the network may be a short-range wireless communication network. For example, the network may be a Bluetooth, Bluetooth Low Energy (BLE), or Wifi communication network.

[0033] In addition, the input / output interfaces 11, 101, and 201 can receive user input and display output data. The input / output interface 11 of the insulin patch 10 according to one embodiment can receive an input signal from the user and display operation information of the insulin patch on the display. In one embodiment, the input / output interface 201 of the medical service providing server 200 can receive a user feedback signal and output control result data of the insulin patch 10.

[0034] Also, in other embodiments, the insulin patch 10, the insulin patch control device 100, and the medical service providing server 200 may include more components than those in FIG. 2. However, it is not necessary to explicitly show most of the conventional components. For example, the insulin patch control device 100 may include a battery and a charging device that supplies power to the internal components of the user terminal, may be embodied to include at least some of the above-mentioned input / output devices, or may further include other components such as a transceiver, a GPS (Global Positioning System) module, various sensors, and a database. Also, although not shown in FIG. 2, the insulin patch 10 may include a blood glucose measurement sensor for measuring blood glucose levels, and may further include an insulin injection tube in other embodiments.

[0035] FIG. 3 is a diagram for explaining an insulin patch control method according to one embodiment of the present invention.

[0036] In step S110, the insulin patch control device can obtain blood glucose information from the blood glucose sensor of the insulin patch. In one embodiment, the insulin patch can obtain blood glucose information of the user through the blood glucose sensor. The type of the blood glucose sensor according to some embodiments of the present invention is not limited, and may be, for example, a continuous glucose monitoring (CGM) sensor. In one embodiment, the blood glucose information includes the blood glucose level of the user, and in some embodiments, may further include information regarding the blood glucose measurement time and blood glucose measurement method.

[0037] In step S120, the insulin patch control device inputs the acquired blood glucose information into a first neural network, and can estimate pump parameters of the insulin patch based on the output data of the first neural network described above. In one embodiment, the first neural network to which the blood glucose information is input can estimate the insulin injection amount according to the blood glucose level included in the blood glucose information. In this case, the amount of insulin injected through the insulin patch can be adjusted according to the pump parameters. Therefore, the above-mentioned first neural network to which the blood glucose information is input according to one embodiment can estimate optimal pump parameters based on the blood glucose information acquired in real time.

[0038] In another embodiment, the insulin patch can estimate pump parameters independently of the insulin patch control device based on the measured blood glucose information, and can adjust the insulin injection amount based on the estimated pump parameters. That is, the insulin patch control method according to this embodiment can stably inject insulin to the user even if a communication problem occurs between the insulin patch and the insulin patch control device or a problem occurs with the insulin patch control device. A more detailed description will be given later with reference to FIG. 5.

[0039] In step S130, the insulin patch control device can send a control signal to the insulin patch based on the pump parameters estimated in step S120. In one embodiment, the insulin patch control device and the insulin patch can communicate through a network. As mentioned above, the network communication method between the insulin patch control device and the insulin patch is not limited, but in one example, the above-mentioned network can be a short-range wireless communication network, such as a Bluetooth communication network.

[0040] In another embodiment, the insulin patch control device can perform a simulation based on the insulin infusion algorithm using the estimated pump parameters before sending a control signal to the insulin patch. The pump parameters can also be re-estimated based on the simulation results. For example, if the simulation results show that the user's predicted blood glucose level is outside the valid range, it can be determined that the estimated pump parameters are invalid, and the pump parameters can then be re-estimated using the neural network described above. The insulin patch control method according to this embodiment can improve the accuracy of the insulin patch. A more detailed description will be given later in FIG. 6.

[0041] In step S140, the insulin patch control device can obtain control result data from the insulin patch. In one embodiment, the control result data of the insulin patch can include at least one of blood glucose information measured over time, pump parameters estimated according to the measured blood glucose information, the amount of insulin injected according to the pump parameters, and blood glucose information after insulin injection. It should be noted that the type and form of data included in the control result data of the insulin patch according to some embodiments of the present invention are not limited. For example, the control result data of the insulin patch may be one of time series data on blood glucose levels, time series data on pump parameters, or time series data on insulin injection amounts, and can include data in a form consisting of a pair of insulin injection amounts and blood glucose levels, and can include data in a form consisting of a pair of pump parameters and blood glucose levels.

[0042] In step S150, the insulin patch control device can input the blood glucose information, the pump parameters, and the control result data of the insulin patch described above into the second neural network, and in step S160, the insulin patch control device can estimate the user pathology diagnosis result data of the insulin patch based on the output data of the second neural network. More specifically, the insulin patch control device according to one embodiment can input the blood glucose information and the pump parameters into the second neural network to obtain the first pathology diagnosis result data, and then perform feedback on the first pathology diagnosis result data described above based on the control result data of the insulin patch. For example, the first pathology diagnosis result data and the control result data of the insulin patch can be used to retrain the second neural network described above. As an example, the insulin patch control device can perform reinforcement learning of the second neural network using the control result data of the insulin patch. In this case, the control result data of the insulin patch can include a control signal of the insulin patch based on the pump parameters and blood glucose information due to insulin injection, and in some embodiments, can further include the user's diabetes diagnosis result information.

[0043] In another embodiment, an external server connected to the insulin patch control device via a network can execute the above-mentioned steps S150 to S160. In this embodiment, the external server may be a hospital server or a medical service provider server. In this case, the external server may receive a user feedback signal for the pump parameters of the insulin patch, the simulation result of the insulin patch, or the pathological diagnosis result data of the user. In this case, the user is a user of the hospital server and the medical service provider server, and may be, for example, a doctor, a nurse, or a medical staff. After that, the feedback signal of the external server according to some embodiments of the present invention can be re-inputted to the above-mentioned first neural network or second neural network to re-train the neural network.

[0044] Hereinafter, a series of steps for controlling an insulin patch according to one embodiment of the present invention will be described in detail with reference to FIG.

[0045] The insulin patch 10 according to one embodiment of the present invention can perform network communication with a user terminal 20 that can monitor and / or control the insulin patch. In one example, the insulin patch 10 and the user terminal 20 can perform Bluetooth communication. An application for monitoring and / or controlling the insulin patch 10 can be installed in the user terminal 20 according to the present embodiment, and the user terminal 20 can receive user input for the insulin patch in real time through the application. In this case, according to one embodiment, the insulin patch 10 can estimate pump parameters independently only when communication with the insulin patch control device 100 as well as the user terminal 20 is cut off.

[0046] The insulin patch control device 100 according to one embodiment can estimate pump parameters based on user information of the insulin patch. That is, the insulin patch control device 100 can estimate pump parameters of one or more insulin patches 10 at the same time, and can estimate pump parameters for each insulin patch 10 by further using user information of each insulin patch 10. This may enable personalized insulin patch control. Also, the insulin patch control device 100 can generate a virtual insulin patch corresponding to each insulin patch and perform a simulation of the insulin patch using the estimated pump parameters. Since ensuring stability is very important in injecting insulin, the insulin patch control device 100 according to some embodiments of the present invention can stably control the insulin patch through such a process.

[0047] The external device 200 that acquires the pathological diagnosis result data of the insulin patch user from the insulin patch control device 100 according to one embodiment may further include an additional user terminal 210 that can receive a feedback signal regarding the pathological diagnosis result data. In this case, the user who inputs the feedback signal regarding the pathological diagnosis result data may be a medical staff member such as a doctor. This can improve the accuracy of the pump parameter estimation neural network executed by the insulin patch control device 100 or the external device 200.

[0048] Hereinafter, the steps executed by the insulin patch unit 300, the insulin patch control device 100, and the pathological diagnosis result data analysis unit 400 shown in Fig. 4 will be described in more detail with reference to Figs. 5 to 7. It should be noted that the entity that executes each step described in Figs. 5 to 7 is not necessarily limited to the specific electronic device shown in Fig. 4.

[0049] Hereinafter, with reference to FIG. 5, a process in which the insulin patch independently estimates pump parameters in one embodiment of the present invention will be described in detail.

[0050] In step S210, blood glucose information can be obtained through a CGM sensor. However, it is noted that the CGM sensor is just one example of a blood glucose measurement sensor, and the blood glucose measurement sensor and blood glucose measurement method of the insulin patch according to some embodiments of the present invention do not necessarily utilize a CGM sensor.

[0051] In step S220, the pump parameters can be obtained through a designated algorithm based on the blood glucose information. In one embodiment, the designated algorithm may be similar or identical to the first neural network described above, but if the subject performing step S220 is an insulin patch, it may be a simplified algorithm with less computational complexity than the first neural network to save computing resources. As an example, when an error occurs in a user terminal communicating with the insulin patch, an error occurs in an insulin patch control device, or a communication failure occurs in a network, the insulin patch according to the present embodiment can independently estimate pump parameters. Therefore, when the insulin patch according to the present embodiment cannot receive an insulin patch control signal or pump parameter information from the outside for a designated time, it can independently estimate pump parameters for determining the insulin injection amount. That is, the insulin patch according to the present embodiment can independently determine the insulin injection amount without communication with an external device.

[0052] In step S230, a pump parameter validity verification can be performed. That is, before controlling the insulin patch using the acquired pump parameters, a minimum validity verification to ensure stability can be performed. For example, it can be confirmed whether the acquired pump parameters are pump parameters within a specified stable range. In this case, the specified stable range may be a fixed value input by a user, a value determined based on a previous pump parameter record, or a value estimated through a specified algorithm as in step S220. In one embodiment, it goes without saying that the insulin patch can perform the pump parameter validity verification even when receiving an insulin patch control signal or pump parameters from an insulin patch control device. That is, the insulin patch control method according to this embodiment can perform the above-mentioned pump parameter validity verification in an insulin patch including an insulin injecting unit rather than an insulin patch control device in order to improve stability.

[0053] The insulin patch can be controlled based on the pump parameters whose validity was verified in step S240, the CGM sensor and the insulin patch control device can be controlled based on the control results, and insulin patch control result data can be sent to the insulin patch control device. In another embodiment, if the validity of the pump parameters is not verified in step S230, the above-mentioned steps S210 to S230 can be repeatedly executed.

[0054] Hereinafter, with reference to FIG. 6, a process in which the insulin patch control device according to an embodiment of the present invention executes a simulation will be described in detail.

[0055] In step S310, the insulin patch control device can perform validity verification of the acquired blood glucose information. Before inputting the blood glucose information acquired from the insulin patch into the neural network, the insulin patch control device can perform validity verification on the blood glucose information. For example, it can be confirmed whether the acquired blood glucose information is within a predetermined valid range. This allows the insulin patch control device to check whether an error has occurred in the insulin patch. The above-mentioned valid range of blood glucose information may be a fixed value input by the user, a value determined based on past blood glucose information records for each user, or a value estimated through a predetermined algorithm.

[0056] In step S320, the insulin patch control device can input the blood glucose information to the first neural network to estimate pump parameters of the insulin patch. In one embodiment, the first neural network may be a neural network trained using the user's blood glucose information before insulin injection, the insulin injection amount, the insulin pump parameters, and the user's blood glucose information after insulin injection, or may be a neural network trained by transfer learning using another neural network trained using the insulin injection amount and blood glucose information. Furthermore, in one embodiment, the first neural network may be a neural network trained further using user information. This allows the insulin patch control device to estimate pump parameters by further considering the degree and speed of blood glucose change for each user.

[0057] In step S330, the insulin patch control device can classify the output data of the first neural network and select pump parameters included in the output data according to the specified conditions. In one embodiment, the specified conditions for selecting the pump parameters can be determined based on user information. As an example, the insulin patch control device can use the user information to select a more appropriate pump parameter from pump parameters for injecting a small amount of insulin over a long period of time and pump parameters for injecting a large amount of insulin over a short period of time.

[0058] In step S340, the insulin patch control device can set the pump parameters of the insulin patch to the pump parameters selected in step S330, and in step S350, can run a simulation for a virtual insulin patch through an insulin injection algorithm based on the pump parameters set.

[0059] In one embodiment, if the simulation result shows that the user's predicted blood glucose level is outside the valid range, the insulin patch control device can determine that the estimated pump parameters are not valid. In this case, the insulin patch control device can perform a simulation of the insulin patch using pump parameters not selected in step S330, and can repeatedly perform steps S310 to S340 described above and perform a simulation using re-estimated pump parameters. The accuracy and stability of the pump parameters estimated through the insulin patch control method according to this embodiment can be improved.

[0060] Hereinafter, a process of obtaining pathological diagnosis result data according to an embodiment of the present invention will be described in detail with reference to FIG.

[0061] In step S410, the insulin patch control device can input the obtained blood glucose information and the estimated pump parameters into a second neural network.

[0062] Then, in step S420, the insulin patch control device can retrain the second neural network using the output data of the second neural network and the control result data of the insulin patch already obtained. In one embodiment, the control result data of the insulin patch can include at least one of blood glucose information measured over time, pump parameters estimated according to the measured blood glucose information, the insulin injection amount injected according to the pump parameters, and blood glucose information after insulin injection.

[0063] More specifically, the insulin patch control device according to one embodiment can input blood glucose information and pump parameters to a second neural network to obtain the first pathological diagnosis result data, and then perform feedback on the first pathological diagnosis result data based on the insulin patch control result data. For example, the second neural network can be retrained using the first pathological diagnosis result data and the insulin patch control result data.

[0064] As an example, the insulin patch control device can perform reinforcement learning of the second neural network using the control result data of the insulin patch, where the control result data of the insulin patch can include the control signal of the insulin patch based on the pump parameters and blood glucose information by insulin injection, and in some embodiments, can further include the user's diabetes diagnosis result information.

[0065] The above-described devices may be implemented with hardware components, software components, and / or a combination of hardware and software components. For example, the devices and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications that run on the operating system. Furthermore, the processing device may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device may be described as being one, but a person skilled in the art will understand that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, the processing device may include multiple processors or one processor and one controller. Additionally, other processing configurations are possible, such as parallel processors.

[0066] The software may include a computer program, code, instructions, or a combination of one or more of these, and may configure or instruct a processing device to operate in a desired manner, either individually or collectively. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual device, computer storage medium or device, or transmitted signal wave, to be interpreted by the processing device or to provide instructions or data to the processing device. The software may be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored in one or more computer-readable recording media.

[0067] The method according to the embodiment may be implemented in the form of program instructions that can be executed via various computer means and recorded on a computer-readable recording medium. The computer-readable recording medium may include, alone or in combination, program instructions, data files, data structures, and the like. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs, DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include not only machine code, such as produced by a compiler, but also high-level language code that can be executed by a computer using an interpreter, or the like. The hardware devices may be configured to operate as one or more software modules to perform the operations of the embodiment, and vice versa.

[0068] Although the embodiments have been described above with reference to the limited embodiments and drawings, those skilled in the art will appreciate that various modifications and variations may be made from the above description, for example, the described techniques may be performed in a different order than described, and / or the components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different manner than described, or may be substituted or replaced by other components or equivalents, and still achieve suitable results.

[0069] Accordingly, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims

1. A method for controlling an insulin pump, which is executed by a control device including a computing device and communicates via a network, comprising: inputting blood glucose information acquired by a blood glucose sensor into a first neural network of the computing device; estimating pump parameters based on output data of the first neural network; selecting and simulating one of the estimated pump parameters; If the simulation result indicates that the predicted blood glucose level is outside a valid range, determining that the selected pump parameters are invalid and selecting other pump parameters from the estimated pump parameters or re-estimating the pump parameters; and sending a control signal to the insulin pump if the predicted blood glucose level is within a valid range as a result of the simulation. If the insulin pump cannot receive the control signal for a specified time due to a communication failure of the network or a problem with the control device, a processor communicating with the insulin pump independently estimates the pump parameters through an algorithm based on the acquired blood glucose information; A method of controlling an insulin pump comprising sending a control signal to an insulin pump based on the estimated pump parameters.

2. After the control device transmits the control signal, the control device obtains a control result; inputting the acquired blood glucose information and the estimated pump parameters into a second neural network to acquire first pathology diagnosis result data; using the control result to perform feedback to a second neural network in relation to the first pathology diagnosis result data; The insulin pump control method according to claim 1 , further comprising a step of estimating user pathology diagnosis result data based on output data of the second neural network.

3. The step of inputting the acquired blood glucose information into a first neural network includes: transmitting the acquired blood glucose information to an external server and receiving a feedback signal from the external server regarding the validity of the blood glucose information; The simulating step includes: The insulin pump control method of claim 2 , further comprising the steps of transmitting the simulation result to an external server and receiving a feedback signal regarding the simulation result from the external server.

4. 3. The insulin pump control method of claim 2, further comprising: acquiring pump parameters through a specified algorithm based on the blood glucose information acquired through the blood glucose sensor; verifying the validity of the acquired pump parameters; and then generating a control signal.

5. An insulin pump control system including a controller, a processor, an insulin pump, and a blood glucose sensor, The processor, Acquire blood glucose information measured by the blood glucose sensor, input the acquired blood glucose information to a first neural network of a computing device included in the control device, estimate pump parameters of the insulin pump based on output data of the first neural network, and select one of the estimated pump parameters to perform a simulation; If the simulation results in the predicted blood glucose level being outside a valid range, the selected pump parameters are determined to be invalid, and other pump parameters are selected from the estimated pump parameters or the pump parameters are re-estimated; If the simulation results in a predicted blood glucose value being within a valid range, the insulin pump control system sends a control signal to the insulin pump to obtain a control result; inputs the obtained blood glucose information and the estimated pump parameters into a second neural network of the computing device to obtain first pathology diagnosis result data; and uses the control result to perform feedback to the second neural network in relation to the first pathology diagnosis result data.

Images