Electronic device and control method thereof

Abstract

This electronic device comprises: a sensor for sensing information indicating a state of the electronic device; a memory for storing a first neural network model and a second neural network model, which are pre-trained to classify an error type of the electronic device, and one or more instructions; and a processor which is connected to the memory to operate, and executes the one or more instructions stored in the memory, wherein the one or more instructions, when executed by the processor, cause the electronic device to perform: acquiring first error type information and second error type information by inputting the information indicating the state of the electronic device to each of the first neural network model and the second neural network model; identifying an error category of the electronic device on the basis of the first error type information and the second error type information; and controlling an operation of the electronic device on the basis of the error category, and wherein the first neural network model is a model pre-trained using first training data for multiple error types, and the second neural network model is a model pre-trained using second training data, which is obtained by processing the first training data, in order to distinguish between error categories to which the respective multiple error types belong.

Description

Electronic device and method of controlling the same

[0001] The present disclosure relates to an electronic device and a control method thereof, and more particularly, to an electronic device and a control method thereof capable of accurately identifying errors in the electronic device using a plurality of neural network models learned in different ways.

[0002] Electronic devices can perform various functions (or operations) by utilizing various internal components. Errors in electronic devices can occur due to aging or failure of internal components. These errors may have a minor impact on the operation of the electronic device, or conversely, they may prevent the device from performing certain functions.

[0003] As errors that can occur in electronic devices are diverse, and the operation of the electronic device may vary depending on each error state, a method that can accurately classify each error was required.

[0004] According to one embodiment of the present disclosure, an electronic device includes a sensor that detects information indicating a state of the electronic device, a memory that stores a first neural network model and a second neural network model pre-learned to classify an error type of the electronic device and one or more instructions, and a processor that is connected to the memory and operates and executes one or more instructions stored in the memory, wherein the one or more instructions, when executed by the processor, input the information indicating the state of the electronic device into each of the first neural network model and the second neural network model to obtain first error type information and second error type information, identify an error category of the electronic device based on the first error type information and the second error type information, and control an operation of the electronic device based on the error category, wherein the first neural network model is a model pre-learned using first learning data for a plurality of error types, and the second neural network model is a model pre-learned using second learning data that is obtained by processing the first learning data to distinguish between error categories to which each of the plurality of error types belongs.

[0005] According to one embodiment of the present disclosure, the first neural network model and the second neural network model are models that output probability values ​​corresponding to each of a plurality of error types, and the one or more instructions, when executed by the processor, cause the electronic device to perform the following: the probability values ​​of each of the plurality of error types included in the first error type information are smaller than a first reference value corresponding to a criterion for error judgment, and at least one probability value included in the first error type information is larger than a second reference value that is smaller than the first reference value, and the error category of the electronic device can be confirmed using the probability values ​​of each of the plurality of error types included in the second error type information.

[0006] According to one embodiment of the present disclosure, the first neural network model and the second neural network model are models that output probability values ​​corresponding to each of a plurality of error types, and the one or more instructions, when executed by the processor, cause the electronic device to perform the following: if at least one probability value among the plurality of probability values ​​in the first error type information is greater than or equal to a first reference value that serves as a reference for error determination, an error category corresponding to the error type greater than or equal to the first reference value is identified as an error category of the electronic device; if the plurality of probability values ​​in the first error type information are less than a second reference value that is less than the first reference value, it is identified that there is no error in the electronic device; and if at least one probability value among the plurality of probability values ​​in the first error type information has a value between the first reference value and the second reference value, and if at least one probability value among the plurality of probability values ​​in the second error type information is greater than or equal to a third reference value that serves as a reference for error determination, an error category corresponding to the error type greater than or equal to the third reference value is identified as an error category of the electronic device.

[0007] According to one embodiment of the present disclosure, the one or more instructions, when executed by the processor, may cause the electronic device to perform the following: determine one of no error, a first error category, a second error category, and a third error category based on the acquired first error type information and the second error type information; and, if the determined error category is the first error category, control the operation of the electronic device to operate in a first mode that limits the performance of a driving device corresponding to the information indicating the state of the electronic device; if the determined error category is the second error category, control the operation of the electronic device to operate in a second mode that limits the performance of the driving device more than the first mode; and if the determined error category is the third error category, control the operation of the electronic device to not operate the driving device.

[0008] According to one embodiment of the present disclosure, the electronic device further includes a display, and the one or more instructions, when executed by the processor, control the display to display information about the identified error category, such that the electronic device performs the following:

[0009] According to one embodiment of the present disclosure, the electronic device further includes an input device for receiving a control command, wherein the one or more instructions, when executed by the processor, cause the electronic device to perform the following: if the identified error category is a first error category and the control command is a command corresponding to the performance of the driving device, the electronic device can control the display to display information indicating that the control command cannot be performed.

[0010] According to one embodiment of the present disclosure, the electronic device further includes a driving device including a motor and an inverter that provides driving power to the motor, wherein the sensor senses a plurality of current values ​​within the driving device, and the one or more instructions, when executed by the processor, determine an error category for at least one of the motor and the inverter based on the first error type information and the second error type information, such that the electronic device performs the following:

[0011] According to one embodiment of the present disclosure, the first learning data may be data sampled to have a similar number of data for each error type among the collected data, and the second learning data may be learning data in which data within a preset similarity range with a second error type belonging to the second error category among the first learning data of the first error type belonging to the first error category and data within a preset similarity range with the first error type among the second learning data of the second error type are removed.

[0012] According to one embodiment of the present disclosure, the second learning data may include augmented data corresponding to one of the first error type and the second error type, such that the learning data corresponding to the first error category and the learning data corresponding to the second error category have a similarity within a predetermined threshold value.

[0013] According to one embodiment of the present disclosure, data corresponding to the first error category and the second error category in the first learning data are data based on the information indicating the state of the electronic device, and the first learning data may further include experimental data for a plurality of error types in the third error category.

[0014] According to one embodiment of the present disclosure, a method for controlling an electronic device includes the steps of detecting information indicating a state of the electronic device, inputting the information indicating the state of the electronic device into a first neural network model and a second neural network model, which are pre-trained to classify an error type of the electronic device, to obtain first error type information and second error type information, confirming an error category of the electronic device based on the first error type information and the second error type information, and controlling an operation of the electronic device based on the error category, wherein the first neural network model is a model pre-trained using first learning data for a plurality of error types, and the second neural network model is a model pre-trained using second learning data that is obtained by processing the first learning data to distinguish between error categories to which each of the plurality of error types belongs.

[0015] According to one embodiment of the present disclosure, the first neural network model and the second neural network model are models that output probability values ​​corresponding to each of a plurality of error types, and the step of checking the error category may include checking the error category of the electronic device using the probability values ​​of each of the plurality of error types included in the first error type information based on the probability values ​​of each of the plurality of error types included in the first error type information being smaller than a first reference value corresponding to a criterion for error judgment and at least one probability value included in the first error type information being larger than a second reference value that is smaller than the first reference value.

[0016] According to one embodiment of the present disclosure, the first neural network model and the second neural network model are models that output probability values ​​corresponding to each of a plurality of error types, and the step of checking the error category includes: if at least one probability value among the plurality of probability values ​​in the first error type information is greater than or equal to a first reference value that serves as a criterion for error determination, the error category corresponding to the error type greater than or equal to the first reference value may be checked as an error category of the electronic device; if the plurality of probability values ​​in the first error type information are smaller than a second reference value that is less than the first reference value, it may be checked that there is no error in the electronic device; and if at least one probability value among the plurality of probability values ​​in the first error type information has a value between the first reference value and the second reference value, and if at least one probability value among the plurality of probability values ​​in the second error type information is greater than or equal to a third reference value that serves as a criterion for error determination, the error category corresponding to the error type greater than or equal to the third reference value may be checked as an error category of the electronic device.

[0017] According to one embodiment of the present disclosure, the step of checking the error category may include checking one of the error categories of no error, the first error category, the second error category, and the third error category based on the acquired first error type information and the second error type information, and the step of controlling may include controlling the operation of the electronic device to operate in a first mode that limits the performance of a driving device corresponding to the information indicating the state of the electronic device when the checked error category is the first error category, controlling the operation of the electronic device to operate in a second mode that limits the performance of the driving device more than the first mode when the checked error category is the second error category, and controlling the operation of the electronic device to operate in a third mode when the checked error category is the third error category.

[0018] According to one embodiment of the present disclosure, a non-transitory computer-readable recording medium storing a program for executing a control method in an electronic device includes a step of detecting information indicating a state of the electronic device, a step of inputting the information indicating the state of the electronic device into a first neural network model and a second neural network model, which are pre-trained to classify an error type of the electronic device, to thereby obtain first error type information and second error type information, a step of identifying an error category of the electronic device based on the first error type information and the second error type information, and a step of controlling an operation of the electronic device based on the error category, wherein the first neural network model is a model pre-trained using first learning data for a plurality of error types, and the second neural network model is a model pre-trained using second learning data that is obtained by processing the first learning data to distinguish between error categories to which each of the plurality of error types belongs.

[0019] The above and other aspects, features, and advantages of the embodiments of the present disclosure will become more apparent from the following description taken in conjunction with the accompanying drawings. In the accompanying drawings:

[0020] FIG. 1 is a diagram illustrating the operation of an electronic device according to one or more embodiments of the present disclosure;

[0021] FIG. 2 is a drawing for explaining the configuration of an electronic device according to one or more embodiments of the present disclosure;

[0022] FIG. 3 is a drawing for explaining another configuration of an electronic device according to one or more embodiments of the present disclosure;

[0023] FIG. 4 is a diagram illustrating an example of current information collected from a sensor according to one or more embodiments of the present disclosure;

[0024] Figure 5 is a drawing for explaining examples of normal and fault states according to examples of collected current information.

[0025] FIG. 6 is a diagram illustrating various error types according to one or more embodiments of the present disclosure;

[0026] FIG. 7 is a diagram illustrating an example of a method of processing learning data collected according to one or more embodiments of the present disclosure;

[0027] FIG. 8 is a diagram illustrating an example of a method of processing learning data collected according to one or more embodiments of the present disclosure;

[0028] FIG. 9 is a diagram illustrating an example of a method of processing learning data collected according to one or more embodiments of the present disclosure;

[0029] FIG. 10 is a diagram illustrating learning data used for each of the two neural network models according to one or more embodiments of the present disclosure;

[0030] FIG. 11 is a diagram illustrating an example of a configuration of a neural network model according to one or more embodiments of the present disclosure;

[0031] FIG. 12 is a diagram illustrating an example of a configuration of a neural network model according to one or more embodiments of the present disclosure;

[0032] FIG. 13 is a diagram illustrating an example of a configuration of a neural network model according to one or more embodiments of the present disclosure;

[0033] FIG. 14 is a drawing for explaining a control method according to one or more embodiments of the present disclosure;

[0034] FIG. 15 is a flowchart illustrating an error diagnosis method using multiple neural network models according to one or more embodiments of the present disclosure;

[0035] FIG. 16 is a diagram for explaining different control operations for each error category according to one or more embodiments of the present disclosure;

[0036] FIG. 17 is a diagram illustrating a user control command and an error category-specific control operation according to one or more embodiments of the present disclosure;

[0037] FIG. 18 is a drawing for explaining an operation example when an electronic device according to one or more embodiments of the present disclosure is a refrigerator;

[0038] FIG. 19 is a drawing for explaining an example of operation when an electronic device according to one or more embodiments of the present disclosure is an air conditioner, and

[0039] FIG. 20 is a drawing for explaining an example of operation when an electronic device according to one or more embodiments of the present disclosure is a washing machine.

[0040] The present embodiments may be modified and have various embodiments, and specific embodiments are illustrated and described in detail in the drawings. However, this is not intended to limit the scope to specific embodiments, and it should be understood that various modifications, equivalents, and / or alternatives of the embodiments of the present disclosure are included. In connection with the description of the drawings, similar reference numerals may be used for similar components.

[0041] In describing the present disclosure, if it is determined that a specific description of a related known function or configuration may unnecessarily obscure the gist of the present disclosure, a detailed description thereof will be omitted.

[0042] In one or more embodiments, the following embodiments may be modified in various other forms, and the scope of the technical concepts of the present disclosure is not limited to the following embodiments. Rather, these embodiments are provided to more faithfully and completely convey the technical concepts of the present disclosure to those skilled in the art.

[0043] The terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the scope of the rights. Singular expressions include plural expressions unless the context clearly dictates otherwise.

[0044] In this disclosure, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part, or combination thereof described in this disclosure, but do not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0045] In the present disclosure, when a component is said to be “connected,” “coupled,” “supported,” or “contacted” with another component, this includes not only cases where the components are directly connected, coupled, supported, or contacted, but also cases where the components are indirectly connected, coupled, supported, or contacted through a third component.

[0046] When we say that a component is "on" another component, this includes not only cases where the component is in contact with the other component, but also cases where there is another component between the two components.

[0047] In this disclosure, expressions such as “A or B,” “at least one of A and / or B,” or “one or more of A or / and B” can include all possible combinations of the listed items. For example, “A or B,” “at least one of A and B,” or “at least one of A or B” can all refer to instances where (1) at least one A is included, (2) at least one B is included, or (3) at least one A and at least one B are included.

[0048] The expressions “first,” “second,” “first,” or “second,” etc., used in this disclosure can describe various components, regardless of order and / or importance, and are only used to distinguish one component from another, but do not limit the components.

[0049] When it is said that a component (e.g., a first component) is “(operatively or communicatively) coupled with / to” or “connected to” another component (e.g., a second component), it should be understood that the component may be directly coupled to the other component, or may be coupled via another component (e.g., a third component).

[0050] On the other hand, when it is said that a component (e.g., a first component) is "directly connected" or "directly connected" to another component (e.g., a second component), it can be understood that no other component (e.g., a third component) exists between said component and said other component.

[0051] The expression "configured to" as used in the present disclosure may be used interchangeably with, for example, "suitable for," "having the capacity to," "designed to," "adapted to," "made to," or "capable of." The term "configured to" may not necessarily mean only "specifically designed to" in terms of hardware.

[0052] Instead, in some contexts, the phrase "a device configured to" may mean that the device, in conjunction with other devices or components, is "capable of" performing A, B, and C. For example, the phrase "a processor configured (or set) to perform A, B, and C" may refer to a dedicated processor (e.g., an embedded processor) for performing those operations, or a general-purpose processor (e.g., a CPU or application processor) that can perform those operations by executing one or more software programs stored in a memory device.

[0053] In the embodiments, a 'module' or 'part' performs at least one function or operation and may be implemented as hardware or software, or as a combination of hardware and software. In one or more embodiments, a plurality of 'modules' or a plurality of 'parts' may be integrated into at least one module and implemented as at least one processor, except for a 'module' or 'part' that needs to be implemented as a specific hardware.

[0054] According to various embodiments, operations performed by a module, program or other component may be executed sequentially, in parallel, iteratively or heuristically, or at least some operations may be executed in a different order, omitted, or other operations may be added.

[0055] The various elements and areas in the drawings are schematically drawn. Therefore, the technical concept of the present invention is not limited by the relative sizes or spacings drawn in the attached drawings.

[0056] An electronic device according to various embodiments of the present disclosure may be at least one of various types of home appliances. For example, the electronic device may include, but is not limited to, at least one of a refrigerator, a dishwasher, an electric range, an electric oven, an air conditioner, a clothes manager, a washing machine, a dryer, an air conditioner, a cleaning robot, a vacuum cleaner, and a microwave oven, as illustrated. That is, the aforementioned home appliances are merely examples, and in addition to the aforementioned home appliances, a device that performs a specific function through the operation of a motor (or heater, etc.) or a device that includes an inverter that provides power to specific electronic components (e.g., a motor, heater, steam generator), etc. may also be included in an electronic device according to one or more embodiments of the present disclosure. In one or more embodiments, the aforementioned electronic device may be a non-home appliance, such as a computer or a television. Embodiments of the present disclosure may also be applied to mobile devices, such as vehicles.

[0057] Hereinafter, with reference to the attached drawings, embodiments according to the present disclosure will be described in detail so that a person having ordinary knowledge in the technical field to which the present disclosure pertains can easily implement the present disclosure.

[0058] FIG. 1 is a diagram illustrating the operation of an electronic device according to one or more embodiments of the present disclosure.

[0059] Referring to FIG. 1, a refrigerator, an example of an electronic device (100), is illustrated. Such a refrigerator may include a main body. Below, the configuration of the refrigerator will be described first, followed by descriptions of various errors that may occur in the refrigerator and error classification methods.

[0060] The "body" may include an inner case, an outer case disposed on the outside of the inner case, and an insulating material provided between the inner case and the outer case.

[0061] The "inner case" may include at least one of a case, a plate, a panel, or a liner forming a storage compartment. The inner case may be formed as a single body, or may be formed by assembling a plurality of plates. The "outer case" may form the outer appearance of the main body, and may be joined to the outer side of the inner case so that insulation is placed between the inner case and the outer case.

[0062] "Insulation" can insulate the interior and exterior of a storage room so that the temperature inside the storage room can be maintained at a set temperature without being affected by the external environment of the storage room. According to one or more embodiments, the insulation can include foam insulation. The foam insulation can be formed by injecting and foaming urethane foam, a mixture of polyurethane and a foaming agent, between the inner and outer layers.

[0063] According to one or more embodiments, the insulation may additionally include a vacuum insulation material in addition to the foam insulation, or the insulation may consist solely of the vacuum insulation material instead of the foam insulation. The vacuum insulation material may include a core material and an outer shell material that accommodates the core material and seals the interior under a vacuum or near-vacuum pressure. However, the insulation material is not limited to the foam insulation or vacuum insulation material described above, and may include various materials that can be used for insulation.

[0064] A "storage room" may include a space defined by an interior wall. The storage room may further include an interior wall defining a corresponding space. The storage room may store various items, such as food, medicine, and cosmetics, and the storage room may be configured to be open on at least one side for the entry and exit of items.

[0065] A refrigerator may include one or more storage compartments. When a refrigerator has two or more storage compartments, each compartment may have a different purpose and be maintained at different temperatures. To achieve this, each storage compartment may be separated from the others by a partition wall containing insulation.

[0066] The storage room may be designed to maintain an appropriate temperature range depending on its intended use, and may include a "refrigerator," a "freezer," or a "variable temperature room," which are distinguished by their intended use and / or temperature range. A refrigerator may be maintained at a temperature appropriate for refrigerating items, and a freezer may be maintained at a temperature appropriate for freezing items. "Refrigeration" may mean cooling items to a temperature that does not freeze them, and for example, a refrigerator may be maintained at a temperature ranging from 0 degrees Celsius to +7 degrees Celsius. "Freezing" may mean cooling items to freeze or maintain them in a frozen state, and for example, a freezer may be maintained at a temperature ranging from -20 degrees Celsius to -1 degree Celsius. A variable temperature room may be used as either a refrigerator or a freezer, at the user's option or not.

[0067] In addition to names such as "refrigerator," "freezer," and "variable temperature room," a storage room may also be called by various other names such as "vegetable room," "fresh room," "cooling room," and "ice room." The terms "refrigerator," "freezer," and "variable temperature room" used hereinafter should be understood to encompass storage rooms having corresponding uses and temperature ranges.

[0068] According to one or more embodiments, the refrigerator may include at least one door configured to open and close an open side of a storage compartment. The door may be configured to open and close each of one or more storage compartments, or a single door may be configured to open and close multiple storage compartments. The door may be installed on the front of the main body in a pivotal or sliding manner.

[0069] The "door" may be configured to seal the storage compartment when the door is closed. The door may include insulation, similar to the body, to insulate the storage compartment when the door is closed.

[0070] According to one or more embodiments, the door may include a door outer panel forming a front of the door, a door inner panel forming a rear of the door and facing the storage compartment, an upper cap, a lower cap, and door insulation provided on the interior of these.

[0071] The door inner panel may be provided with a gasket that seals the storage compartment by contacting the front of the body when the door is closed. The door inner panel may include a dyke that protrudes rearward to accommodate a door basket for storing items.

[0072] According to one or more embodiments, the door may include a door body and a front panel detachably coupled to a front side of the door body and forming a front surface of the door. The door body may include a door outer panel forming a front surface of the door body, a door inner panel forming a rear surface of the door body and facing a storage compartment, an upper cap, a lower cap, and door insulation provided inside these.

[0073] Depending on the arrangement of the door and storage compartment, refrigerators can be classified into French door type, side-by-side type, bottom mounted freezer (BMF), top mounted freezer (TMF), or single-door refrigerator.

[0074] According to one or more embodiments, the refrigerator may include a cold air supply device arranged to supply cold air to the storage compartment.

[0075] A "cold air supply device" may include a system of machines, devices, electronic devices and / or combinations thereof that can generate cold air and guide the cold air to cool a storage room.

[0076] According to one or more embodiments, a cold air supply device can generate cold air through a refrigeration cycle including compression, condensation, expansion, and evaporation processes of a refrigerant. To this end, the cold air supply device can include a refrigeration cycle device having a compressor, a condenser, an expansion device, and an evaporator capable of driving a refrigeration cycle. According to one or more embodiments, the cold air supply device can be configured with a motor for driving the compressor and an inverter for supplying power to the motor. Alternatively, instead of the above-described compressor or the like, the cold air supply device can be configured with a semiconductor such as a thermoelectric element and an inverter for supplying power to the thermoelectric element. Such a thermoelectric element can cool a storage room through heat generation and cooling through the Peltier effect.

[0077] According to one or more embodiments, the refrigerator may include a machine room in which at least some components belonging to the cold air supply device are arranged.

[0078] The "machine room" may be designed to be partitioned and insulated from the storage room to prevent heat generated by components placed within the machine room from being transferred to the storage room. The interior of the machine room may be configured to be in communication with the exterior of the main body to dissipate heat from components placed within the machine room.

[0079] According to one or more embodiments, the refrigerator may include a dispenser provided on the door to provide water and / or ice. The dispenser may be provided on the door so that it is accessible to a user without having to open the door.

[0080] According to one or more embodiments, a refrigerator may include an ice making device configured to produce ice. The ice making device may include an ice making tray configured to store water, an ice separator configured to separate ice from the ice making tray, and an ice bucket configured to store ice produced in the ice making tray.

[0081] According to one or more embodiments, the refrigerator may include a control unit for controlling the refrigerator.

[0082] The "control unit" may include a memory that stores or memorizes a program and / or data for controlling the refrigerator, and a processor that outputs a control signal for controlling a cold air supply device, etc. according to the program and / or data memorized in the memory.

[0083] Memory stores or records various information, data, commands, programs, etc. necessary for the operation of the refrigerator. Memory can store temporary data generated during the generation of control signals for controlling components within the refrigerator. Memory may include at least one of volatile memory and non-volatile memory, or a combination thereof.

[0084] The processor controls the overall operation of the refrigerator. The processor can control components of the refrigerator by executing programs (e.g., instructions) stored in memory. The processor may include a separate NPU that performs the operations of an artificial intelligence model. In one or more embodiments, the processor may include a central processing unit (CPU), a graphics processing unit (GPU), or the like. The processor may generate control signals for controlling the operation of the cold air supply unit. For example, the processor may receive temperature information of the storage compartment from a temperature sensor and generate a cooling control signal for controlling the operation of the cold air supply unit based on the temperature information.

[0085] In one or more embodiments, the processor may process user input of a user interface and control the operation of the user interface based on programs and / or data stored / stored in memory. The user interface may be provided using an input interface and an output interface. The processor may receive user input from the user interface. In one or more embodiments, the processor may transmit display control signals and image data to the user interface for displaying an image on the user interface in response to the user input.

[0086] In one or more embodiments, the processor can diagnose whether each component within the electronic device has an error (or is malfunctioning). Specifically, the processor can use current information measured from a specific component to determine whether the component has an error. In one or more embodiments, the current information may be information indicating the status of the electronic device. For example, the current information may indicate an error code associated with a component of the electronic device (e.g., a compressor). In one or more embodiments, the current information may provide an operating value range of the component of the electronic device, and the processor can refer to a table of operating values ​​to determine whether the component is operating properly. In one or more embodiments, the processor can limit or suspend operations for the corresponding component. Specific diagnostic methods will be described later.

[0087] In one or more embodiments, the fault may be based on a hardware abnormality of a particular configuration (open circuit, short circuit, failure of a particular component), a software abnormality of a particular configuration, etc., and such an error may also be referred to as an error or a failure.

[0088] The processor and memory may be provided as a single unit or separately. The processor may include one or more processors. For example, the processor may include a main processor and at least one subprocessor. The memory may include one or more memories.

[0089] According to one or more embodiments, a refrigerator may include a processor and memory that control all components included in the refrigerator, and may include multiple processors and multiple memories that individually control the components of the refrigerator. For example, the refrigerator may include a processor and memory that control the operation of a cooling air supply device based on the output of a temperature sensor. In one or more embodiments, the refrigerator may separately include a processor and memory that control the operation of a user interface based on user input.

[0090] The communication module can communicate with external devices, such as servers, mobile devices, and other home appliances, via a nearby access point (AP). The AP can connect the local area network (LAN) where the refrigerator or user device is connected to the wide area network (WAN) where the server is connected. The refrigerator or user device can then connect to the server via the WAN.

[0091] The communication module can output information about an error (or failure) in a specific configuration within the electronic device to another electronic device (e.g., a user terminal, a home server, etc.).

[0092] The input interface may include keys, a touchscreen, a microphone, or other input devices known to those skilled in the art. The input interface may receive user input and transmit it to the processor.

[0093] The output interface may include a display, speaker, or other output device known to those skilled in the art. The output interface may output various notifications, messages, information, etc. generated by the processor. For example, the output interface may output information about an error (or failure) identified in a specific component within the electronic device.

[0094] In the above, the components within a refrigerator have been described assuming that the electronic device (100) is a refrigerator. However, as previously explained, the electronic device of the present invention may be a device other than a refrigerator. Therefore, if the electronic device (100) is implemented as a device other than a refrigerator, it is self-evident that the components of the other device may be implemented instead of the components of the refrigerator described above.

[0095] Below, a diagnostic method using the present disclosure is described.

[0096] As described above, the refrigerator operates the compressor by rotating the motor, and the motor is operated by the driving power supplied by the inverter. Therefore, the output current of each component within the inverter (or the current in a specific loop) can be measured and analyzed to determine whether the inverter (or motor) is faulty. If the refrigerator operates the compressor with a three-phase motor, the driving power for the three-phase motor can be generated through the operation of six switching elements. The detailed configuration and operation thereof are described later in Fig. 4.

[0097] The following description assumes a case where a three-phase motor and six switch elements are used, but the following method can be modified and applied when a motor other than a three-phase motor is used or when a different number of switch elements is used.

[0098] Faults occurring in inverters (or motors) can generally be classified into one of the following four exemplary categories. Those skilled in the art will recognize that the number of categories is not limited to four and can include any number of fault categories. The number of categories may vary depending on the type of electronic device.

[0099] In one or more embodiments, the first error category is due to a sensor anomaly, and there may be, for example, two types of such errors. For example, the first error is an error (scale failure) in which a current value higher than an experimentally or computationally input current value (e.g., a current value 1.2 times the applied current) is measured. The second error is an error (i.e., offset failure) in which the measured peak value of the current sensor is measured higher than a preset value (e.g., +0.25 A) than in a normal state, and the offset of the measured peak current value changes. Errors due to such sensor anomalies are referred to as the first error category.

[0100] In one or more embodiments, the second fault category is a failure of a switching element within the inverter to turn on. Since there are six switching elements as described above, for example, there are six types of faults (i.e., each fault corresponds to a faulty switching element). These fault types are hereinafter referred to as the second fault category. However, since there are multiple switching elements within such an inverter, even if only one switching element fails, operation is still possible, but an overcurrent may occur. However, if multiple switching elements fail, this may be classified as a separate fault category. As will be understood by those skilled in the art, an overcurrent may refer to an electrical condition that occurs when the normal load current of a circuit exceeds the rated current of the equipment or the current capacity of the conductor. An overcurrent may also be caused by an overload, a short circuit, a ground fault, etc.

[0101] In one or more embodiments, the third error category is an error that occurs when there is an abnormality in the current transmission between the inverter and the motor, such as an open circuit between the motor and the inverter or an open circuit within the motor. There are, for example, three types of such errors (i.e., each error corresponds to a line of each of the three phases of the motor), and these types of errors are referred to as the third error category below.

[0102] In one or more embodiments, the fourth error category is one in which the switching elements within the inverter do not turn off. Since there are six switching elements as described above, there may be six types of errors, and these types of errors are referred to as the fourth error category below. Unlike the second error category, these types of errors require immediate fault control in that even if only one switching element fails, an overcurrent is immediately possible.

[0103] In one or more embodiments, the error type is classified based on an anomaly or error in a specific configuration, and the error category refers to a set of error types that have different causes (i.e., configurations) but must be handled in the same manner or with the same corresponding action. These error types may be referred to as error numbers, fault numbers, or any other suitable error diagnostics, and the numbers and types described above may be utilized in various modified forms.

[0104] Error categories may be classified based on the configuration (or manner) in which the error occurred, as described above, but they may also be grouped based on how the error occurred and how it was dealt with, as described above.

[0105] According to one or more embodiments, although the above-described error types are divided into four error categories, during implementation, the error categories may be divided according to the degree of risk of the error. For example, the first risk type may be divided into a case where the degree of risk is low, the second risk type may be divided into a case where the degree of risk is medium, and the third risk type may be divided into a case where the degree of risk is high and requires immediate operation interruption. In such a case, the first and second error types may be divided into the first risk type, the third to eleventh error types may be divided into the second risk type, and the twelfth to seventeenth error types may be divided into the third risk type. In one or more embodiments, the risk types may also be referred to as error states or risk categories, failure categories, failure types, or other suitable failure designations.

[0106] Among the errors described above, the first and second error types corresponding to the first error category are due to an abnormality in the sensor, and there is no problem with the inverter or motor that directly plays a role in the driving operation. Therefore, it is an abnormality that does not cause a great problem even if normal operation is performed.

[0107] However, except for the first error category, the remaining error categories are problems with the inverter or motor that directly play a role in the driving operation, and problems may occur during normal operation.

[0108] In particular, the fourth fault category is a more dangerous (or important) fault than the second or third fault categories in that it is a fault that can cause the motor to overheat because the switch element is always on.

[0109] While each error type does not require a precise distinction within the same error category, in one or more embodiments, different error categories carry different levels of risk. Therefore, accurately classifying an error into which error type it belongs is beneficial for providing accurate information about the operating status of an electronic device.

[0110] However, the errors described above are often the first error category in actual occurrence cases, and accordingly, the proportion of learning data belonging to the first error category is very high in the learning data collected for fault diagnosis.

[0111] When training a neural network model under the aforementioned conditions, failure-related data may not be collected uniformly, potentially leading to errors belonging to the second (or third) error category being classified as the first error category. Conversely, errors belonging to the first error category may also be classified as the second (or third) error category.

[0112] However, as described above, errors corresponding to the first error category and errors corresponding to the second error category (or third error category) have different degrees of failure, and thus, there is a difference in subsequent control operations. Therefore, a method for accurately distinguishing between the first error category and the second error category (or third error category) was required.

[0113] If, as described above, the classification of error categories cannot be performed with certainty, the electronic device must always operate under the assumption of the worst case scenario, which causes inconvenience to the user of the electronic device in that the device itself becomes inoperable even when normal operation is not impaired due to a sensor abnormality.

[0114] To address these issues, the present disclosure accurately distinguishes the above-described error types (or error categories) by using multiple neural network models trained with different learning data.

[0115] For example, the first neural network model used in the present disclosure is a model trained using training data for multiple error types, and the second neural network model is a model trained using second training data, which is obtained by processing the first training data used to train the first neural network model, to distinguish between error categories to which each of the multiple error types belongs. Specifically, the second neural network model is a model that can distinguish between the first error category and the second error category (or the third error category) more accurately than the first neural network model.

[0116] In one or more embodiments, 'learning data' refers to data used for learning the neural network model described above, and may be actual measured data, simulation data, etc. In the following, learning data is processed in various ways, and for the purpose of distinction, the expression 'learning data' is used in the above-described meaning, data obtained through the collection of various cases is referred to as 'raw data', data prepared to be used for learning the first neural network model using raw data is referred to as 'first learning data', and data processed from the first learning data to be used for learning the second neural network model is referred to as 'second learning data'.

[0117] In the above, the method of identifying the error type of an error using a neural network model and performing a control operation based on the error category belonging to the identified error type has been described. The reason for this is that, as described above, the form of the current information measured when the first switch element has a non-turn-off error is different from the form of the current information measured when the second switch element has a non-turn-off error. Therefore, in one or more embodiments, a method of first identifying the error type and then identifying the error category to which the identified error type belongs has been used.

[0118] However, in one or more embodiments, instead of training a neural network model to distinguish error types, a method may be used to train a neural network model by merging training data by error category. In other words, a neural network model may be trained so that its output is not a probability value for each error type, but rather a probability value for a specific error category, and such a neural network may be used.

[0119] By utilizing these two neural network models, the electronic device can accurately identify errors occurring within the electronic device, identifying the type (or category) of errors. The specific configuration and operation of this electronic device (100) are described below with reference to FIGS. 2 and 3.

[0120] Meanwhile, the above description only describes the diagnosis of motor or inverter failures. However, since errors can occur in various devices or components other than the aforementioned inverter (or the motor driving the compressor), the electronic device (100) can also perform error diagnosis for various components by providing a neural network model corresponding to each component as described above.

[0121] FIG. 2 is a drawing for explaining the configuration of an electronic device according to one or more embodiments of the present disclosure.

[0122] Referring to FIG. 2, the electronic device (100) may include a plurality of sensors (110), an input device (160), a memory (120), and a processor (130).

[0123] In one or more embodiments, the sensor (110) may detect a plurality of current information within the electronic device (100). The plurality of current information may include information indicating the state of the electronic device (100). For example, current values ​​output from a plurality of nodes (or loops) within an inverter within the electronic device (100) may be detected. In one or more embodiments, the current information may be information indicating whether current is conducted, and may be a current value at a corresponding node (or loop). In implementation, the current value described above may be the current value itself, but may also be a converted value (scaled value), or a voltage value of a specific resistor used to measure the current value may be used.

[0124] In one or more embodiments, the sensing operation of the sensor (110) may be performed based on a control command of the processor (130) described below, and may be automatically measured in preset cycle units and provided to the processor (130). In one or more embodiments, each sensor information (or sensing value) may be information measured at the corresponding moment, or an average value of the corresponding cycle unit may be used. In one or more embodiments, the preset cycle unit described above may be 20 ms, but is not limited thereto.

[0125] While the above description uses only current information, in one or more embodiments, voltage information other than current information may be used. Furthermore, while one or more embodiments describe a single sensor detecting multiple current information, multiple sensors may be used during implementation.

[0126] The memory (120) may store at least one instruction regarding the electronic device (100). In one or more embodiments, the memory (120) may store an O / S (Operating System) for driving the electronic device (100). Such instructions may include instructions for identifying an error type described below, instructions for controlling the operation of the electronic device according to the identified error type, instructions for controlling various components of the electronic device, instructions for controlling the configuration of the electronic device in an error state, instructions for storing information collected from a sensor as learning data, and other suitable information.

[0127] The memory (120) may include a semiconductor memory such as a flash memory or a magnetic storage medium such as a hard disk. For example, various software modules for operating the electronic device (100) according to various embodiments of the present disclosure may be stored in the memory (120), and the processor (130) may execute various software modules stored in the memory (120) to control the operation of the electronic device (100). That is, the memory (120) is accessed by the processor (130), and data reading / writing / modifying / deleting / updating, etc. may be performed by the processor (130).

[0128] According to one or more embodiments, the term memory (120) may be used to mean memory (120), ROM, RAM within the processor (130), or a memory card (e.g., micro SD card, memory stick) mounted on the electronic device (100).

[0129] The memory (120) may store the first and second neural network models that have been trained. For example, the neural network models may be implemented as a convolutional neural network (CNN), a long short-term memory (LSTM), a deep neural network (DNN), a recurrent neural network (RNN), a restricted boltzmann machine (RBM), a deep belief network (DBN), or a bidirectional recurrent deep neural network (BRDNN). Those skilled in the art will appreciate that the present embodiment is not limited to these examples, and may include any appropriate neural network or machine learning module known to those skilled in the art.

[0130] These neural network models are computing systems inspired by the neural networks of human or animal brains, and may also be referred to as learning models, machine learning models, artificial intelligence models, deep learning models, or other suitable machine learning models.

[0131] In one or more embodiments, two neural network models are used. The two neural network models may differ only in the training data used for training, and may have the same internal model. The two neural network models may also differ in the training data and the models themselves. For example, both the first and second neural network models may be CNN-based, or the first neural network model may be CNN-based, and the second neural network model may be DNN-based. This example is not an example, and various combinations of methods may be used.

[0132] The learning data used in the two neural network models is described below with reference to FIGS. 7 to 10, and an implementation example of the neural network model is described in detail in FIGS. 11 to 15.

[0133] In one or more embodiments, the memory (120) may store information collected from the sensor. Specifically, the information collected in this manner may be used to train the neural network model described above. Accordingly, if an error is detected, the current information used to detect the error may be stored in the memory (120).

[0134] The processor (130) controls the overall operation of the electronic device (100). For example, the processor (130) can control the overall operation of the electronic device (100) by executing at least one instruction stored in the memory (120) as described above.

[0135] The processor (130) may be composed of one or more processors. In one or more embodiments, the one or more processors (130) may include at least one of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and an NPU (Neural Processing Unit), but is not limited to the examples of the processor (130) described above.

[0136] A CPU is a general-purpose processor capable of performing not only general calculations but also artificial intelligence calculations. Its multi-layered cache structure allows for the efficient execution of complex programs. CPUs are advantageous for serial processing, allowing for the organic linking of previous and subsequent calculation results through sequential calculations. A general-purpose processor is not limited to the examples described above, except where specifically designated as a CPU.

[0137] A GPU is a processor for large-scale calculations, such as floating-point operations used in graphics processing, and can perform large-scale calculations in parallel by integrating a large number of cores. In particular, a GPU may be advantageous compared to a CPU in parallel processing methods, such as convolution operations. In one or more embodiments, a GPU may be used as a co-processor (130) to supplement the functions of a CPU. The processor for large-scale calculations is not limited to the examples described above, except in cases where it is specifically referred to as a GPU.

[0138] An NPU is a processor specialized in artificial intelligence operations using artificial neural networks, and each layer constituting the artificial neural network can be implemented in hardware (e.g., silicon). In one or more embodiments, the NPU is designed specifically according to the company's required specifications, so it has less freedom than a CPU or GPU, but can efficiently process the artificial intelligence operations requested by the company. Meanwhile, as a processor specialized in artificial intelligence operations, the NPU can be implemented in various forms, such as a Tensor Processing Unit (TPU), an Intelligence Processing Unit (IPU), or a Vision Processing Unit (VPU). The artificial intelligence processor is not limited to the examples described above, except in cases where it is specifically referred to as an NPU.

[0139] In one or more embodiments, one or more processors (130) may be implemented as a System on Chip (SoC). In this case, the SoC may further include, in addition to one or more processors (130), a memory (120), and a network interface such as a bus for data communication between the processor (130) and the memory (120).

[0140] When a plurality of processors (130) are included in a SoC (System on Chip) included in an electronic device (100), the electronic device (100) may perform operations related to artificial intelligence (e.g., operations related to learning or inference of an artificial intelligence model) by using some of the plurality of processors (130). For example, the electronic device (100) may perform operations related to artificial intelligence by using at least one of a GPU, an NPU, a VPU, a TPU, and a hardware accelerator specialized in artificial intelligence operations such as convolution operations and matrix multiplication operations among the plurality of processors (130). However, this is merely an example, and it is of course possible to process operations related to artificial intelligence by using a CPU or a general-purpose processor (130).

[0141] In one or more embodiments, the electronic device (100) may perform operations related to functions related to artificial intelligence by utilizing multiple cores (e.g., dual cores, quad cores, etc.) included in one processor (130). In particular, the electronic device (100) may perform artificial intelligence operations, such as convolution operations and matrix multiplication operations, in parallel by utilizing multiple cores included in the processor (130).

[0142] One or more processors (130) are controlled to process input data according to predefined operation rules or artificial intelligence models stored in the memory (120). The predefined operation rules or artificial intelligence models are characterized by being created through learning.

[0143] In one or more embodiments, "created through learning" means that a predefined set of operating rules or an artificial intelligence model with desired characteristics is created by applying a learning algorithm to a plurality of learning data. This learning may be performed on the device itself, where the artificial intelligence according to the present disclosure is performed, or may be performed through a separate server / system. For example, the learning model may be pre-trained by a remote server using data from multiple electronic devices. The pre-trained learning model may be downloaded to the electronic device (100 of FIG. 1). Here, the learning model may be further refined based on specific data regarding the electronic device (100).

[0144] An artificial intelligence model may be composed of multiple neural network layers. At least one layer has at least one weight value and performs its operation through the operation result of the previous layer and at least one defined operation. Examples of neural networks include a convolutional neural network (CNN), a deep neural network (DNN), a recurrent neural network (RNN), a restricted boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), deep Q-networks, and a transformer. The neural networks in the present disclosure are not limited to the above-described examples unless otherwise specified.

[0145] In one or more embodiments, a learning algorithm is a method for training a given target device (e.g., a robot) using a plurality of learning data sets so that the given target device can make decisions or predictions on its own. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. However, the learning algorithm of the present disclosure is not limited to the examples described above, and any suitable learning algorithm known to those skilled in the art may be included.

[0146] In one or more embodiments, the processor (130) may use current information to identify error types. Specifically, the processor (130) may use multiple modules to perform processes related to various embodiments of the present disclosure. This will be described below with reference to FIGS. 11 to 13 .

[0147] The plurality of modules may be implemented as hardware modules or software modules, and at least some of the modules may include a neural network model. For convenience of explanation, the following description will assume that all of the plurality of modules are implemented through the memory (120) and processor (130) of the electronic device (100). However, depending on the embodiment, at least some of the plurality of modules may be implemented by an external device or server.

[0148] The processor (130) can acquire error types using two neural network models and determine an error status based on the acquired error types. For example, the processor (130) can input current information into a first neural network model to acquire a first error type, input current information into a second neural network model to acquire a second error type, and determine the error type of the electronic device based on the acquired first and second error types.

[0149] In one or more embodiments, the processor (130) may use two neural network models sequentially or simultaneously. For example, the processor (130) may input current information into a first neural network model to obtain a first error type, determine whether a second neural network model needs to be used based on the obtained first error type, and if the second neural network model needs to be used, input current information into the second neural network model to obtain a second error type, and perform confirmation of the final error type based on the obtained second error type. An example of this is described in detail in FIG. 12, and an example of simultaneous use is described later in FIG. 11.

[0150] In one or more embodiments, the processor (130) controls the operation of the electronic device based on the identified error category (or error condition). For example, if no error condition is identified, the processor (130) may operate in a normal mode. This normal mode may be referred to as a normal mode, a normal mode, a normal state, a normal state, etc.

[0151] In one or more embodiments, when an error condition corresponding to the first error category is identified, the processor (130) may be set to operate in a first mode that limits and utilizes the performance of the driving device corresponding to the current information. This first mode may be referred to as a first safety mode, a first abnormality mode, a first error mode, a first safe operation mode, etc., and is a mode that utilizes a configuration related to the above-described error, but can operate within a range that does not impose a large load on the operation of the configuration related to the error. In one or more embodiments, the processor (130) may notify the user of the occurrence of the above-described error.

[0152] In one or more embodiments, if an error condition corresponding to the second error category is identified, the processor (130) may be set to operate in a second mode that further limits the performance of the driving device corresponding to the current information. This second mode may be referred to as a second safety mode, a second abnormality mode, a second error mode, a second safe operation mode, etc.

[0153] When operating in this second mode, the processor (130) may notify the user of the error status and operate in the second mode by limiting the performance of the driving device according to the user's selection, or may operate in a state in which the driving device is not used.

[0154] In one or more embodiments, if an error condition corresponding to the third error category is identified, the processor (130) may be set to a stop mode so that the driving device corresponding to the current information does not operate. At this time, the processor (130) may notify the user that the error has occurred.

[0155] Although the above description describes performing a control operation corresponding to the immediately identified error category, in one or more embodiments, the processor (130) may determine an intermediate error state and then perform a control operation based on the error state. For example, if the first error category is identified, the processor (130) may determine that it is a first-stage error state. If the second or third error category is identified, the processor (130) may determine that it is a second-stage error state that is more serious than the first-stage error state, and if the fourth error category is identified, the processor (130) may determine that it is a very serious error state.

[0156] As described above, the electronic device according to one or more embodiments utilizes a neural network model for general error classification along with a neural network model with enhanced error category discrimination performance, enabling more accurate classification of error types. Accordingly, adaptive operation is also facilitated based on the identified error types.

[0157] Although only a simple configuration of the electronic device (100) is illustrated above, various other configurations may be included in one or more embodiments. This will be described below with reference to FIG. 3.

[0158] FIG. 3 is a drawing for explaining another configuration of an electronic device according to one or more embodiments of the present disclosure.

[0159] Referring to FIG. 3, the electronic device (100) may include a sensor (110), a memory (120), a processor (130), a driving device (140), a communication device (150), an input device (160), and a display (170).

[0160] The sensor (110), memory (120), and processor (130) have been previously described in FIG. 2, and only the parts that differ from the operations described above will be described below.

[0161] The driving device (140) is a configuration for performing a specific function (or service) of the electronic device. If the electronic device (100) is a device that performs a specific function through rotational motion, the driving device (140) may include a motor and an inverter.

[0162] For example, if the electronic device is a refrigerator or an air conditioner, the driving device (140) may include a motor for rotating a compressor, etc., and an inverter for providing driving power to the motor. In one or more embodiments, if the electronic device is a washing machine, the driving device may include a motor for rotating a drum, and an inverter for supplying driving power to the motor. In one or more embodiments, if the electronic device is a microwave oven, the driving device may include a generator for generating electromagnetic waves, and an inverter for supplying power to the generator, and the electronic device may include a magnetic field generating coil, and an inverter for supplying power to the coil. In the following, for ease of explanation, it is assumed that the driving device includes a motor and an inverter.

[0163] The motor may be configured for rotation and may be a DC motor, a BLDC motor, or a 3-phase motor.

[0164] An inverter can provide driving power to a three-phase motor. The following description assumes that driving power is provided to a three-phase motor, and thus uses an example of six switching elements. However, if a configuration other than the three-phase motor described above is used, a circuit configuration corresponding to that configuration may be used. The specific configuration and operation of the inverter are described with reference to Fig. 4.

[0165] The sensor (110) can measure multiple current values ​​within each loop of the above-described inverter.

[0166] The communication device (150) is formed to connect the electronic device (100) to an external device (specifically, a terminal device, a home server, an external server, etc.), and can be connected by a short-range wireless communication method (e.g., Bluetooth, WiFi, WiFi Direct) as well as a long-range wireless communication method (e.g., wireless communication such as GSM, UMTS, LTE, WiBRO, etc.).

[0167] The communication device (150) may include at least one of a WiFi module, a Bluetooth module, a wireless communication module, an NFC module, and a UWB (Ultra-Wide Band) module. Specifically, the WiFi module and the Bluetooth module may each perform communication in the WiFi or Bluetooth manner. When using a WiFi module or a Bluetooth module, various connection information, such as an SSID, may be first transmitted and received, and then communication may be established using this, after which various pieces of information may be transmitted and received.

[0168] In one or more embodiments, the wireless communication module may perform communication according to various communication standards, such as IEEE, Zigbee, 3G (3rd Generation), 3GPP (3rd Generation Partnership Project), LTE (Long Term Evolution), 5G (5th Generation), etc. In one or more embodiments, the NFC module may perform communication in an NFC (Near Field Communication) manner using a 13.56 MHz band among various RF-ID frequency bands, such as 135 kHz, 13.56 MHz, 433 MHz, 860-960 MHz, 2.45 GHz, etc.

[0169] The communication device (150) can receive a neural network model learned from an external device, receive learning data required for learning the neural network model, or transmit learning data collected from the electronic device (100) to the external device.

[0170] Alternatively, when utilizing a neural network model stored on an external server, the communication device (150) may transmit current information collected from the sensor and / or user operation commands to the external server. In response, the communication device (150) may receive error information identified from the external server.

[0171] When an error is identified in the electronic device, the communication device (150) can transmit information about the identified error to an external device (e.g., a user terminal device, a home server, or a manufacturer server).

[0172] In one or more embodiments, the communication device (150) may receive various control commands for the electronic device through an external device (e.g., a user terminal device).

[0173] The input device (160) receives user commands. These user commands may be control commands for controlling the functions of the electronic device. For example, if the electronic device is a refrigerator, these may be temperature settings for the refrigerator compartment and the freezer compartment. Such control commands may also be input via the communication device (150), as previously described.

[0174] In one or more embodiments, the input device (160) may receive an operating state of the electronic device in response to an error. For example, in the case of the first error category, which is a state in which limited operation is possible, the input device (160) may receive a user setting as to whether to perform a general operation, perform an operation, or perform a limited operation in the state of the first error category. In one or more embodiments, the input device (160) may also receive a setting as to whether to perform a limited operation or not to perform an operation for a configuration in which an error has occurred in the state of the second error category.

[0175] In one or more embodiments, such limitations in driving performance may be related to the driving device causing the error. For example, if the electronic device (100) operates the refrigerator and freezer compartments as separate compressors, and an error occurs only in the motor (or inverter) associated with the refrigerator compartment, the electronic device (100) may perform the limited operations described above only for the refrigerator compartment where the error occurred, and may operate in a normal mode for the freezer compartment where the error did not occur.

[0176] The display (170) can display various types of information supported by the electronic device (100). In one or more embodiments, the display (170) may be a display such as an LCD, and may also be implemented as a touch screen that can perform the functions of the input device described above.

[0177] The display (170) may display the operating status of the electronic device (100). In one or more embodiments, the display (170) may display information on identified error information and methods for resolving the error (e.g., AS contact information, etc.).

[0178] The display (170) can display various functions that can be performed on the electronic device. In one or more embodiments, when the processor (130) is operating in the first mode or the second mode, it can display only the functions that can be performed, excluding the functions that cannot be performed in the first mode or the second mode. In one or more embodiments, when a control command for a function that cannot be performed in the first mode or the second mode is input from the user, the processor (130) can control the display (170) to display a message notifying that the function cannot be performed.

[0179] While various configurations that the electronic device (100) may include are illustrated and described in FIG. 3, some of the aforementioned configurations may be omitted during implementation, and other configurations not illustrated may be additionally provided. For example, a microphone, speaker, or other suitable configuration may be further included to receive voice input of control commands from a user, or to output various information about the vacuum cleaner as audio.

[0180] FIG. 4 is a diagram illustrating an example of current information collected from a sensor according to one or more embodiments of the present disclosure. For example, FIG. 4 is a diagram illustrating a case where an electronic device uses a three-phase motor and six switching elements to drive the three-phase motor.

[0181] Referring to FIG. 4, the driving device (140) includes a three-phase motor (141) and an inverter circuit (145). The configuration of the driving device is based on the assumption that the driving device operates a compressor within a refrigerator. If the electronic device (100) is a device other than a refrigerator or is intended to drive a component other than a compressor, the configuration within the driving device (140) may vary.

[0182] The three-phase motor (141) receives driving power through three lines. Accordingly, the three-phase motor (141) rotates by the driving power and can provide the corresponding rotational power to, for example, a compressor.

[0183] The inverter circuit (145) includes six switching elements. Specifically, two switching elements connected in series constitute one switching unit, and the structure has three such switching units arranged in parallel. Accordingly, a first driving power (Va) can be generated in the first line (i.e., the first switching unit), a second driving power (Vb) can be generated in the second line (i.e., the second switching unit), and a third driving power (Vc) can be generated in the third line (i.e., the third switching unit). The first to third driving power sources can have AC waveforms whose phases are 120 degrees different from each other.

[0184] The sensor (110) can measure current values ​​at multiple locations within the inverter described above. For example, the sensor (110) can be i ap , i bp , i cp , i am , i bm , i cm The current value corresponding to each can be measured. In one or more embodiments, i ap , i bp , i cp is the phase current flowing into the motor at nodes a, b, and c respectively, and i am , i bm , icm can be the current flowing from the motor to ground at each of nodes a, b, and c. The form of each current value is described below with reference to Fig. 5.

[0185] Figure 5 is a diagram for explaining examples of normal and fault states according to examples of collected current information.

[0186] Referring to FIG. 5, the current values ​​of the first group among the current values ​​of the above-described sensors (e.g., i ap , i bp , i cp ) and the current value of the second group (e.g., i am , i bm , i cm) are symmetrical in a normal state, as shown in Fig. 5. Specifically, in the normal waveform (510), the upper waveform is the current values ​​of the first group, and the lower waveform is the current values ​​of the second group.

[0187] However, if we look at the fault condition, we can see that there are cases where some of the six values ​​converge to a value of 0. In addition to this case, in other fault conditions, there are cases where the waveform shape is similar but the size is different, so it is difficult to determine whether it is normal or an error simply by looking at whether the current waveforms of the two groups are symmetrical or not, as described above.

[0188] In this regard, one or more embodiments of the present disclosure classify errors using a neural network model.

[0189] The above describes 17 error types used in an environment such as Fig. 4.

[0190] In one or more embodiments, the first error is a scale error, in which the measured current value is measured to be a predetermined value (e.g., 1.2 times) greater than the applied current value. The second error is an offset error, in which the peak value of the measured current is measured to be a predetermined value (e.g., 0.25 A) greater than the steady-state peak value. These first and second errors can be classified into the first error category.

[0191] In one or more embodiments, the third through eighth errors are non-turn-on abnormalities for each of the six switch elements described above, the third error is an abnormality in the first switch element, the fourth error is an abnormality in the second switch element, the fifth error is an abnormality in the third switch element, the sixth error is an abnormality in the fourth switch element, the seventh error is an abnormality in the fifth switch element, and the eighth error is an abnormality in the sixth switch element. These third through eighth errors can be classified into the second error category.

[0192] In one or more embodiments, the ninth through eleventh errors are abnormalities in which the connection between the inverter and the motor is broken. The ninth error is when the connection between node A and the motor is broken, the tenth error is when the connection between node B and the motor is broken, and the eleventh error is when the connection between node C and the motor is broken. These ninth through eleventh errors can be classified into the third error category.

[0193] Errors 12 through 17 are non-stop abnormalities for each of the six switch elements described above. Error 12 is an abnormality in the first switch element, error 13 is an abnormality in the second switch element, error 14 is an abnormality in the third switch element, error 15 is an abnormality in the fourth switch element, error 16 is an abnormality in the fifth switch element, and error 17 is an abnormality in the sixth switch element. These errors 12 through 17 can be classified into the fourth error category.

[0194] The error occurrence rates of each of the first to 17th errors are as shown in Figure 6.

[0195] The above-described category classification assumes that a single error type occurs within the same category. Therefore, if multiple errors are detected simultaneously for each of the second or third error categories, such cases may be classified as a separate fifth error category (or sixth error category) rather than the second or third error categories described above. In one or more embodiments, the same treatment may be applied to cases where one error type in the second error category and one error type in the third error category are detected simultaneously.

[0196] FIG. 6 is a diagram illustrating various error types according to one or more embodiments of the present disclosure.

[0197] Referring to Figure 6, it can be seen that the error types corresponding to the first error category have a high frequency, and in particular, the error types corresponding to the fourth error category have a very low frequency.

[0198] In particular, the fourth error category is an error type that can occur in an actual usage environment, but collecting data on that error type is difficult.

[0199] As the frequency of the first error type is too high, and the fourth error category cannot be collected through actual data, the present disclosure performs certain processing on the collected raw data. This will be described below with reference to FIGS. 7 through 10.

[0200] FIG. 7 is a diagram illustrating an example of a method for processing learning data collected according to one or more embodiments of the present disclosure.

[0201] Referring to Fig. 7, examples of undersampling and oversampling operations for non-uniform data are described.

[0202] Undersampling is a method of reducing the number of data (712) acquired in large numbers among the collected data (710) to match the number of data (711) acquired in small numbers, thereby equalizing the total number of data (711, 722). Specifically, in order to acquire data evenly, the average value of the data may be acquired, and among the acquired data in large numbers, only data similar to the average may be selected according to the target number, and the rest may not be selected.

[0203] Conversely, the oversampling operation is a method of augmenting the number of data (732) acquired as a minority among the collected data (730) to match the number of data acquired as a minority (731), thereby equalizing the total number of data (741, 742). In one or more embodiments, the number of data acquired as a minority may be simply duplicated to be the same as the number acquired as a majority.

[0204] In the present disclosure, through the above-described undersampling operation on raw data, the types of errors corresponding to the first error category can be reduced, and the total number of errors in the first and second error categories (or third error categories) can be made approximately equal. For example, the total number of errors corresponding to the first and second error categories can be made within 5% of the total number of errors or within another appropriate threshold.

[0205] Alternatively, the error type corresponding to the first error category among raw data may be undersampled with median value (overall average count) information, and the data corresponding to the second error category (or third error category) among raw data may be oversampled.

[0206] In one or more embodiments, first training data for the first neural network model may be prepared by generating training data for a fourth error category, as described below in FIG. 9, and adding it to the previously oversampled (and / or undersampled) results.

[0207] Below, the process of preparing second learning data for a second neural network model, when first learning data for a first neural network model is prepared, is described.

[0208] FIG. 8 is a diagram illustrating an example of a method for processing learning data collected according to one or more embodiments of the present disclosure.

[0209] Referring to Figure 8, data corresponding to the first error category in the first neural network model and data corresponding to the second error category (or third error category) are sometimes adjacent to each other. Due to this proximity, there may be instances where data corresponding to the second error category is judged as an error corresponding to the first error category.

[0210] Therefore, to better distinguish between the first error category and the second error category (or the third error category), the first training data is processed in the following manner.

[0211] First, in one or more embodiments, when the first learning data (810) is prepared in the same manner as in FIG. 7, data less than a certain distance value can be identified using the Euclidean distance value between the first error category and the second error category data (820), and data within the identified certain distance can be removed from the first learning data (830).

[0212] In one or more embodiments, the Euclidean distance value is used, but other distance values, such as Manhattan Distance and Minkowski Distance, may be used during implementation. In one or more embodiments, various methods other than the calculation method described above may be applied to clearly distinguish between categories.

[0213] In one or more embodiments, the electronic device may perform a process of identifying principal components (or axial directions, or central coordinates) of data corresponding to a first error category and data corresponding to a second error category, and removing data located at the central portion.

[0214] In one or more embodiments, after some data has been deleted, some data of the first error category may be removed, and augmented data (840) corresponding to the first error category may be added. In one or more embodiments, the augmented data may be modified original data, in order to increase the number of data samples.

[0215] In one or more embodiments, the preparation of the first and second learning data and the training of the neural network model may be performed on the electronic device (100) side described above, but may also be performed on an external device such as a manufacturer's server.

[0216] FIG. 9 is a diagram illustrating an example of a method for processing learning data collected according to one or more embodiments of the present disclosure.

[0217] Referring to Figure 9, an error in which a portion of a switch element remains on makes it difficult to obtain data corresponding to the actual error. In other words, data corresponding to the fourth error category may not be available in the raw data.

[0218] However, since these types of errors are errors that cause overload to the motor, it is essential to be able to detect and classify them.

[0219] In one or more embodiments, the error type may be considered to be a case where a specific current value has an extreme value. As illustrated in FIG. 9, data for the fourth error category may be generated by creating a data set in which one of the multiple current values ​​has an extreme value.

[0220] In one or more embodiments, the fourth error category may have extreme values, as described above, which may affect the overall feature average depending on the use of the data. Therefore, for the fourth error category, only a minimal number of values ​​may be used, as described in FIG. 7, without the need for matching the number of values ​​similar to other error types.

[0221] In one or more embodiments, when implemented, the learning operation is performed using only one data type, and then it is confirmed whether the neural network model operates normally, and if the classification for the fourth error category is not performed well, the number of data for the fourth error category may be gradually increased so that only the minimum number is used.

[0222] FIG. 10 is a diagram illustrating learning data used for each of the two neural network models according to one or more embodiments of the present disclosure.

[0223] Referring to FIG. 10, the number of error types in the first learning data (1010) used for learning the first neural network model is shown, and the number of error types in the second learning data (1020) used for learning the second neural network model is shown.

[0224] In Fig. 6, the data ratio corresponding to the first error category is greater than that of the second error category (or the third error category), but by using the oversampling or undersampling described above, it can be confirmed that the number of errors between the first error category to the third error category in the first learning data (1010) is similar.

[0225] As described above, the data corresponding to the 4th error category has extreme values, so it can be confirmed that the number of errors in the 4th error category in the first learning data (1010) includes only the minimum number.

[0226] In the case of the second learning data (1020), as described in the previous FIG. 8, since the data corresponding to the first error category within a preset distance from the first learning data (1010) has been removed, it can be confirmed that the number of error types corresponding to the first error category in the second learning data (1020) is smaller than the number of error types corresponding to the first error category in the first learning data (1010).

[0227] Therefore, in order to more accurately distinguish between the first and second error categories, the second learning data (1020) has data corresponding to intermediate values ​​deleted, and thus, by using the second learning data, it is possible to train a neural network model that can more accurately distinguish whether the error is in the first error category or the second error category (or the third error category).

[0228] When the first learning data (1010) is prepared in this way, the first neural network model can be trained using the first learning data (1010), and the second neural network model can be trained using the second learning data (1020).

[0229] In one or more embodiments, the following describes specific operations for performing fault diagnosis using a first neural network model and a second neural network model learned in the above-described manner.

[0230] FIG. 11 is a diagram illustrating an example configuration of a neural network model according to one or more embodiments of the present disclosure.

[0231] As illustrated in FIG. 11, when a plurality of sensing values ​​(i.e., sensing information) are received from a sensor (110), the preprocessing module (1110) can process (or process) at least some of the received sensing values ​​to obtain a data set to be input into a neural network model.

[0232] In the above, it has been described that the information provided to the preprocessing module (1110) is input through sensors. However, in one or more embodiments, some of the information may be information stored in memory (120) or preset data values, rather than hardware sensors.

[0233] The preprocessing module (1110) can combine the sensed values ​​(or measured values) of a certain period of time (e.g., 100 ms when the data collection cycle is 20 ms), determine which of the sensed values ​​among the acquired sensed values ​​to include in the data set, and can also obtain the data set by assigning weights to the sensed values. In one or more embodiments, the preprocessing module (1110) can perform an operation of using an average value accumulated over a certain period of time for the sensed values, assigning a specific weight value to the corresponding values, or assigning a specific scaling value (e.g., scaling up a value of 0.001 to 1). In one or more embodiments, the data can be encoded or embedded to suit the purpose of the neural network model (1120, 1130).

[0234] The 'first neural network model (1120)' and 'second neural network model (1130)' illustrated in FIG. 11 refer to artificial intelligence models that include a neural network trained to acquire information indicating an error type based on current information. As illustrated in FIG. 11, when a data set is received from the preprocessing module (1110), each of the first neural network model (1120) and the second neural network model (1130) can output probability values ​​for each of a plurality of error types.

[0235] For example, the first neural network model (1120) and the second neural network model (1130) can be implemented as a CNN (Convolutional Neural Network), LSTM (Long Short-Term Memory), DNN (Deep Neural Network), RNN (Recurrent Neural Network), RBM (Restricted Boltzmann Machine), DBN (Deep Belief Network), BRDNN (Bidirectional Recurrent Deep Neural Network), etc., and the two neural network models can have the same structure, that is, only the learning data can be different. Alternatively, the two neural network models can be models with different structures.

[0236] In one or more embodiments, when 'type information' is used as a general term for information indicating the type of error, it may also be referred to as error type information, error probability information, probability information, etc. Specifically, the type information may be expressed as probability values ​​for multiple error types, but may also include information about whether an error exists or at least one of the types of errors.

[0237] In the above, it was explained that the probability value of each error type is output as the output of the neural network model, but when implemented, an additional configuration (e.g., softmax, etc.) may be added to the neural network model to output only the error type value with the highest probability value, or the judgment operation in the control module (1140) described below may be performed within the neural network model.

[0238] The control module (1140) can determine whether an error exists and, if an error exists, the corresponding error type by using the first type information (i.e., probability values ​​for each of multiple error types) output from the first neural network model (1120) and the second type information output from the second neural network model (1130).

[0239] In one or more embodiments, the control module (1140) can use two reference values ​​to determine the type of error. In one or more embodiments, a first reference value (e.g., 0.7) is a reference value that determines an error if the value is greater than or equal to the first reference value, and a second reference value (e.g., 0.5) is a reference value that determines a non-error if the value is less than the first reference value (e.g., a type indicating that no error has occurred). In one or more embodiments, if the above-described numerical values ​​are not exemplary, other values ​​may be used in addition to the above-described values, and instead of using two reference values, only one reference value may be used, and it is also possible to use three or more reference values.

[0240] Accordingly, the control module (1140) can first determine that there is no error (e.g., a type in which no error occurred) if all probability values ​​within the first type information are less than the second reference value.

[0241] In one or more embodiments, the control module (1140) may determine that there is an error of an error type greater than the first reference value if at least one probability value among the probability values ​​within the first type information is greater than the first reference value. In one or more embodiments, the control module (1140) may additionally perform accurate error classification based on the probability value within the second type information if there is a case where the probability value of an error type belonging to a different error category than the corresponding error type is greater than the second reference value.

[0242] For example, if the probability value of the first error type corresponding to the first error category in the first type information is 0.7 or greater, and the probability value of the third error type corresponding to the second error category is within the range of 0.7 to 0.5, the control module (1140) can identify the error type based on the probability value in the second type information for accurate classification.

[0243] In one or more embodiments, if there is an error type greater than a preset third reference value among the error values ​​in the second type information, the control module (1140) may determine that the error type greater than the third reference value is an error. Alternatively, the control module (1140) may check the largest error value in the second type information and determine the error type corresponding to the largest error value as the corresponding error.

[0244] For example, if the control module (1140) determines that the probability value of an error type belonging to a different error category than the error category to which the error type larger than the first reference value belongs is within the range of the first reference value to the second reference value, or if, overall, there is no error type larger than the first reference value and the probability value is within the range of the first to second reference values, the control module (1140) can use the second neural network model to determine the error type based on the probability value of the second neural network model.

[0245] For example, if the probability value of an error type belonging to the second neural network model is greater than or equal to a third reference value, the control module (1140) may determine that an error of the error type greater than or equal to the third reference value exists. In one or more embodiments, the third reference value may be the same as the first reference value described above (e.g., 0.7). However, other reference values ​​may be used during implementation.

[0246] If the probability value of the first neural network model for the first error category is greater than or equal to the first reference value, and the probability value of the first neural network model for the second error category is within the first to third reference values, the control module (1140) may preferentially determine that there is an error in the electronic device, and may also identify the error type having the highest probability value among the probability values ​​output from the second neural network model as the currently occurring error type.

[0247] Therefore, in summary of the above, first, a method may be used to identify the error type by comprehensively utilizing each probability value of the first neural network model and each probability value of the second neural network model. Second, the first neural network model may be used to only check for the presence of an error in the electronic device, and if an error is determined to exist, the second neural network model may be used to identify the error. Third, a method may be used to identify the error type using the first neural network model, but additionally use the second neural network model only when the error type is difficult to distinguish (e.g., the first error category and the second error category), thereby accurately distinguishing between the two error categories.

[0248] When an error type (or error category) is identified, the control module (1140) can perform a control action corresponding to the corresponding error category. The control action for each error category has been previously described in FIG. 2, and thus, a redundant description will be omitted.

[0249] The above describes a method of simultaneously (or jointly) inputting current information to the first and second neural network models. However, during implementation, it is also possible to sequentially use the first and second neural network models. This will be described below with reference to Figure 12.

[0250] FIG. 12 is a diagram illustrating an example configuration of a neural network model according to one or more embodiments of the present disclosure.

[0251] As illustrated in FIG. 12, when a plurality of sensing values ​​(i.e., sensing information) are received from the sensor (110), the preprocessing module (1210) can process (or process) at least some of the received sensing values ​​to obtain a data set to be input into the neural network model.

[0252] The "first neural network model (1220)" refers to an artificial intelligence model that includes a neural network trained to acquire information indicating an error type based on current information. The first neural network model (1220) can receive current information as input and output first type information.

[0253] The control module (1240) can primarily determine whether an error exists and, if so, the corresponding error type based on the first error type information output from the first neural network model (1220). For example, it can determine whether at least one probability value among a plurality of probability values ​​is greater than or equal to a second reference value.

[0254] If at least one probability value is greater than or equal to the first reference value, the control module (1240) can determine that the error category corresponding to the error type having a probability value greater than or equal to the first reference value is abnormal.

[0255] If the total probability value is not greater than the first reference value, the control module (1240) may additionally cause the second neural network model (1230) to output second type information so that fault diagnosis using the second neural network model is performed.

[0256] Accordingly, when additional second type information is acquired, the control module (1240) can check each probability value within the acquired second type information to determine whether an error or malfunction exists. For example, if at least one probability value among multiple probability values ​​is greater than or equal to a third reference value, the error can be determined as an abnormality in an error category corresponding to an error type having a probability value greater than or equal to the third reference value.

[0257] In one or more embodiments, the third reference value may be the same as or different from the first reference value described above. In one or more embodiments, the use of the second neural network model may also involve not utilizing a reference value. That is, if an error is identified using the first neural network model, a failure of the type of error having the highest probability value among the probability values ​​output by the second neural network model may be determined to have occurred.

[0258] FIG. 13 is a diagram illustrating an example configuration of a neural network model according to one or more embodiments of the present disclosure.

[0259] Specifically, FIG. 13 is a diagram illustrating an example of using a neural network model provided on an external server in one or more embodiments.

[0260] Referring to FIG. 13, when receiving current information from a sensor, the preprocessing module (1310) can process (or process) at least a portion of the data included in the current information to obtain a data set to be input into a neural network model.

[0261] The data set can be transmitted to an external device (200) via a communication device (1320).

[0262] The external device (200) can receive a data set via the communication device (210). The processor (220) of the external device (200) can control the communication device (210) to generate type information using the first and second neural network models that have been learned, and transmit the generated type information to the electronic device (100). In one or more embodiments, the processor (220) may not transmit the type information as is, but may identify whether an error has occurred or the category of an error that has occurred based on the type information, and transmit the identified result to the electronic device (100) via the communication device (210).

[0263] When type information is received through the communication device (1320), the control module (1330) can check whether an error exists and / or the error type (or error category) based on the received type information.

[0264] In one or more embodiments, the control module (1330) may control the electronic device (100) to operate in a control mode corresponding to the identified error type (or error category).

[0265] In Fig. 13, the neural network model described in Fig. 11 is described assuming that it is deployed on an external device, but when implemented, the neural network model of the embodiment described in Fig. 12 can also be implemented in a form in which it is deployed on an external server.

[0266] The above describes the use of two neural network models. However, depending on the type of error category, implementation can also utilize three or more neural networks. Furthermore, the above-described operations can be performed with a single neural network model. For example, a single neural network model can be individually trained with the first and second learning models described above, so that the operations performed by the two neural network models in Figure 11 described above can be performed with a single neural network model.

[0267] FIG. 14 is a drawing for explaining a control method in an electronic device according to one or more embodiments of the present disclosure.

[0268] Referring to FIG. 14, the electronic device (or processor) detects internal current information.

[0269] In one or more embodiments, the electronic device inputs current information into each of a first neural network model and a second neural network model trained to classify an error type of the electronic device to obtain first error type information and second error type information.

[0270] In one or more embodiments, the first neural network model and the second neural network model are models that output probability values ​​corresponding to each of a plurality of error types. In one or more embodiments, the first neural network model may be a model pre-trained using first learning data for the plurality of error types, and the second neural network model may be a model pre-trained using second learning data obtained by processing the learning data to distinguish between error categories to which each of the plurality of error types belongs. The learning process and characteristics of each neural network model have been described above, and thus, a redundant description thereof will be omitted.

[0271] In one or more embodiments, the electronic device identifies an error category of the electronic device based on the acquired first error type information and the acquired second error type information. For example, the acquired first error type information may be used first to identify the error category, and then the acquired second error type information may be additionally identified to identify the error category. Alternatively, the error category may be identified after both the first error type information and the second error type information have been acquired.

[0272] In the first case, if the probability values ​​of each of the plurality of error types included in the first error type information are smaller than the first reference value corresponding to the criterion for error judgment and at least one probability value included in the first error type information is larger than the second reference value that is smaller than the first reference value, the error category of the electronic device can be identified using the probability values ​​of each of the plurality of error types included in the second error type information.

[0273] In the second case, if at least one probability value among the plurality of probability values ​​in the first error type information is greater than or equal to the first reference value that serves as the criterion for error judgment, the error category corresponding to the error type greater than or equal to the first reference value can be identified as the error category of the electronic device.

[0274] In one or more embodiments, if a plurality of probability values ​​within the first error type information are less than a second reference value that is less than the first reference value, the electronic device can determine that there is no error in the electronic device.

[0275] In one or more embodiments, if at least one probability value among the plurality of probability values ​​in the first error type information has a value between the first reference value and the second reference value, and if at least one probability value among the plurality of probability values ​​in the second error type information is greater than or equal to a third reference value that serves as a reference for error determination, the electronic device can identify an error category corresponding to an error type greater than or equal to the third reference value as an error category of the electronic device.

[0276] In one or more embodiments, the electronic device may control the operation of the electronic device based on the identified error category (S1440). For example, if the identified error category is a first error category, the operation of the electronic device may be controlled to operate in a first mode that limits the performance of the driving device corresponding to the current information. If the identified error category is a second error category, the operation of the electronic device may be controlled to operate in a second mode that further limits the performance of the driving device than the first mode. If the identified error category is a third error category, the operation of the electronic device may be controlled to prevent the driving device from operating.

[0277] When an error category is identified in this manner, the electronic device may display information about the identified category. In one or more embodiments, the electronic device may perform a control action selected by the user when the first error category or the second error category is identified.

[0278] In one or more embodiments, when a control command for a configuration related to an error is input during operation in the first to third modes as described above, the electronic device may display information indicating that the control command cannot be performed if the identified error category is the first error category and the control command corresponds to the performance of the driving device.

[0279] As described above, the control method according to the present embodiment utilizes two neural network models, particularly the second neural network model, which more clearly distinguishes between error categories, thereby enabling more accurate classification of errors in electronic devices. In one or more embodiments, the method enables accurate error classification, and thus also enables the execution of control actions corresponding to the classified errors.

[0280] FIG. 15 is a flowchart illustrating an error diagnosis method using multiple neural network models according to one or more embodiments of the present disclosure. Specifically, FIG. 15 is a flowchart illustrating an example of an error diagnosis method when multiple neural network models are sequentially used.

[0281] Referring to FIG. 15, multiple current information can be input into a first neural network model to obtain first error type information. By checking each probability value within the obtained error type information, the presence of an error or failure can be confirmed. For example, it can be confirmed whether at least one probability value among the multiple probability values ​​is greater than or equal to a second reference value (S1510-Y).

[0282] If at least one probability value is greater than or equal to the first reference value, it can be determined as an abnormality of an error category corresponding to an error type having a probability value greater than or equal to the first reference value (S1530).

[0283] If the overall probability value is not greater than the first reference value, additional fault diagnosis using a second neural network model can be performed (S1530-N). For example, by inputting multiple current information into the second neural network model, second fault type information can be obtained, and the presence or absence of an error or fault can be confirmed using each probability value within the obtained fault type information.

[0284] In one or more embodiments, if at least one probability value among a plurality of probability values ​​is greater than or equal to a third reference value, it may be determined as an abnormality of an error category corresponding to an error type having a probability value greater than or equal to the third reference value (S1520-Y).

[0285] In one or more embodiments, the third reference value may be the same as or different from the first reference value described above. In one or more embodiments, the use of the second neural network model may also involve not utilizing a reference value. That is, if an error is identified using the first neural network model, a failure of the type of error having the highest probability value among the probability values ​​output by the second neural network model may be determined to have occurred.

[0286] Meanwhile, in the illustrated example, fault diagnosis is performed using the second neural network model even when a fault is not diagnosed based on the first error type information of the first neural network model. However, when implemented, if all probability values ​​within the first error type information are lower than or equal to the second reference value, it may be determined that there is no fault and the second neural network model may not be used.

[0287] In one or more embodiments, if at least one probability value in the first error type information is greater than or equal to the first reference value, the error type having the first reference value is determined. However, if the error type belongs to the first error category or the second error category that requires precise distinction, the error type may be additionally confirmed again using the second neural network model.

[0288] In one or more embodiments, although the illustrated example describes using two neural network models sequentially, for faster execution, it is also possible to input current information into two neural network models, obtain both the first error type information and the second error type information, and then perform only the above-described judgment operation based on FIG. 15.

[0289] FIG. 16 is a diagram for explaining different control operations for each error category according to one or more embodiments of the present disclosure.

[0290] Referring to Fig. 16, it is determined whether a failure has occurred (S1610). This step can be performed in the same manner as the operation of Fig. 16 described above.

[0291] When a fault is confirmed (S1610-Y), it is checked whether the fault category that occurred is the first fault category (S1620). If the fault category that occurred is the first fault category (S1620-Y), the electronic device can operate in the first mode.

[0292] In one or more embodiments, it is determined whether the error category that occurred is a second error category (S1630). If the error category that occurred is a second error category (S1630-Y), the user selects an operating mode (S1640). Depending on the user's selection, the second mode may be used, or the operation of the configuration in which the error occurred may be halted.

[0293] In the illustrated example, it is described that the second mode or the driving operation is performed according to the user's selection, but when implemented, it is also possible to operate in the second mode first, and then switch to maintaining the second mode or stopping the driving operation according to the user's settings.

[0294] If the error category is other than the first to third error categories (i.e., if the error category that occurred is the fourth error category), the performance of the function corresponding to the error can be adjusted.

[0295] FIG. 17 is a diagram illustrating user control commands and error category-specific control operations according to one or more embodiments of the present disclosure.

[0296] Referring to Fig. 17, a user control command is input. Such a user control command may be a menu selection for controlling the function of an electronic device or a control command for changing the settings of a specific function.

[0297] When a user's control command is input, it is checked whether the current operating state (or operating mode) of the electronic device is stopped (S1710). In other words, it is possible to check whether an error has occurred due to the second error category or the third error category as described above.

[0298] If in a stopped state (S1710-Y), the electronic device may indicate to the user that operation for a specific function is currently stopped. In one or more embodiments, if the electronic device falls under the third error category as described above, the electronic device may notify the user that operation is not possible, and if the electronic device falls under the second error category, the electronic device may not perform operation or display a menu prompting the user to select whether to operate in the second mode (S1720).

[0299] If not in a stationary state (S1710-N), the electronic device can determine whether it is operating through safe operation 1 (S1730). That is, the electronic device can determine whether it is operating in the first mode as described above, whether the error that occurred corresponds to the first error category, and whether the user's control command is an operation that puts a load on the motor (S1740).

[0300] For example, an electronic device may determine that an action to increase the temperature of a refrigerator is not a load-inducing action, but an action to decrease the temperature of the refrigerator or a control command to increase the rotation speed of a motor from the current level are load-inducing actions.

[0301] If the user's control command does not affect the load, the electronic device can perform the action according to the user's request.

[0302] If the user's control command affects the load, the electronic device may notify the user that it cannot perform the request (S1740-Y).

[0303] FIG. 18 is a drawing for explaining an example of operation when an electronic device according to one or more embodiments of the present disclosure is a refrigerator.

[0304] The following description assumes that a user control command is input while operating in the second mode, but the following operation can be similarly applied while operating in the first mode.

[0305] First, when operating in the second mode (S1805), if the user selects a menu, turns on the touch screen to input a control command, or operates an input device, the electronic device (100) can display only currently available items. For example, if a refrigerator operates with a single compressor and can only partially maintain temperature in the refrigerator compartment, partially maintain temperature in the freezer compartment, or only maintain temperature in the variable temperature compartment, only menu items for the operable configuration can be displayed.

[0306] If only the freezer (or refrigerator) is available (S1815, S1820), the electronic device displays operable items related to the freezer operation, and if the temperature set by the user is a temperature that currently loads the motor (S1830). In one or more embodiments, the electronic device may indicate that the operation cannot be performed (S1835). If necessary, the electronic device may display a notification that the operation of the corresponding configuration is to be stopped (S1840), provide the notification to the user terminal, or provide a notification requesting that the plug be controlled as there is a dangerous situation (S1845).

[0307] If only a variable temperature room can be used (S1825), the electronic device can display only the items that the user can operate in relation to the operation of the variable temperature room and determine whether the user's setting operation can be performed (S1850). That is, the electronic device can determine whether the user's operation command is a setting that adds additional load to the current motor (S1850, S1860) and display information about functions (or types of internal components) that cannot be used (S1855, 1835). For example, if the variable temperature room can be used by changing it to a refrigerator or a freezer, the electronic device can determine whether it can be operated as a freezer or a refrigerator in the current state, and if it can be used only as a refrigerator, it can notify that it can be used only as a refrigerator.

[0308] If the refrigerator cannot be used even with the temperature set by the user, the electronic device may display a message indicating that the items in the refrigerator cannot be stored.

[0309] When performing such operations, the refrigerator may operate at a minimum motor speed to maintain only the minimum temperature, and may guide the user to perform a fault-remediation operation to enable quick A / S. In one or more embodiments, additional operations may be performed to prevent unnecessary power consumption in such a state, such as turning off the refrigerator lights or disabling the defrost function.

[0310] FIG. 19 is a drawing for explaining an example of operation when an electronic device according to one or more embodiments of the present disclosure is an air conditioner.

[0311] First, it is assumed that an error corresponding to the first error category has occurred in the air conditioner and it is operating in the first mode (S1910).

[0312] When the electronic device is in operation in the first mode, if the user selects a menu, turns on the touch screen to input a control command, or operates an input device, the electronic device (100) can display only the currently available items (S1920). For example, an air conditioner may have a strong cooling mode, a cooling mode, a delivery mode, an air purification mode, etc., and may display that the remaining modes, except for the strong cooling mode that puts a burden on the motor, are selectable.

[0313] In one or more embodiments, if the user selects the cooling mode (S1930-Y), the electronic device can check the user's set temperature or blower strength to determine whether the set temperature or blower strength input by the user causes additional load on the motor (S1940). In one or more embodiments, the electronic device can perform the user control command if it is within the range that does not cause additional load, and otherwise, can indicate that the corresponding function cannot be performed (S1950).

[0314] If the user selects a fan mode or air purification mode other than the cooling mode, the electronic device can execute the user command as is.

[0315] The above assumes a case where an error corresponding to the first error category occurs, but an error corresponding to the second error category can also operate similarly to the above-described case. That is, the items that the user can select are limited to the blower mode and the air purification mode excluding the cooling operation, and control can be performed only within the corresponding operation.

[0316] FIG. 20 is a drawing for explaining an example of operation when an electronic device according to one or more embodiments of the present disclosure is a washing machine.

[0317] First, the washing machine's error can be checked. If the error falls under the third error category, the washing machine can be displayed as unable to perform any operation. If the error falls under the second error category, only items other than high-speed washing and spin-drying, which place a significant load on the motor, can be selected. If the error falls under the first error category, only items that assume the maximum spin-drying speed or a certain weight of laundry cannot be performed can be displayed.

[0318] When a user washing command is input on a restricted displayed item (S2010), an operation according to the input washing command can be performed (S2030). In one or more embodiments, in the first error category, the rotation speed of the motor can be controlled to operate within a range that does not use the maximum speed as described above, and in the second error category, the washing operation, etc. can be controlled to be performed only at a speed lower than the maximum speed described above.

[0319] Meanwhile, the methods according to at least some of the various embodiments of the present disclosure described above may be implemented in the form of an application that can be installed on an existing electronic device.

[0320] In one or more embodiments, the methods according to at least some of the various embodiments of the present disclosure described above may be implemented solely through a software upgrade or a hardware upgrade to an existing electronic device.

[0321] In one or more embodiments, the methods according to at least some of the various embodiments of the present disclosure described above may also be performed via an embedded server provided in the electronic device, or via an external server of at least one of the electronic devices.

[0322] According to one or more embodiments of the present disclosure, the various embodiments described above can be implemented as software including instructions stored in a machine-readable storage medium that can be read by a machine (e.g., a computer). The device can include an electronic device (e.g., an electronic device (A)) according to the disclosed embodiments, which is a device that can call instructions stored from the storage medium and operate according to the called instructions. When an instruction is executed by a processor, the processor can perform a function corresponding to the instruction directly or by using other components under the control of the processor. The instruction can include code generated or executed by a compiler or an interpreter. The machine-readable storage medium can be provided in the form of a non-transitory storage medium. In one or more embodiments, a 'non-transitory storage medium' means a tangible device and does not include a signal (e.g., an electromagnetic wave), and this term can be used to refer to cases where data is stored semi-permanently in the storage medium and cases where data is stored temporarily. No distinction is made. For example, a 'non-transitory storage medium' may include a buffer in which data is temporarily stored. In one or more embodiments, the methods according to the various embodiments disclosed in the present document may be provided as a computer program product. The computer program product may be traded between sellers and buyers as a commodity. The computer program product may be distributed in the form of a machine-readable storage medium (e.g., compact disc read only memory (CD-ROM)), or may be distributed online (e.g., downloaded or uploaded) through an application store (e.g., Play Store™) or directly between two user devices (e.g., smartphones).In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) may be temporarily stored or temporarily created in a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or an intermediary server.

[0323] Various embodiments of the present disclosure may be implemented as software including instructions stored in a machine-readable storage medium that can be read by a machine (e.g., a computer). The device may include an electronic device (e.g., an electronic device (100)) according to the disclosed embodiments, which is a device that can call instructions stored in the storage medium and operate according to the called instructions.

[0324] When the above-described instruction is executed by the processor, the processor may perform the function corresponding to the instruction directly or by utilizing other components under the control of the processor. The instruction may include code generated or executed by a compiler or interpreter.

[0325] Although the preferred embodiments of the present disclosure have been illustrated and described above, the present disclosure is not limited to the specific embodiments described above, and various modifications may be made by a person having ordinary skill in the art to which the present disclosure pertains without departing from the gist of the present disclosure as claimed in the claims, and such modifications should not be understood individually from the technical idea or prospect of the present disclosure.

Claims

1. In electronic devices, A sensor for detecting information indicating the status of the electronic device; A memory storing one or more instructions and a first neural network model and a second neural network model trained to classify error types of the electronic device; and A processor connected to the memory and operable to execute one or more instructions stored in the memory; The one or more instructions, when executed by the processor, cause the electronic device to: By inputting the information representing the state of the electronic device into each of the first neural network model and the second neural network model, first error type information and second error type information are obtained, Checking the error category of the electronic device based on the first error type information and the second error type information, Controlling the operation of the electronic device based on the above error category, The above first neural network model is, A pre-trained model using the first learning data for multiple error types, The above second neural network model is, An electronic device which is a model trained using second learning data processed from the first learning data to distinguish between error categories to which each of the above plurality of error types belongs.

2. In paragraph 1, The above first neural network model and the above second neural network model are models that output probability values ​​corresponding to each of a plurality of error types. The one or more instructions, when executed by the processor, cause the electronic device to: An electronic device that determines an error category of the electronic device by using the probability values ​​of each of the plurality of error types included in the first error type information, based on the probability values ​​of each of the plurality of error types included in the first error type information being smaller than a first reference value corresponding to a criterion for error judgment and at least one probability value included in the first error type information being larger than a second reference value that is smaller than the first reference value.

3. In paragraph 1, The above first neural network model and the above second neural network model are models that output probability values ​​corresponding to each of a plurality of error types. The one or more instructions, when executed by the processor, cause the electronic device to: If at least one probability value among the plurality of probability values ​​in the first error type information is greater than or equal to the first reference value that serves as the criterion for error judgment, the error category corresponding to the error type greater than or equal to the first reference value is confirmed as the error category of the electronic device, If the plurality of probability values ​​within the above first error type information are smaller than the second reference value which is smaller than the first reference value, it is determined that there is no error in the electronic device, An electronic device that determines an error category corresponding to an error type equal to or greater than the third reference value as an error category of the electronic device, if at least one probability value among a plurality of probability values ​​in the first error type information has a value between the first reference value and the second reference value, and if at least one probability value among a plurality of probability values ​​in the second error type information is equal to or greater than a third reference value that serves as a reference for error judgment.

4. In paragraph 1, The one or more instructions, when executed by the processor, cause the electronic device to: Based on the first error type information and the second error type information obtained above, one of the error categories of no error, the first error category, the second error category and the third error category is confirmed, If the above-determined error category is the first error category, the operation of the electronic device is controlled to operate in a first mode that limits the performance of the driving device corresponding to the information indicating the state of the electronic device, If the above-determined error category is a second error category, the operation of the electronic device is controlled to operate in a second mode that further limits the performance of the driving device than the first mode; An electronic device that controls the operation of the electronic device so that the driving device does not operate if the above-determined error category is a third error category.

5. In paragraph 4, display; including more; The one or more instructions, when executed by the processor, cause the electronic device to: An electronic device controlling said display so as to display information about the identified error category.

6. In paragraph 5, further comprising an input device for receiving a control command; The one or more instructions, when executed by the processor, cause the electronic device to: An electronic device that controls the display so that information indicating that the control command cannot be performed is displayed when the above-determined error category is a first error category and the control command is a command corresponding to the performance of the driving device.

7. In paragraph 1, Further comprising a driving device including a motor and an inverter providing driving power to the motor; The above sensor senses multiple current values ​​within the driving device, The one or more instructions, when executed by the processor, cause the electronic device to: An electronic device that determines an error category for at least one of the motor and the inverter based on the first error type information and the second error type information.

8. In paragraph 1, The above first learning data is sampled data so that each error type has a similar number of data among the collected data, An electronic device in which the second learning data is learning data in which data within a preset similar range of the second error type belonging to the second error category are removed from the first learning data of the first error type belonging to the first error category, and data within a preset similar range of the first error type belonging to the second error category are removed from the second learning data of the second error type.

9. In paragraph 8, The second learning data is An electronic device including augmented data corresponding to one of the first error type and the second error type, such that the training data corresponding to the first error category and the training data corresponding to the second error category have a similarity within a predetermined threshold value.

10. In paragraph 8, The data corresponding to the first error category and the second error category in the first learning data are data based on the information indicating the status of the electronic device, An electronic device wherein the first learning data further includes experimental data for multiple error types within a third error category.

11. In a method for controlling an electronic device, A step of detecting information indicating the status of the electronic device; A step of obtaining first error type information and second error type information by inputting the information representing the state of the electronic device into each of a first neural network model and a second neural network model trained to classify the error type of the electronic device; A step of confirming an error category of the electronic device based on the first error type information and the second error type information; and A step of controlling the operation of the electronic device based on the above error category; The above first neural network model is, A pre-trained model using the first learning data for multiple error types, The above second neural network model is, A control method in which a model is learned by using second learning data processed from the first learning data to distinguish between error categories to which each of the above plurality of error types belongs.

12. In paragraph 11, The above first neural network model and the above second neural network model are models that output probability values ​​corresponding to each of a plurality of error types. The steps to check the above error categories are: A control method for confirming an error category of the electronic device by using the probability values ​​of each of the plurality of error types included in the first error type information, based on the probability values ​​of each of the plurality of error types included in the first error type information being smaller than a first reference value corresponding to a criterion for error judgment and at least one probability value included in the first error type information being larger than a second reference value that is smaller than the first reference value.

13. In paragraph 11, The above first neural network model and the above second neural network model are models that output probability values ​​corresponding to each of a plurality of error types. The steps to check the above error categories are: If at least one probability value among the plurality of probability values ​​in the above first error type information is greater than or equal to the first reference value that serves as the criterion for error judgment, the error category corresponding to the error type greater than or equal to the first reference value is confirmed as the error category of the electronic device, If the plurality of probability values ​​within the above first error type information are smaller than the second reference value which is smaller than the first reference value, it is determined that there is no error in the electronic device, A control method for confirming an error category corresponding to an error type equal to or greater than the third reference value as an error category of the electronic device, when at least one probability value among a plurality of probability values ​​in the first error type information has a value between the first reference value and the second reference value, and at least one probability value among a plurality of probability values ​​in the second error type information is equal to or greater than a third reference value that serves as a reference for error judgment.

14. In paragraph 11, The steps to check the above error categories are: Based on the first error type information and the second error type information obtained above, one of the error categories of no error, the first error category, the second error category and the third error category is confirmed, The above controlling step is, If the above-determined error category is the first error category, the operation of the electronic device is controlled to operate in a first mode that limits the performance of the driving device corresponding to the information indicating the state of the electronic device, If the above-determined error category is a second error category, the operation of the electronic device is controlled to operate in a second mode that further limits the performance of the driving device than the first mode; A control method for controlling the operation of the electronic device so that the driving device does not operate if the above-determined error category is a third error category.

15. In a non-transitory computer-readable recording medium storing a program for executing a control method in an electronic device, The above control method is, A step of detecting information indicating the status of the electronic device; A step of obtaining first error type information and second error type information by inputting the information representing the state of the electronic device into each of a first neural network model and a second neural network model trained to classify the error type of the electronic device; A step of confirming an error category of the electronic device based on the first error type information and the second error type information; and A step of controlling the operation of the electronic device based on the above error category; The above first neural network model is, A pre-trained model using the first learning data for multiple error types, The above second neural network model is, A computer-readable recording medium which is a model learned by using second learning data processed from the first learning data to distinguish between error categories to which each of the above plurality of error types belongs.

Images