Cleaner operating according to floor characteristics and operating methods thereof

Abstract

This cleaner comprises: a suction module; a memory storing a plurality of artificial intelligence models; and a processor, wherein the processor identifies at least one operation section corresponding to state data among a plurality of operation sections, identifies floor characteristics by using an artificial intelligence model corresponding to the identified operation section among the plurality of artificial intelligence models, and adjusts a driving state of the suction module on the basis of the identified floor characteristics.

Description

Vacuum cleaners that operate according to floor characteristics and their operating methods

[0001] The present invention relates to a vacuum cleaner and a method thereof that identify the characteristics of a floor on which cleaning work is performed and operate accordingly.

[0002] Advances in electronic technology have enabled a variety of vacuum cleaners to become part of our everyday lives. Users can use these vacuum cleaners in a variety of floor environments. For example, they can clean on a variety of surfaces, including carpets, hard floors, and mats. They can also clean flat surfaces, curved surfaces, and corners.

[0003] Therefore, the need for a vacuum cleaner that can accurately identify floor characteristics and operate appropriately accordingly has arisen.

[0004] According to at least one embodiment of the present disclosure, a vacuum cleaner includes a suction module, a memory storing a plurality of artificial intelligence models corresponding to a plurality of operation sections that divide a vacuum cleaner operation used for cleaning into a plurality of sections, and a processor. The processor operates the suction module to suck up foreign substances, collects status data related to the operation of the suction module and stores the collected data in the memory, identifies at least one operation section corresponding to the status data among the plurality of operation sections, identifies a floor characteristic on which the cleaning operation is performed using an artificial intelligence model corresponding to the identified operation section among the plurality of artificial intelligence models, and adjusts the operation state of the suction module based on the identified floor characteristic.

[0005] In addition, according to at least one embodiment of the present disclosure, a method for operating a vacuum cleaner including a suction module includes a step of driving the suction module to suck up foreign substances, a step of collecting and storing status data related to the driving of the suction module, a step of identifying at least one operation section corresponding to the status data among a plurality of operation sections in which an operation of the vacuum cleaner used for cleaning is divided into a plurality of sections, and a step of identifying a floor characteristic of a floor surface on which the cleaning is performed using an artificial intelligence model corresponding to the identified operation section among a plurality of artificial intelligence models that are learned and stored to correspond to each of the plurality of operation sections, and a step of adjusting the operation state of the suction module based on the identified floor characteristic.

[0006] In addition, a non-transitory readable recording medium according to at least one embodiment of the present disclosure stores a program for performing the identification method, including the steps of collecting status data related to the cleaning operation when the cleaner performs the cleaning operation, identifying at least one operation section corresponding to the status data among a plurality of operation sections in which the operation of the cleaner performing the cleaning operation is divided into a plurality of sections, and identifying a floor characteristic on which the cleaning operation is performed using an artificial intelligence model corresponding to the identified operation section among a plurality of artificial intelligence models that are learned and stored to correspond to each of the plurality of operation sections.

[0007] FIG. 1 is a drawing for explaining an example of the external configuration of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0008] FIG. 2 is a block diagram illustrating a configuration of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0009] FIG. 3 is a diagram illustrating an example of current data, which is one of the status data of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0010] FIG. 4 is a diagram illustrating an example of a software structure of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0011] FIG. 5 and FIG. 6 are diagrams for explaining an example of a learning method of an interval identification model using the interval identification model of FIG. 4.

[0012] FIG. 7 is a drawing for explaining the operation of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0013] FIG. 8 is a diagram illustrating an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using an artificial intelligence model.

[0014] FIG. 9 is a diagram illustrating an example of an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using a plurality of artificial intelligence models.

[0015] FIG. 10 is a diagram illustrating another example of an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using multiple artificial intelligence models.

[0016] FIG. 11 is a flowchart illustrating a method of operating a vacuum cleaner according to at least one embodiment of the present disclosure.

[0017] The terms used in the various embodiments of this disclosure have been selected from widely used, current terms, taking into account the functions of this disclosure. However, these terms may vary depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the description of the relevant disclosure. Therefore, the terms used in this disclosure should be defined based on the meaning of the terms and the overall content of this disclosure, rather than simply their names.

[0018] It should be understood that the various embodiments of the present disclosure and the terminology used therein are not intended to limit the technical features described in the present disclosure to specific embodiments, but include various modifications, equivalents, or substitutes of the embodiments.

[0019] In connection with the description of the drawings, similar reference numerals may be used for similar or related components.

[0020] The singular form of a noun corresponding to an item may include one or more of said items, unless the relevant context clearly indicates otherwise.

[0021] In this disclosure, each of the phrases "A or B", "at least one of A and B", "at least one of A or B", "A, B, or C", "at least one of A, B, and C", and "at least one of A, B, or C" may include any one of the items listed together in the corresponding phrase, or all possible combinations thereof.

[0022] Terms such as "first," "second," or "first" or "second" may be used simply to distinguish one component from another and do not qualify the components in any other respect (e.g., importance or order).

[0023] When a component (e.g., a first component) is referred to as being “coupled” or “connected” to another component (e.g., a second component), with or without the terms “functionally” or “communicatively,” it means that the component can be connected to the other component directly (e.g., wired), wirelessly, or through a third component.

[0024] 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 the present disclosure, but do not preclude the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0025] When a component is said to be “connected,” “coupled,” “supported,” or “in contact with” another component, this includes not only cases where the components are directly connected, coupled, supported, or in contact, but also cases where the components are indirectly connected, coupled, supported, or in contact through a third component.

[0026] 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.

[0027]

[0028] *The term "and / or" includes any combination of multiple related described components or any one of multiple related described components.

[0029] In the present disclosure, a "module" or "part" performs at least one function or operation and may be implemented in hardware or software, or a combination of hardware and software. Furthermore, multiple "modules" or multiple "parts" may be integrated into at least one module and implemented as at least one processor, excluding any "modules" or "parts" that need to be implemented as specific hardware.

[0030] 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.

[0031] Meanwhile, 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 spacing depicted in the attached drawings.

[0032] Meanwhile, the cleaner according to various embodiments of the present disclosure may include at least one of, for example, a vacuum cleaner, a robot cleaner, a handy cleaner, a stick cleaner, etc.

[0033] An embodiment of the present disclosure will be described in more detail with reference to the attached drawings below.

[0034] FIG. 1 is a drawing for explaining an example of the external configuration of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0035] Referring to FIG. 1, a vacuum cleaner (100) according to one embodiment of the present disclosure has the form of a stick-type vacuum cleaner (or upright type). In the illustrated example, an upright type shape in which a suction module is formed integrally with the main body is depicted. However, when implemented, a canister type in which the suction module is provided separately from the main body and connected by an extension tube may be utilized. Alternatively, the vacuum cleaner may be implemented as various types of vacuum cleaners, such as a wirelessly controlled vacuum cleaner, a robot vacuum cleaner, or a handheld vacuum cleaner.

[0036] A vacuum cleaner (100) may include a vacuum cleaner body (10) and a suction module (110). The vacuum cleaner body (10) may accommodate major components such as a motor and a processor. In the present disclosure, the concept of the body assumes separation of components such as the suction module (110) or the stick (20). However, if the above-described components have an integrated form, the components may also be referred to as the body.

[0037] The vacuum cleaner (100) may include a stick (20) connecting the vacuum cleaner body (10) and the suction module (110) and a handle (40) connected to the vacuum cleaner body (10). The stick (20) connects the vacuum cleaner body (10) and the suction module (110), and provides suction force generated by the driving of the motor inside the vacuum cleaner body (10) to the suction module (110), thereby acting as a passage for transferring foreign substances sucked in by the suction module (110) to the inside of the vacuum cleaner body (10). If necessary, the stick (20) and the vacuum cleaner body (10) may be detachable. In addition, the vacuum cleaner body (10) may have a cleaning tool other than the aforementioned stick (20) or suction module (110) attached to it. The handle (40) is a part that is connected to the vacuum cleaner body (10) and may be provided so that a user can hold it and perform the operation of the vacuum cleaner.

[0038] The vacuum cleaner body (10) may include a dust collector (11) and a driving device (12) placed inside it. The dust collector (11) may perform the function of separating and collecting foreign substances from the air sucked in from the suction module (110).

[0039] The driving device (12) may include a motor assembly (50) that generates suction pressure of the cleaner (100). The motor assembly (50) may generate power to generate suction force inside the cleaner body (10).

[0040] The motor assembly (50) includes a first motor, and through the rotation of the first motor

[0041] Here, suction pressure can be generated. Specifically, when a drive command for the first motor is input and power is supplied to the first motor, the impeller rotates by the driving of the first motor. Suction pressure is generated by the rotation of the impeller, and air containing foreign substances can be sucked in through the suction port of the suction module (110) by this suction pressure. As the rotation speed of the first motor increases, the suction pressure increases.

[0042] When the suction strength is set (or determined), the cleaner (100) can control the first motor to rotate at a rotation speed corresponding to the set suction strength described above.

[0043] The suction module (110) may be provided at the lower portion of the cleaner body (10) and positioned so as to be in contact with the floor surface. The suction module (110) may be provided so as to be in contact with the floor surface and to draw dust or foreign substances from the floor surface into the interior of the cleaner body (10) using suction force generated from the motor assembly (50).

[0044] The suction module (110) may include a brush and a second motor. Specifically, when power is supplied to the second motor, the brush may be rotated by the driving of the second motor. The second motor may be a variety of motors, such as a DC (Direct Current) motor, an AC (Alternating Current) motor, or a BLDC (Brushless DC) motor.

[0045] Depending on the implementation example, a driving device for providing driving power to the second motor described above may be separately provided on the suction module (110) side. Alternatively, a driving device for controlling the second motor may be provided in the cleaner body (10).

[0046] The suction module (110) may be described by various terms such as suction head, suction device, suction part, suction system, etc., but is described as a suction module (110) in the present disclosure. In addition, in FIG. 1, the module itself that is detachable from the main body (10) or the stick (20) is referred to as the suction module (110), but the suction module (110) may be implemented in various forms depending on the external structure, size, type, etc. of the cleaner (100). For example, when implemented as a robot cleaner, the suction module (110) may be used as a term that comprehensively refers to the entire configuration that operates to perform cleaning by suction, such as a suction inlet formed on the bottom surface of the robot cleaner, an air duct connected to the suction inlet, a dust collector, and a motor assembly.

[0047] The brush is formed to extend a certain length out of the suction port, so that when the brush rotates, it can strike foreign substances such as dust, dirt, and hair stuck to the floor surface. This allows the foreign substances to be separated from the floor surface and easily sucked up by the suction port.

[0048] Meanwhile, depending on the floor characteristics, the suction strength and brush rotation speed described above can be adjusted to corresponding values ​​and operated accordingly. Floor characteristics refer to the characteristics of the floor on which the vacuum cleaner performs cleaning.

[0049] For example, when operating at maximum suction power on all types of floor surfaces, the floor surface may stick to the suction module (110) on certain floor surfaces, making cleaning difficult. In addition, rapid brush rotation on floor surfaces such as carpets may cause damage to the carpet. In this regard, the vacuum cleaner needs to identify the characteristics of the floor and operate at an appropriate suction strength and brush rotation speed.

[0050] Commonly used floor characteristics include lifted floor, hard floor, hard floor corner, long-pile carpet, short-pile carpet, and mat. "Lifted" refers to a condition away from the floor surface. Carpet is a thick woolen fabric with wool-like hairs woven to create a fluffy surface. Hard floors can be any of the aforementioned mats, carpets, or lifted floors.

[0051] This disclosure assumes cleaning by classifying floor characteristics into the six types described above. However, during implementation, only some of the types described above may be cleaned, and other types of floor surfaces may be cleaned in addition to the examples described above. For example, hard floors may be further subdivided into different types of flooring (marble, wood, concrete, tile, etc.).

[0052] Meanwhile, floor characteristics may be described in various terms such as floor material, type of surface to be cleaned, characteristics of surface to be cleaned, type of floor surface, etc., but in this disclosure, they are described as floor characteristics.

[0053] As described above, identifying the floor characteristics allows the suction module's operating conditions to be adjusted to suit those characteristics. Therefore, accurately identifying the floor characteristics is crucial.

[0054] However, in the case of a vacuum cleaner such as that shown in Fig. 1, a user can utilize the vacuum cleaner in various ways to perform cleaning tasks. That is, the operation of the vacuum cleaner while performing cleaning tasks can be divided into multiple operation sections.

[0055] For example, the user may clean by continuously pushing or pulling the vacuum cleaner (100), or may clean by repeatedly performing the pushing and pulling motions. In other words, there may be a pushing section and a pulling section, etc. In addition, when performing the pushing and pulling motions, there may be a section where the vacuum cleaner temporarily stops at the point where the pushing and pulling motions switch. In addition, the pushing and pulling sections may also be divided into an initial section, a middle section, a late section, etc.

[0056] Specifically, the plurality of cleaning sections may include at least one of various sections, such as at least one first section in which the cleaner (100) performs cleaning while moving in a first direction (for example, a pushing direction), at least one second section in which the cleaner (100) performs cleaning while moving in a second direction opposite to the first direction (for example, a pulling direction), a transition section in which the cleaner (100) switches from the first direction to the second direction (a section in which a pushing motion is switched to a pulling motion), a transition section in which the cleaner (100) switches from the second direction to the first direction (a section in which a pulling motion is switched to a pushing motion), a section in which the cleaner temporarily stops at a specific location while performing cleaning, and a section in which the cleaner stops at a specific location for a certain period of time.

[0057] In this disclosure, the operating section of the vacuum cleaner is described assuming that it operates in six types, but when implemented, the operating section may be divided into other types in addition to the examples described above.

[0058] The vacuum cleaner's suction power and other status data may vary depending on the cleaning section. Therefore, even if data sensed during the moving section and data sensed during the stationary section have similar or identical statistical characteristics, this may not necessarily indicate identical floor characteristics.

[0059] In other words, if data characteristics are not considered for each section of the vacuum cleaner, the accuracy of identifying floor characteristics may be reduced. If floor characteristics are not accurately identified, cleaning cannot be performed in an operating state appropriate for the floor characteristics, which can lead to various problems such as increased noise, damage to carpet fibers, and power waste.

[0060] Therefore, in various embodiments of the present disclosure, floor characteristics can be effectively identified by using multiple artificial intelligence models corresponding to each operating section of the vacuum cleaner.

[0061] Below, a vacuum cleaner (100) and its driving method according to various embodiments of the present disclosure are specifically described.

[0062] FIG. 2 is a block diagram illustrating a configuration of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0063] Referring to FIG. 2, the vacuum cleaner (100) may include a suction module (110), a memory (120), and a processor (130).

[0064] The suction module (110) is configured to suck dust or foreign substances from the floor into the vacuum cleaner (100). As described above, the suction module (110) can suck foreign substances by generating suction force through the driving of the driving device (12) included in the vacuum cleaner (100). The driving method of the suction module (110) has been described in detail in the above-described section, so a redundant description will be omitted.

[0065] The memory (120) is configured to store at least one instruction, O / S (Operating System), program, and data related to the vacuum cleaner (100). For example, an instruction for identifying floor characteristics, an instruction for determining the cleaning suction strength, an instruction for controlling various components of the vacuum cleaner, an instruction for retraining a plurality of pre-stored artificial intelligence models, and the like may be stored in the memory (120).

[0066] The memory (120) can store multiple pre-trained artificial intelligence models. For example, the artificial intelligence models can be implemented as 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., but are not limited to these examples.

[0067] These artificial intelligence models are computing systems implemented based on the neural networks of human or animal brains, and may also be referred to as learning models, machine learning models, neural network models, or deep learning models.

[0068] Specifically, multiple artificial intelligence models corresponding to multiple operation sections that divide the vacuum cleaner operation used for cleaning into multiple sections, or a section identification model learned for operation section identification, may be stored in the memory (120). Since examples of operation sections have been specifically described in the above-described section, a redundant description will be omitted.

[0069] These artificial intelligence models are described in detail in Figures 7 to 10.

[0070] The processor (130) is a component for controlling the overall operation of the vacuum cleaner (100). Specifically, the processor (130) can control the operation and components of the vacuum cleaner (100) by executing at least one instruction stored in the memory (120) as described above. For example, the processor (130) can drive the suction module (110) to suction foreign substances.

[0071] The processor (130) may be implemented as a digital signal processor (DSP), a microprocessor, a GPU (Graphics Processing Unit), an AI (Artificial Intelligence) processor, an NPU (Neural Processing Unit), etc. However, the present invention is not limited thereto, and may include one or more of a central processing unit (CPU), a MCU (Micro Controller Unit), an MPU (micro processing unit), a controller, an application processor (AP), a communication processor (CP), an ARM processor, or may be defined by the relevant term. In addition, the processor (150) may be implemented as a SoC (System on Chip), an LSI (Large Scale Integration) having a built-in processing algorithm, or may be implemented in the form of an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array).

[0072] The processor (130) may be implemented in a form integrated with the memory (120). In FIG. 2, each of the memory (120) and the processor (130) is illustrated as one, but the number of memories (120) and processors (130) may be implemented as multiple.

[0073] The processor (130) can collect status data related to the operation of the suction module and store it in the memory (120).

[0074] The status data may be data on the operating status of the vacuum cleaner (100) that operates to perform cleaning. Specifically, when cleaning is performed using a suction module, time series data on various items such as the internal pressure of the suction module (110) generated by the motor assembly, suction force, current and voltage required to drive the suction module (110) or brush, and rotation speed of the brush included in the suction module (110) may be used as the status data. The current and voltage data required to drive the suction module (110) or brush may be data on the magnitude of the current or voltage applied to operate the motor assembly. As described above, since the brush can be rotated by driving the second motor, the time series data on the current and voltage required to drive the brush may be data measuring the change in the magnitude of the current or voltage applied to the second motor over time.

[0075] However, examples of status data are not limited thereto, and any data that changes in various ways depending on the operation of the vacuum cleaner can be utilized as status data. When the processor (130) drives the suction module according to the user's operation, the processor (130) can directly secure data on changes in the magnitude of the current or voltage applied to the first or second motor of the motor assembly to drive the suction module as status data. In addition, the suction force that varies depending on the magnitude of the current or voltage or the rotation speed of the brush can be directly acquired and stored as status data in the memory (120). However, the present invention is not limited thereto, and the vacuum cleaner (100) may further include at least one sensor for sensing various status data. In this case, the processor (130) may acquire status data based on the sensing value sensed by each sensor and store the status data in the memory (120).

[0076] The processor (130) may store the state data itself in the memory (120), or may perform a preprocessing operation on the collected state data to extract it in the form of feature information and store it in the memory (120).

[0077] Specifically, the processor (130) may perform preprocessing operations, such as filtering operations to remove noise or data in unnecessary frequency bands from collected raw data, or normalizing operations to fit data within a preset normal size range. In addition, the processor (130) may perform various operations on preprocessed state data to extract feature information.

[0078] Feature information is information extracted to identify the characteristics of state data, including the time-series data described above. For example, feature information can include information such as variance, standard deviation, slope, and mean calculated for state data over a specific time interval, or information on change patterns, such as periodicity and peak size. This feature information is described in detail in Figure 3.

[0079] The processor (130) can identify the operating section of the cleaner (100) based on the above-described status data or characteristic information. That is, the processor (130) can identify whether the cleaner (100) is currently in a pushing section, a pulling section, a transition section therebetween, a stationary section, etc. If the characteristics of the status data measured for each operating section are pre-stored in the memory (120), the processor (130) can compare the status data or characteristic information with the information stored in the memory (120) and identify the operating section based on the degree of matching. However, the present invention is not limited thereto, and the processor (130) can also identify the operating section using a section identification model learned to identify the operating section. This will be described in detail again in the following section.

[0080] Once the motion section is identified, the processor (130) can identify the floor characteristics on which the cleaning task is performed using an artificial intelligence model corresponding to the identified motion section. For example, if the identified motion section is a sliding section, the processor (130) can select an artificial intelligence model that has learned various floor characteristics based on the sliding section, and input status data into the artificial intelligence model.

[0081] If there are multiple identified operation sections, the processor (130) may utilize all AI models corresponding to each operation section. For example, the processor (130) may input status data or feature information into each AI model corresponding to each operation section and obtain output values ​​of each AI model. The processor (130) may compare the reliability scores of the obtained output values ​​to identify the characteristics of the floor on which the cleaning task is performed.

[0082] The processor (130) may adjust the operating status of the suction module, such as determining the suction strength based on the identified floor characteristics. For example, if the identified floor characteristic is carpet, the processor (130) may determine a suction strength that has a lower suction strength than the suction strength when the floor is identified as hard floor. Furthermore, even carpets may be classified into short-pile carpets, long-pile carpets, etc., depending on the length of the pile, and may be classified into various types such as wool carpets, cotton carpets, and synthetic carpets depending on the material of the pile. The processor (130) may adjust the operating status differently depending on the detailed type of carpet. For example, if the identified floor characteristic is long-pile carpet, the processor (130) may determine a suction strength that has a lower suction strength than the suction strength when the floor is identified as short-pile carpet to prevent damage to the carpet due to the suction strength. Furthermore, even in the case of wool carpets, the suction strength may be lowered to prevent damage to the pile. On the other hand, synthetic fiber carpets are highly durable, so the suction strength may be maintained the same as that of a general floor or may be increased.

[0083] Here, the suction strength indicates the degree of suction of the vacuum cleaner. If the vacuum cleaner is divided into multiple stages, from minimum to maximum suction power, it may be information indicating one of the stages. For example, if three suction intensities (or suction stages) are supported, the determined suction strength may be one of the three suction intensities described above. This suction strength may also be referred to as suction power, suction level, suction intensity, cleaning intensity, or cleaning power.

[0084] Additionally, the processor (130) can determine the rotation speed of the brush based on the identified floor characteristic information as well as the suction strength. Once the rotation speed of the brush is determined, the processor (130) can drive the second motor so that the brush rotates at the determined rotation speed.

[0085] Meanwhile, when a user command to adjust the suction strength of the vacuum cleaner during operation with the suction strength determined through the above-described process is input, the processor (130) may adjust the current suction strength to a suction strength corresponding to the user command.

[0086] Additionally, the processor (130) can selectively use multiple artificial intelligence models in various ways to accurately identify motion sections and floor characteristics. These operations are described in detail in the sections 9 and 10 described below.

[0087] Although only a simple configuration of the vacuum cleaner (100) is illustrated above, other configurations not illustrated may be additionally provided during implementation. For example, at least one sensor, input device, etc. may be further included, so that sensing values ​​regarding the vacuum cleaner's status may be acquired through the sensor, or a user's control command may be input through the input device.

[0088] FIG. 3 is a diagram showing an example of current data, which is one of the status data of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0089] According to Fig. 3, current data is displayed in the form of a graph that changes in various sizes over time. As described above, the processor (130) can perform various preprocessing operations, such as filtering and normalization. Fig. 3 illustrates an example of current data after preprocessing is completed. The processor (130) can obtain characteristic information by periodically distinguishing and analyzing the change pattern of the current data.

[0090] Specifically, the processor (130) can divide the state data within a certain time interval (Tw, 320) into units of a certain cycle (Tp, 310) and extract characteristic information such as the variance, standard deviation, slope, mean, and change pattern of the data within each cycle. FIG. 3 illustrates a case where Tw is set to 2Tp. According to one embodiment, based on an arbitrary point in time t, the processor (130) can store the state data for the past Tw time in the form of an array in the memory (120). The processor (130) can repeat this operation in a cycle of Tp. Therefore, if Tw is set to 2Tp, half of the state data stored for each Tp cycle may be newly generated data, and the other half may be data stored in the previous cycle.

[0091] The processor (130) can determine which operation section a given section is based on characteristic information within each cycle. For example, when calculating a slope, the processor (130) can determine that a section is a pushing section if the slope of the current data increases in a positive direction, and can determine that a section is a section transitioning from a pushing section to a pulling section or vice versa if the slope increases rapidly in a negative direction.

[0092] The processor (130) may perform this judgment using an interval identification model.

[0093] FIG. 4 is a diagram illustrating the software structure of a vacuum cleaner according to at least one embodiment of the present disclosure. According to FIG. 4, a state data collection module (620), a preprocessing module (630), a section identification model (640), and the like may be stored in the memory (120) of the vacuum cleaner (100).

[0094] The processor (130) can collect status data related to the operation of the suction module by executing the status data collection module (620) while the vacuum cleaner (100) operates the suction module (110) to perform cleaning. If the vacuum cleaner (100) further includes a sensor (not shown), the status data collection module (620) can also obtain status data by analyzing the sensing value of the sensor.

[0095] The processor (130) can execute a preprocessing module (630) to perform preprocessing on the collected status data and extract characteristic information. The preprocessing module (630) can also assign weights to each status data and organize it into a data set. Furthermore, the data can be processed to suit the purpose of the interval identification model (640).

[0096] The processor (130) can identify operation section information by inputting the above-described state data or feature information into the section identification model (640).

[0097] The interval identification model (640) can be learned by state data for each interval or state data for multiple intervals.

[0098] FIG. 5 and FIG. 6 are diagrams for explaining an example of a learning method of an interval identification model using the interval identification model of FIG. 4.

[0099] The processor (130) performs preprocessing work on the collected status data and can divide the operating section of the cleaner into multiple sections based on the preprocessed data.

[0100] Figure 5 shows a state where the cleaning section is divided into four sections based on current data. Specifically, it can be divided into an initial push section (a) where the cleaner is pushed to move, a later push section (b) where a constant force is maintained to push the cleaner that is being pushed, a section (c) where the cleaner is temporarily stopped to transition from a push section to a pull section, and a pull section (d) where the temporarily stopped cleaner is pulled to move.

[0101] FIG. 5 illustrates a case where a specific time interval (Tw, 410) is set to be similar to or larger than the sum of the times of the aforementioned (a), (b), (c), and (d) intervals. In this case, the interval identification model (640) can be trained with a pattern that includes all four of the aforementioned cleaning intervals.

[0102] Figure 6 illustrates a case where a specific time interval (Tw, 510) is set similarly to the times of each of the aforementioned cleaning intervals (a), (b), (c), and (d). In this case, the interval identification model (640) can be selectively trained for each specific cleaning interval. In this case, an algorithm for interval determination can be utilized.

[0103] For example, the processor (130) can calculate the amount of change for the state data through methods such as variance and standard deviation, and if the amount of change is greater than a certain value, identify it as sections (a) and (c), and if the amount of change is less than the certain value, identify it as sections (b) and (d). In addition, the processor (130) can calculate the amount and direction of change through methods such as linear or polynomial regression, and if the amount of change is greater than a certain value and is positive, identify it as section (a), and if the amount of change is greater than a certain value and is negative, identify it as section (c). In addition, if the amount of change is less than a certain value and the previous section is (a), identify it as section (b), and if the amount of change is less than a certain value and the previous section is (c), identify it as section (d).

[0104] However, the method for learning the interval identification model is not limited to the above-described embodiment, and various methods such as deep learning can be applied.

[0105] The learned interval identification model (640) for motion interval identification can, when a preprocessed data set is input, identify an interval corresponding to the input data set and output a probability value indicating which interval the data set corresponds to. When the probability value is output, the motion interval with the highest probability value among the output information can be determined as the current motion interval. For example, the probability value corresponding to the pushing interval can be output as 0.7, and the probability value corresponding to the stopping interval can be output as 0.5. In this case, the motion interval can be determined as the pushing interval with the larger probability value.

[0106] In the above, the information input during the preprocessing task is described as status data related to the operation of the suction module, but it is not necessarily limited thereto, and some status data may include information input through a sensor, information stored in memory (120), etc., or preset data.

[0107] FIG. 7 is a drawing for explaining the operation of a vacuum cleaner according to at least one embodiment of the present disclosure.

[0108] The processor (130) can identify an operation section based on state data or feature information extracted therefrom.

[0109] If there are multiple identified operation sections, the processor (130) may input state data or feature information to each of the artificial intelligence models (740-1 to 740-n) corresponding to each identified operation section. Alternatively, if it is impossible to identify the operation section, the processor (130) may input state data or feature information to all or part of the stored artificial intelligence models.

[0110] Each artificial intelligence model (740-1 to 740-n) that has input information can output floor characteristic information and a reliability score thereof based on the input information.

[0111] The control module (750) is a software module for determining floor characteristics based on the output values ​​of each artificial intelligence model (740-1 to 740-n). For example, the control module (750) may sum the reliability scores obtained through each artificial intelligence model for each floor characteristic, and then determine the floor characteristic with the highest summed reliability score as the current floor characteristic.

[0112] Specifically, when the operating section is identified as a pushing section and a pause section in FIG. 7, the state data in the pushing section can be input to artificial intelligence model 1 (740-1), and the state data in the pause section can be input to artificial intelligence model 2 (740-2). Based on the input state data, if the output values ​​of artificial intelligence model 1 (740-1) are lift reliability 0.7 and mat reliability 0.3, and the output values ​​of artificial intelligence model 2 (740-2) are lift reliability 0.5 and mat reliability 0.5, the final reliability score summed for each material is lift 1.2 and mat 0.8, so the floor characteristic can be ultimately determined as a lift state with a large reliability score.

[0113] However, the operation of the control module (750) is not necessarily limited to this, and according to another embodiment, the control module (750) may determine the floor characteristic information having the highest reliability score among the output information as the current floor characteristic.

[0114] Specifically, when the operating section is identified as a pushing section and a pause section in FIG. 7, the status data in the pushing section can be input to artificial intelligence model 1 (740-1), and the status data in the pause section can be input to artificial intelligence model 2 (740-2). Based on the input status data, if the output value of artificial intelligence model 1 (740-1) is a lift state and a reliability of 0.7, and the output value of artificial intelligence model 2 (740-2) is a mat and a reliability of 0.5, the floor characteristic can be ultimately determined as a lift state with a large reliability score.

[0115] Once the floor characteristics are determined in this manner, the control module (750) can generate control information for adjusting the operating state of the suction module (110) based on the floor characteristic customized information stored in the memory (120). Accordingly, the cleaner (100) can automatically operate with a suction strength that matches the identified floor characteristic information even without separate operation by the user.

[0116] The floor characteristic customized information may include information on the suction strength of the suction module, the rotation speed of the brush, etc., which are set according to the floor characteristics. For example, if the floor characteristic information is input in a lift state, a control command to minimize the suction strength can be generated based on the floor characteristic customized information that minimizes the suction strength in the lift state.

[0117] Meanwhile, although the control module (750) is used as an example to adjust the suction strength, the control module (750) can also adjust various driving states such as the rotation speed of the brush in addition to the suction strength. The control information can be implemented in various forms of data such as a digital control code applied to each component, a PWM (Pulse Width Modulation) control signal, or an analog signal.

[0118] Meanwhile, in the above, it has been described that the control module (750) directly receives the output of the artificial intelligence model and directly determines the floor characteristics, but this is not limited to this, and the determination of the above-described floor characteristics may be performed by a separately provided judgment module or post-processing module. In other words, the control module (750) may be designed as a software module that receives the determined floor characteristics and only performs control operations accordingly.

[0119] FIG. 8 is a diagram illustrating an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using an artificial intelligence model.

[0120] According to FIG. 8, the vacuum cleaner can collect status data related to the operation of the suction module and perform a preprocessing task. When including a software module such as FIG. 4, the preprocessing module (630) can configure a data set (810) in the form of a 2D array of M features and N samples. Here, each feature means status data items or characteristic information such as current and pressure collected during the cleaning task, and a sample represents an observation value of each status data item or characteristic information collected during a specific time interval. Since the preprocessing process has been described in detail in FIG. 3, a redundant description will be omitted.

[0121] The data set (810) configured through the above-described preprocessing process can be input into various artificial intelligence models such as a neural network to determine floor characteristic information.

[0122] An artificial intelligence model may be composed of multiple neural network layers. Each layer has at least one weight value and performs its own operation based on 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.

[0123] The neural network of Fig. 8 may be trained by various learning algorithms such as supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, and by learning data such as state data for each floor characteristic.

[0124] Specifically, a neural network can generally be composed of an input layer (820), one or more hidden layers (830), and an output layer (840). The input layer (820) refers to the first layer that receives data in the neural network, the hidden layers (830) are layers that are located between the input layer and the output layer and play a key role in learning complex characteristics and patterns of data, and the output layer (840) refers to the final layer of the neural network that plays a role in representing the answer derived by the neural network to solve a problem.

[0125] According to Fig. 8, when a data set is input through each node of the input layer (820) of the artificial intelligence model, each node performs a preset operation using weights on the input data and sequentially outputs it to the next layer. The hidden layer (830) can learn the input information and extract floor characteristic information and reliability scores through the learned information. The information extracted from the hidden layer (830) can be finally output through the output layer (840).

[0126] The processor (130) can identify floor characteristics using software modules of various structures and artificial intelligence models as described above.

[0127] FIG. 9 is a diagram illustrating an example of an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using a plurality of artificial intelligence models.

[0128] Referring to FIG. 9, the processor (130) may input state data or feature information into one of a plurality of artificial intelligence models to initially identify floor characteristics. For example, the processor (130) may select an artificial intelligence model trained on various types of floor characteristics from among the plurality of artificial intelligence models to perform initial identification.

[0129] When the first identification is performed, the processor (130) can identify the current operating section using the section identification model, and can finally identify the floor characteristics by selectively using the artificial intelligence model corresponding to the first identified floor characteristics and the identified current operating section. In this case, the section identification model may be a model learned based on state data of various periods as described above, or may be a model learned based on state data for each section in a specific floor characteristic. That is, when the first identification result is a hard floor, the processor (130) can identify the section using an artificial intelligence model learned to identify the section using state data in the hard floor.

[0130] When a section is identified by the section identification model, the processor (130) finally identifies the floor characteristics using an artificial intelligence model corresponding to the identified section.

[0131] In FIG. 9, the processor (130) inputs state data or its characteristic information into artificial intelligence model 1 to first identify it as a hard floor, and then, when the current operating section is identified as a stop state through the section identification model, the process of inputting the first-identified floor characteristic information and section information into artificial intelligence model 3 corresponding to the stop state to finally identify whether the floor characteristic is a hard floor or a corner is illustrated.

[0132] FIG. 10 is a diagram illustrating another example of an operation of a vacuum cleaner according to at least one embodiment of the present disclosure to identify floor characteristics using multiple artificial intelligence models.

[0133] According to FIG. 10, the processor (130) inputs state data into an artificial intelligence model trained to detect multiple floor characteristics among multiple artificial intelligence models to initially identify the floor characteristics, and if the initially identified result is a carpet, the processor may use an artificial intelligence model trained to identify the type of carpet to finally identify the floor characteristics as either a long-pile carpet or a short-pile carpet.

[0134] Specifically, when the status data or feature information is input into an artificial intelligence model 1 trained to identify floor characteristics and the floor characteristic information is identified as carpet, the carpet can be identified as a long-haired carpet based on the status data or feature information identified into an artificial intelligence model 2 trained to identify carpet types.

[0135] As shown in Fig. 10, by identifying the length of the carpet hair, the operating state can be adjusted according to the hair length, thereby providing the user with a satisfactory level of AI function. For example, when a vacuum cleaner is used on a long-pile carpet, if the suction power is too strong, the carpet hair may be sucked into the suction port, blocking the suction port and preventing movement, or the carpet hair may be pulled out. Accordingly, when the processor (130) identifies a long-pile carpet, the processor (130) may lower the suction power to protect the carpet hair or improve the vacuum cleaner's performance.

[0136] FIG. 11 is a flowchart illustrating a method of operating a vacuum cleaner according to at least one embodiment of the present disclosure.

[0137] Referring to Figure 11, the suction module can be driven to suck up foreign substances (S1110). Specifically, the brush included in the suction module rotates to strike foreign substances such as dust stuck to the floor surface, and foreign substances separated from the floor surface can be easily sucked up.

[0138] In addition, status data related to the operation of the suction module can be collected and stored (S1120). For example, when the suction module is operated as described above, status data related to the operation of the suction module, such as the internal pressure of the suction module and the current required to operate the suction module, can be collected and stored.

[0139] In addition, at least one operation section corresponding to the status data can be identified among multiple operation sections (S1130). For example, information such as a pushing section or a pause section among the cleaning operation sections of a vacuum cleaner can be identified, and at this time, the section identification model described above can be used to identify the section.

[0140] Furthermore, the floor characteristics on which cleaning work is performed can be identified using an AI model corresponding to the identified motion section among multiple AI models (S1140). For example, when using an AI model corresponding to the identified motion section (i.e., the embodiment of FIG. 7), the floor characteristics on which cleaning work is performed can be identified more accurately.

[0141] Furthermore, the operating status of the suction module can be adjusted based on the identified floor characteristics (S1150). For example, if the floor characteristic is identified as a corner, the operating status of the suction module can be adjusted to suit the floor characteristics, such as by increasing the suction strength to suck up foreign substances such as dust on the floor surface.

[0142] The various embodiments described above may be implemented as a single embodiment, or at least one of the embodiments may be combined in whole or in part to be implemented together in one device.

[0143] According to the various embodiments described above, by using an artificial intelligence model that learns data for each operating section of a vacuum cleaner, the accuracy of judgment on floor characteristics can be increased, thereby providing consumers with a high level of satisfaction with the artificial intelligence function.

[0144] Various embodiments of the present disclosure may be implemented as software stored in a machine-readable storage media that can be installed or connected to a smartphone, a user terminal device, or other various electronic devices (e.g., a computer).

[0145] Specifically, when a vacuum cleaner performs a cleaning task, a non-transitory readable storage medium storing software for sequentially performing the steps of collecting status data related to the cleaning task, identifying at least one operation section corresponding to the status data among a plurality of operation sections in which the operation of the vacuum cleaner performing the cleaning task is divided into a plurality of sections, and identifying the characteristics of the floor on which the cleaning task is performed using an artificial intelligence model corresponding to the identified operation section among a plurality of artificial intelligence models that are learned and stored to correspond to each of the plurality of operation sections may be provided.

[0146] A device equipped with such a non-transitory readable medium can perform various operations, such as collecting state data, identifying an operating section, identifying floor characteristics, and adjusting the operating state of a suction module, as described in the various embodiments described above.

[0147] In the context of non-transitory readable storage media, 'non-transitory' means that the storage medium does not contain signals and is tangible, but does not distinguish between whether data is stored semi-permanently or temporarily on the storage medium.

[0148] Alternatively, a program for performing the method according to the various embodiments described above may be distributed online through an application store. In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily generated on a storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

[0149] Each component (e.g., a module or a program) according to various embodiments may be composed of one or more entities, and some of the aforementioned sub-components may be omitted, or other sub-components may be further included in various embodiments. Alternatively or additionally, some components (e.g., a module or a program) may be integrated into a single entity, which may perform the same or similar functions as those performed by each of the respective components prior to integration. Operations performed by a module, program, or other component according to various embodiments 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.

[0150] 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 skilled 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 the vacuum cleaner, suction module; A memory storing multiple artificial intelligence models corresponding to multiple operation sections that divide the vacuum cleaner operation used for cleaning into multiple sections; and Processor; including; The above processor By driving the above suction module, foreign substances are sucked in, Collect status data related to the operation of the above suction module and store it in the memory, Identifying at least one operation section corresponding to the state data among the above multiple operation sections, The floor characteristics of the floor surface on which the cleaning is performed are identified by using an artificial intelligence model corresponding to the identified operation section among the above multiple artificial intelligence models. A cleaner that adjusts the driving state of the suction module based on the identified floor characteristics.

2. In paragraph 1, The above processor, Performing preprocessing on the above state data to extract feature information, and identifying at least one operation section from among the plurality of operation sections based on the feature information, The above status data is, It includes time series data for at least one of the internal pressure of the suction module, the suction force, the current and voltage required to drive the suction module or the brush included in the suction module, and the rotation speed of the brush. The above feature information is, A cleaner that includes information on at least one feature among variance, standard deviation, slope, mean, and change pattern for data during a preset unit period in the above time series data.

3. In paragraph 2, The above memory is, To identify the motion segment, the learned segment identification model is stored. The above processor, A vacuum cleaner that inputs the above state data or the above feature information into the section identification model and identifies the operation section of the vacuum cleaner based on the output value thereof.

4. In paragraph 2, The above processor, A vacuum cleaner, wherein when there are multiple operation sections corresponding to the above state data, the state data or the feature information is input into each of the artificial intelligence models corresponding to each operation section, and the reliability scores of the output values of each of the artificial intelligence models are compared to identify the characteristics of the floor on which the cleaning task is performed.

5. In paragraph 2, The above processor, By inputting the state data or the feature information into one of the above multiple artificial intelligence models, the floor characteristics are first identified, The current operation section of the vacuum cleaner is identified using a section identification model learned and stored in the memory to identify the operation section. A vacuum cleaner that selectively uses an artificial intelligence model corresponding to the first identified floor characteristic and the current operating section among the plurality of artificial intelligence models to finally identify the floor characteristic.

6. In paragraph 1, The above processor, The above state data is input into an artificial intelligence model trained to detect floor characteristics among the above multiple artificial intelligence models to first identify the floor characteristics, A vacuum cleaner, wherein if the result identified in the first stage is a carpet, the vacuum cleaner uses an artificial intelligence model learned to identify the type of carpet among the plurality of artificial intelligence models to finally identify the floor characteristic as one of a long-pile carpet and a short-pile carpet.

7. In paragraph 1, The above multiple action sections are: The cleaner comprises at least one first section in which cleaning is performed while moving in a first direction, at least one second section in which cleaning is performed while moving in a second direction opposite to the first direction, a transition section in which the cleaner switches from the first direction to the second direction, a transition section in which the cleaner switches from the second direction to the first direction, a section in which the cleaner temporarily stops at a specific location while performing cleaning, and a section in which the cleaner stops at a specific location for a certain period of time or longer. The above floor characteristics are, A vacuum cleaner, wherein the vacuum cleaner is in a lift state off the floor surface, including at least one of a hard floor, a hard floor corner point, a long pile carpet, a short pile carpet and a mat.

8. In a method of operating a vacuum cleaner including a suction module, A step of driving the above suction module to suck up foreign substances; A step of collecting and storing status data related to the operation of the above suction module; A step of identifying at least one operation section corresponding to the state data among a plurality of operation sections in which the operation of the vacuum cleaner used for cleaning is divided into a plurality of sections, and identifying the floor characteristics of the floor surface on which the cleaning is performed by using an artificial intelligence model corresponding to the identified operation section among a plurality of artificial intelligence models that have been learned and stored to correspond to each of the plurality of operation sections; and An operating method comprising: a step of adjusting the driving state of the suction module based on the identified floor characteristics.

9. In paragraph 8, The step of identifying the above floor characteristics is: A step of extracting feature information by performing preprocessing work on the above state data; A step of identifying at least one operation section among the plurality of operation sections based on the above characteristic information; The above status data is, It includes time series data for at least one of the internal pressure of the suction module, the suction force, the current and voltage required to drive the suction module or the brush included in the suction module, and the rotation speed of the brush. The above feature information is, An operating method including information on at least one feature among variance, standard deviation, slope, mean, and change pattern for data during a preset unit period in the above time series data.

10. In paragraph 9, The step of identifying the above floor characteristics is: A step of inputting the state data or the feature information into a learned and stored section identification model to identify the motion section; An operating method, comprising: a step of identifying an operating section of the cleaner based on an output value of the section identification model.

11. In paragraph 9, The step of identifying the above floor characteristics is: If there are multiple operation sections corresponding to the above state data, a step of inputting the state data or the feature information into each of the artificial intelligence models corresponding to each operation section; and An operating method, comprising: a step of identifying the characteristics of a floor on which the cleaning work is performed by comparing the reliability scores of the output values of each of the artificial intelligence models; 12. In paragraph 9, The step of identifying the above floor characteristics is: A step of first identifying the floor characteristics by inputting the state data or the feature information into one of the plurality of artificial intelligence models; A step of identifying the current operation section of the vacuum cleaner by using a learned and stored section identification model to identify the operation section; An operating method, comprising: a step of selectively using an artificial intelligence model corresponding to the first identified floor characteristic and the current operating section among the plurality of artificial intelligence models to finally identify the floor characteristic; 13. In paragraph 8, The step of identifying the above floor characteristics is: A step of first identifying the floor characteristic by inputting the state data into an artificial intelligence model learned to detect the floor characteristic among the plurality of artificial intelligence models; An operating method comprising: a step of finally identifying the floor characteristic as one of a long-pile carpet and a short-pile carpet by using an artificial intelligence model learned to identify the type of carpet among the plurality of artificial intelligence models if the result identified in the first step is a carpet; 14. In paragraph 8, The above multiple action sections are: The cleaner comprises at least one first section in which cleaning is performed while moving in a first direction, at least one second section in which cleaning is performed while moving in a second direction opposite to the first direction, a transition section in which the cleaner switches from the first direction to the second direction, a transition section in which the cleaner switches from the second direction to the first direction, a section in which the cleaner temporarily stops at a specific location while performing cleaning, and a section in which the cleaner stops at a specific location for a certain period of time or longer. The above floor characteristics are, A method of operation, wherein the cleaner is in a lift state off the floor surface, including at least one of a hard floor, a hard floor corner point, a long pile carpet, a short pile carpet and a mat.

15. In a non-transitory readable recording medium having stored therein a program for performing an identification method for identifying floor characteristics by a vacuum cleaner, The above identification method is, When the above cleaner performs a cleaning task, a step of collecting status data related to the cleaning task; A step of identifying at least one operation section corresponding to the status data among a plurality of operation sections in which the operation of the cleaner performing the cleaning task is divided into a plurality of sections; and A step of identifying the floor characteristics on which the cleaning work is performed by using an artificial intelligence model corresponding to the identified action section among a plurality of artificial intelligence models that have been learned and stored to correspond to each of the plurality of action sections; A non-transitory readable recording medium comprising:

Images