Device for collecting or transmitting data for dehydration control or updating dehydration control model using collected data, or method therefor

Abstract

The present invention proposes a device configured to collect learning-required data for dehydration control and report the collected learning-required data to a server. The device may be configured to comprise: a memory configured to store code for collecting learning-required data while performing the dehydration control and for reporting the collected learning-required data to the server; and a processor configured to execute the code for performing an operation for collecting the learning-required data and reporting the collected learning-required data.

Description

Device or method for collecting, transmitting, or updating a dehydration control model using collected data for dehydration control

[0001] The present invention relates to a device or method for collecting, transmitting, or updating a dehydration control model using collected data for dehydration control. More specifically, the present invention relates to a device, method, or system configured to select and store data to be relearned from among data for dehydration control, and to update a dehydration control model by relearning the data.

[0002] The spin control of a washing machine is to remove moisture from laundry by controlling the rotation speed of the motor or drum.

[0003] Existing washing machine dehydration technologies utilizing reinforcement learning required model training based on experience gathered from actual washing machine operations. Therefore, implementing these technologies before mass production required a large number of washing machines and a significant amount of training resources. Furthermore, conducting experiments using actual washing machines presented significant challenges regarding experimental time.

[0004] To overcome this, this proposal aims to collect data from washing machines installed in various environments, manage the data in an integrated manner, develop data-based washing machine dehydration technology, and provide a method for improving performance.

[0005] The present invention proposes a device or method configured to collect or transmit data for dehydration control.

[0006] The present invention proposes a device or method configured to collect data for dehydration control or update a dehydration control model using the data.

[0007] In addition, the present invention proposes a system comprising a device configured to collect or transmit data for dehydration control and a device configured to update a dehydration control model using the data for dehydration control.

[0008] The problems to be solved by the present invention are not limited to the problems to be solved above, and other problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which the present invention belongs from the description below.

[0009] A device configured to collect learning-necessary data for dehydration control according to the present invention and report the collected learning-necessary data to a server is proposed, the device comprising: a memory configured to store code for collecting learning-necessary data while performing dehydration control and reporting the collected learning-necessary data to the server; and a processor configured to execute the code for performing an operation for collecting the learning-necessary data and reporting the collected learning-necessary data, wherein the operation may include collecting dehydration control data according to the performance of dehydration control following the initiation of a dehydration process, selecting learning-necessary data from among the collected dehydration control data, and transmitting the learning-necessary data to the server.

[0010] Additionally or alternatively, the learning-needed data may include, among the collected dehydration control data, dehydration control data collected for a dehydration process in which a vibration measurement value exceeds a reference value, or, among the collected dehydration control data, dehydration control data in which a difference between a probability value for a first command for the dehydration control and a probability value for a second command for the dehydration control is within a threshold value.

[0011] Additionally or alternatively, the probability value can be obtained through a dehydration control model for the dehydration control.

[0012] Additionally or alternatively, the operation may include tagging dehydration control data collected for a dehydration process in which a vibration measurement exceeds a reference value, or tagging specific dehydration control data among dehydration control data collected for a dehydration process in which the rotation of the drum does not reach a rotation target value within a preset time.

[0013] Additionally or alternatively, the learning needs data may include the tagged dehydration control data.

[0014] Additionally or alternatively, the specific dehydration control data may include dehydration control data in which a difference between a probability value for the first command for the dehydration control and a probability value for the second command for the dehydration control is within a threshold value in the collected dehydration control data.

[0015] Additionally or alternatively, the operation may include transmitting the learning need data to the server when the learning need data reaches a preset amount.

[0016] Additionally or alternatively, the dehydration control data may include status data including measurement sensor values ​​related to the dehydration control and operation data including probability values ​​for rotation increase and decrease commands for the dehydration control.

[0017] Additionally or alternatively, no learning need data may be selected for dehydration processes in which the vibration measurement does not exceed the reference value and the rotation of the drum reaches the rotation target value within a preset time.

[0018] Additionally or alternatively, the operation may include deleting the collected dehydration control data or the learning-needs data after the learning-needs data has been transmitted to the server.

[0019] Additionally or alternatively, the learning needs data may be used to update a dehydration control model for the dehydration control.

[0020] According to another embodiment of the present invention, a method is proposed that is configured to collect learning-necessary data for dehydration control and report the collected data to a server, the method comprising: a step of collecting dehydration control data according to the performance of dehydration control upon initiation of a dehydration process; a step of selecting learning-necessary data from the collected dehydration control data; and a step of transmitting the learning-necessary data to the server.

[0021] According to another embodiment of the present invention, a computer-readable medium storing code configured to execute by a computer or processor a method configured to collect learning-necessary data for dehydration control as described above and report the collected data to a server is proposed.

[0022] According to another embodiment of the present invention, a system is proposed that is configured to collect and report learning-necessary data for dehydration control, and to relearn the learning-necessary data to update a dehydration control model for the dehydration control, the system comprising: a client device configured to collect and report learning-necessary data; and a server device configured to relearn the learning-necessary data to update the dehydration control model, wherein the client device may be configured to include a memory configured to store code for collecting learning-necessary data while performing dehydration control and reporting the collected learning-necessary data to the server; and a processor configured to execute the code for performing an operation for collecting the learning-necessary data and reporting the collected learning-necessary data.

[0023] The above problem solving methods are only some of the embodiments of the present invention, and various embodiments reflecting the technical features of the present invention can be derived and understood by a person having ordinary knowledge in the relevant technical field based on the detailed description of the present invention described below.

[0024] The present invention has the following technical effects.

[0025] By collecting and learning data related to real-world dehydration control from multiple washing machines or multiple dehydration control devices, a model for dehydration control that takes into account various environments can be built.

[0026] In addition, rather than collecting all dehydration control-related data, it is efficient in terms of data transmission, reception, storage, and management by collecting it selectively.

[0027] Additionally, better dehydration performance can be achieved by updating the dehydration control model.

[0028] The effects according to the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art to which the present invention pertains from the detailed description of the invention below.

[0029] The accompanying drawings, which are included as part of the detailed description to aid in understanding the present invention, provide embodiments of the present invention and, together with the detailed description, explain the technical idea of ​​the present invention.

[0030] Figure 1 illustrates a perspective view of a washing machine according to the present invention.

[0031] Figure 2 illustrates a block diagram of a washing machine according to the present invention.

[0032] Figure 3 shows the dehydration control process and the rotation speed of the drum or motor according to the present invention.

[0033] Figure 4 shows an overall flow chart consisting of data collection, transmission, learning, and dehydration control model update and distribution for dehydration control according to the present invention.

[0034] Figures 5 to 7 illustrate a flowchart for a data collection and tagging method for dehydration control according to the present invention.

[0035] Figure 8 illustrates a flowchart for a data transmission method for dehydration control according to the present invention.

[0036] Figure 9 illustrates a flowchart for the mutual operation between a washing machine or a dehydration control device and a server according to the present invention.

[0037] Figure 10 is a graph showing dehydration performance according to an updated dehydration control model obtained by relearning data collected for dehydration control according to the present invention.

[0038] FIG. 11 illustrates a block diagram of a device for data collection and tagging or transmission for dehydration control according to the present invention.

[0039] Figure 12 illustrates a block diagram of a device for updating a dehydration control model for dehydration control according to the present invention.

[0040] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components will be given the same reference numbers and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, a detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present invention.

[0041] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0042] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0043] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0044] In this application, 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 specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0045]

[0046] Fig. 1 illustrates a perspective view of a washing machine according to the present invention. Fig. 2 illustrates a block diagram of a washing machine according to the present invention.

[0047] A washing machine according to one embodiment of the present invention may include a cabinet (10) forming an exterior, a tub (not shown), a drum (30), and a door (60) provided to open and close the drum (30) to load or remove clothing, which is an object to be treated, into or from the drum. Accordingly, it can be said that the door is provided to open and close the object loading port (61) of the cabinet (10).

[0048] A tub is provided inside the cabinet (10) to accommodate the drum (30). The drum (30) is rotatably provided inside the tub and accommodates laundry. An opening is provided at the front of the drum (30), and laundry is fed into the drum (30).

[0049] The washing machine further includes a driving unit configured to rotate the drum (30) within the tub. The driving unit includes a motor (41), and the motor (41) includes a stator and a rotor. The rotor is connected to a rotation shaft, and the rotation shaft is connected to the drum (30) so as to rotate the drum (30) within the tub.

[0050] The washing machine according to the present embodiment may include a user interface (UI) 80. The UI may include various buttons, a rotary knob, and more particularly, a display. Through the UI, the user may input object processing information to the washing machine. Furthermore, through the UI, the washing machine may provide the user with information on the object processing currently being performed and entered by the user.

[0051] In particular, the display can be implemented as a touch display, allowing both user input and display of information on the washing machine.

[0052] The display can display text, numbers, or images, and, as described below, time-series images, augmented reality images, or animations. Therefore, users can intuitively grasp the current processing information and status of the washing machine.

[0053] When a user selects a specific washing course through the UI (80), the control unit (100) performs washing according to the selected washing course.

[0054] First, the water supply valve (23) is controlled to supply washing water to the tub. Then, the water level sensor (26) is controlled to supply an appropriate amount of washing water to the tub.

[0055] When the water supply is finished, the control unit drives the motor (41) to perform washing. That is, washing is performed by rotating the drum and using detergent, washing water, and the mechanical power of the drum. At this time, the circulation pump (24) may be operated to increase washing efficiency. The circulation pump (24) performs the function of pumping washing water from the bottom of the tub and resupplying it to the top of the drum. Since washing is not performed while the laundry inside the drum is submerged in the washing water, the detergent water is supplied to the laundry more effectively, thereby increasing washing efficiency.

[0056] The washing machine according to the present embodiment may include a communication module (90). Through the communication module (90), the washing machine can communicate with an external server and transmit and receive information. The washing machine can transmit and receive information with a user's terminal through the external server.

[0057] For example, a user can input remote control commands via an external terminal. These remote control commands are transmitted to the washing machine via a server, allowing the washing machine to be remotely controlled.

[0058] When a user remotely commands a washing machine to process clothes, the washing machine can transmit current status information to a server while processing the clothes. The server can then transmit this information to the user's external terminal. Therefore, the user can easily access current laundry processing information through the external terminal.

[0059] In addition, the washing machine can perform updates of software or firmware, etc. from the server through the communication module (90). As described below, the washing machine according to one embodiment of the present invention can collect learning data for spin-drying control. This learning data can be used to update a spin-drying control model for spin-drying control. The updated spin-drying control model can be shared or updated through the server. For example, the AI ​​module (200) can store the spin-drying control model received from an external server, or transmit the spin-drying control value according to the spin-drying control model, specifically, a command for increasing or decreasing the rotation speed of the drum or motor, to the control unit (100).

[0060] Details will be provided later.

[0061] The above washing machine may include a vibration sensor (70).

[0062] The rotating shaft that rotates the drum (30) penetrates the tub and is connected to the drum provided within the tub. Accordingly, the vibration of the drum is transmitted to the tub. The vibration transmitted to the tub is transmitted to the cabinet (10), and as the drum vibrates, the entire washing machine vibrates.

[0063] A vibration damping device may be provided to reduce the transmission of drum vibrations to the cabinet through the tub. The vibration damping device may include a spring and a damper.

[0064] However, the vibration reduction effect of these vibration damping devices is inevitably limited. Therefore, when the drum rotates at high speeds, significant vibrations are generated, which are inevitably transmitted to the tub and cabinet. This excessive vibration can be exacerbated when laundry inside the drum is not evenly distributed and remains eccentric.

[0065] Accordingly, a vibration sensor or UB sensor (70) may be provided to detect excessive vibration when it occurs. The vibration sensor may be provided to detect the amplitude in a normal state (when the tub is stationary). In order to optimally detect the vibration amount of the tub, the vibration sensor (70) may be provided at the upper end of the tub. In particular, the vibration sensor (70) may be provided at the upper rear end or the upper front end of the tub.

[0066] Meanwhile, according to one embodiment of the present invention, an acceleration sensor or a gyro sensor (75) may be included. The gyro sensor (75) can sense linear displacement along three axes and angular displacement along three axes. Therefore, it can be referred to as a six-axis sensor. Acceleration changes can be calculated through changes in linear displacement and angular displacement for each axis.

[0067] The above-mentioned gyro sensor (75) can effectively detect and calculate the results of vibration. This is because vibration physically occurs in three dimensions, and thus all vibration displacements can be detected and calculated through the six-axis sensor. In other words, the results of vibration generation can be detected and calculated as a whole.

[0068] In order to effectively detect the vibration results of the tub, the gyro sensor (75) is also preferably provided at the top of the tub. That is, it is preferably located at the very top of the tub. In addition, for effective displacement detection, the gyro sensor (75) is preferably located near the rear or front end of the tub.

[0069] Here, it can be seen that the gyro sensor (75) is a type of vibration sensor. Therefore, by applying the gyro sensor (75), the vibration sensor (70) described above may be omitted. This is because the vibration sensor (70) may output any one of a plurality of displacements output from the gyro sensor (75), for example, a vertical linear displacement.

[0070] However, the vibration sensor (70) may be installed at a different location from the gyro sensor (75) to sense the vibration value. In addition, a plurality of vibration sensors (70) may be installed at the front and rear ends of the tub to sense the vibration value more accurately through the phase difference.

[0071] Dehydration can be defined as the process of centrifugally separating moisture from clothing by rotating a drum at high speed. Therefore, it is desirable to perform high-speed dehydration after the clothing is evenly distributed within the drum. In other words, the flow of clothing must be evenly distributed within the drum before high-speed dehydration is performed. This is crucial for preventing vibration and noise through eccentricity relief and for system protection, as well as for effective dehydration. This is because if the fabric is not properly distributed, the high-speed dehydration process can be delayed or even fail. This not only prevents proper dehydration but also increases the total washing time. Furthermore, there is a risk of incomplete dehydration, which can lead to a decrease in the dehydration effect and user satisfaction.

[0072] For this reason, it is very important to determine whether the foam is properly distributed before performing high-speed dehydration and to perform high-speed dehydration by properly distributing the foam.

[0073] The above washing machine may include a motor control module (45) that controls the operation of the motor (41). The motor control module (45) may control the current value and voltage value applied to the motor so that the motor rotates at a target RPM to rotate the drum.

[0074] The above motor control module (45) may be provided to directly control the operation of the motor through the control of the control unit (processor or main processor, 100). In addition, the motor control module (45) may be provided to calculate the current RPM of the motor and the current value currently applied to the motor through feedback control. That is, the motor control module (45) may output the current drum RPM and the current value applied to the motor.

[0075] The control unit (100) transmits the target RPM of the drum, i.e., the requested RPM or the command RPM, to the motor control module (45) according to the control sequence, and the motor control module (45) controls the current RPM to follow the requested RPM through feedback control. Of course, depending on the requested RPM, a plurality of dehydration sections can be distinguished as in Fig. 3. Fig. 3 shows the process of dehydration control and the rotation speed of the drum or motor according to the present invention. Referring to the illustrated rotation speed of the drum or motor, the rotation speed is not controlled linearly in the dehydration section.

[0076] Accordingly, the control unit knows the current requested RPM and, through the motor control module (45), knows the current drum RPM and the current value applied to the motor.

[0077] In an ideal, vibration-free environment, the current applied to the motor, the requested RPM, and the current RPM can be identically matched. That is, the applied current value corresponding to a specific requested RPM is specified, and when a specific current value is applied, the current RPM can be said to be equal to the specific requested RPM. In other words, the requested RPM and the current RPM can be substantially identical.

[0078] However, vibration inevitably occurs, and the greater the vibration value, the greater the discrepancy between the requested RPM, the applied current, and the current RPM. Of course, this discrepancy can be minimized through feedback control. However, feedback control is different from vibration suppression or excessive vibration prevention. In other words, while performing feedback control, drum rotation must be controlled to prevent excessive vibration.

[0079] As the drum rotates, vibration inevitably occurs. This vibration increases as the eccentricity of the laundry inside the drum increases. Furthermore, for the same eccentricity, the higher the drum rotation speed, the greater the vibration.

[0080] Therefore, the prerequisite for vibration generation is drum rotation, and the factor that rotates the drum and determines its RPM is the current value applied to the motor. The value corresponding to the current value is the requested RPM, and the current RPM varies in conjunction with the vibration.

[0081] Therefore, the above current value, requested RPM and current RPM can be said to be vibration inducing factors.

[0082] Meanwhile, when vibration occurs, the vibration value can be detected through a vibration sensor. In other words, the UB value can be considered a vibration result factor detected through the vibration sensor. Furthermore, since vibration occurs through three-axis linear displacement and three-axis angular displacement, the six values ​​detected through the gyro sensor can also be considered vibration result factors.

[0083] In an embodiment according to the present invention, a vibration inducing factor or a vibration resulting factor can be configured as a single data set, which can be referred to as “status data” related to spin-drying control. For example, the status data for spin-drying control can include at least one of the current RPM of the drum, the command RPM (target RPM) of the drum, the degree of balance of the laundry (UB), X, Y, Z-axis gravity sensor values ​​(GyroX, GyroY, GyroZ), X, Y, Z-axis acceleration sensor values ​​(AccX, AccY, AccZ), X, Y, Z-axis displacement of the front of the washing machine (DistFx, DistFy, DistFz), and Y, Z-axis displacement of the rear of the washing machine (DistRy, DistRz). In addition, the status data for spin-drying control can further include at least one of current data (Iq) applied to the motor, the temperature of the motor, the water level frequency in the drum, the X-axis displacement of the rear of the washing machine, and laundry quality analysis data.

[0084] The control unit (100) can perform dehydration control based on the above-described status data. More specifically, a dehydration control model for controlling the optimal RPM value of the drum that can perform the dehydration process while maintaining the laundry balance based on the above-described status data can be used for dehydration control. This dehydration control model can be generated through deep reinforcement learning and can be updated by relearning status data additionally collected according to the operation or use of the washing machine. The dehydration control model can be included in the AI ​​module (200). The dehydration control model can be generated and updated from an external server and transmitted to the washing machine through the communication module (90) described above.

[0085] The control unit (100) can be configured to control the AI ​​module (200), i.e., the motor control module (45) according to the dehydration control model.

[0086]

[0087] FIG. 4 illustrates an overall flowchart for a dehydration control model update, comprising data collection, transmission, learning, and dehydration control model update and distribution according to the present invention. The dehydration control model update according to FIG. 4 may be performed by a device for dehydration control model update. For example, the device for dehydration control model update may include the washing machine described above, but the present invention is not limited thereto. For example, the dehydration control device may also be implemented as a software module or control module attached or installed to the washing machine. The dehydration control device according to the present invention will be described below with reference to FIG. 11.

[0088] In another embodiment, the dehydration control model update according to FIG. 4 may be performed by two devices for dehydration control model update. For example, the update may be performed by a washing machine or a software module or control module attached or installed to the washing machine, and a server device. In this case, data collection and tagging (S410) and training-required data transmission (reporting) (S420) may be performed by the washing machine or a software module or control module attached or installed to the washing machine, while retraining (S430) and dehydration control model update and distribution (S440) may be performed by the server device.

[0089] Below, a description will be given of cases where a dehydration control model update is performed by a washing machine, a software module or control module attached or installed to the washing machine, and a server device. Furthermore, the washing machine, or a software module or control module attached or installed to the washing machine, will be referred to simply as a washing machine.

[0090] The washing machine can be configured to perform dehydration control according to a dehydration control model, that is, to collect dehydration control data during the dehydration process and to tag specific data (S410).

[0091] In this way, tagging specific data from the collected dehydration control data can be referred to as the process of selecting data required for learning. The dehydration control data will be described later with reference to Tables 1 and 2.

[0092] The washing machine may be configured to transmit or report tagged collected data, i.e., learning-required data, to a server device (S420). The transmission or reporting of the learning-required data may occur when a preset condition is satisfied. For example, the tagged collected data may be transmitted or reported to the server device when the amount of learning-required data exceeds a threshold amount. Frequent transmission of learning-required data may waste transmission resources, and frequent retraining of a small amount of learning-required data may increase the computational load or power consumption of the server device. Therefore, the washing machine may be configured to wait until the tagged collected data exceeds a threshold amount before transmitting or reporting to the server device.

[0093] The server device may be configured to relearn the received tagged collection data to update the dehydration control model (S430). In this case, the server device may be configured to relearn not only the tagged collection data received from a single washing machine, but also the tagged collection data received from multiple washing machines.

[0094] Additionally, the server device can be configured to update the dehydration control model and distribute the updated dehydration control model to the washing machine(s) (S440).

[0095] In this way, a dehydration control model capable of retraining can overcome the environmental and temporal limitations of a dehydration control model (hereinafter, "basic dehydration control model") created or constructed in a laboratory during the development phase of a washing machine. Specifically, since it is difficult to perform dehydration control in a laboratory setting based on various user environments, a basic dehydration control model can be created or constructed by analyzing status and operation data collected during the dehydration process for a limited number of laundry samples. According to the updating of the dehydration control model according to the present invention, the basic dehydration control model can be updated using the results of dehydration control of multiple washing machines, i.e., tagged collected data. Accordingly, the amount of learning data can be expanded, and the retrained dehydration control model is expected to exhibit superior dehydration control performance. The performance of the dehydration control model according to the present invention is illustrated in FIG. 10 .

[0096] In addition, if data collected indiscriminately in the usage environment of an actual washing machine is collected, a lot of resources may be consumed in terms of data transmission, maintenance, or management, but the present invention transmits or reports only specific collected data, i.e., tagged collected data, according to preset conditions, so that high-quality learning data can be collected.

[0097] In addition, if additional user information can be obtained, a customized dehydration control model for each user can be created, updated, and distributed using the tagged collected data, thereby providing hyper-personalized dehydration control.

[0098] Additionally, there is an advantage in reducing development resources even during the development stage of the dehydration control model.

[0099]

[0100] Figures 5 to 7 illustrate a flowchart for a data collection and tagging method for dehydration control according to the present invention. As previously described, data collection and tagging for dehydration control can be performed by the washing machine.

[0101] Data collection and tagging will be described with reference to Figure 5.

[0102] The washing machine can initiate spin-drying control or initialize spin-drying control (S4110). The distinction between initiating spin-drying control and initializing spin-drying control is made because, even if the spin-drying cycle has not begun, if the drum or tub of the washing machine cannot rotate at the commanded RPM due to vibration, spin-drying control is performed again from the initial commanded RPM (e.g., 0).

[0103] The washing machine may be configured to perform dehydration control using a dehydration control model and collect dehydration control data, such as status data (S4120). The dehydration control data collected by the washing machine may include not only status data but also operation data. The operation data represents the results inferred by the dehydration control model with the status data as input. Specifically, it may include a probability value for increasing the RPM from the current RPM and a probability value for decreasing the RPM.

[0104] The following shows an example of status data.

[0105] Time information Dehydration control model name Status / operation classification information Current RPM of the drum Command RPM of the drum Balance of the drum 1st sensor value 2nd sensor value… Nth sensor value T1M1S2031-3812740… 5

[0106] Additionally, the following shows an example of motion data.

[0107] Time information Dehydration control Model name Status / operation classification information… RPM increase probability RPM decrease probability T2M1A… -0.3102-0.3043

[0108]

[0109] State data and motion data form a pair. That is, RPM information, the degree of balance of the laundry, sensor values, etc. of the state data are used as input data of the spin-drying control model according to the present invention, and motion data can be obtained as output data of the spin-drying control model. Therefore, the washing machine can perform spin-drying control using a pair of state data and motion data. The unit that distinguishes a pair of state data and motion data can be referred to as a "step." The washing machine can perform spin-drying control through multiple steps.

[0110] The washing machine derives operation data based on state data from K seconds prior to the present, which can be repeated or performed according to a preset cycle.

[0111] The washing machine performs control in the direction indicated by the larger probability value of the two probability values ​​(i.e., increase or decrease). If the difference between the two probability values ​​is large, the reliability of the inference result is expected to be high. However, if the two probability values ​​are the same or similar, the reliability of the inference result is expected to be low.

[0112] The washing machine may be configured to determine whether vibration measurements obtained from a vibration sensor or similar device exceed a preset threshold (S4130). If the vibration measurements do not exceed the threshold, the data collection and tagging (S410) process may be terminated.

[0113] If the vibration measurement exceeds the above threshold, the washing machine may be configured to determine that the spin control attempt has failed. If the vibration measurement exceeds the above threshold, the washing machine may be configured to perform tagging on data collected since the start or initialization of spin control (S4140).

[0114] As an example, tagging may be performed on state data and motion data (i.e., collected data) collected after the start or initialization of dehydration control.

[0115] As another example, tagging may be performed on collected data for a preset or predetermined period of time after the start or initialization of dehydration control. For example, tagging may be performed on collected data up to a preset period of time prior to the point at which a vibration measurement is determined to exceed the threshold.

[0116] The tagged collected data may be stored locally on the washing machine. Furthermore, if the washing machine is equipped with an image sensor capable of recognizing the type of laundry inside the drum or using other means, the washing machine may be configured to add the recognized laundry type information to the tagged collected data.

[0117] The washing machine may be configured to determine whether the number of spin-drying control attempts exceeds a preset number (S4150). If the number of spin-drying control attempts exceeds the preset number, the washing machine may be configured to determine that the spin-drying control has failed. Accordingly, the method illustrated in FIG. 5 may be terminated. If the number of spin-drying control attempts does not exceed the preset number, the washing machine may be configured to initialize the spin-drying control. Accordingly, the washing machine may be configured to return to the spin-drying cycle at S4110.

[0118] When the data collection and tagging method illustrated in Figure 5 is performed, tagged collected data can be stored locally on the washing machine. Of course, in some cases, the collected data may not be tagged. The tagged collected data can be used as training data for future updates to the dehydration control model.

[0119]

[0120] Data collection and tagging will be described with reference to Figure 6.

[0121] The washing machine can initiate spin-drying control or initialize spin-drying control (S4110). The distinction between initiating spin-drying control and initializing spin-drying control is made because, even if the spin-drying cycle has not begun, if the drum or tub of the washing machine cannot rotate at the commanded RPM due to vibration, spin-drying control is performed again from the initial commanded RPM (e.g., 0).

[0122] A washing machine may be configured to perform dehydration control using a dehydration control model and collect dehydration control data such as status data (S4120). The dehydration control data collected by the washing machine may include not only status data but also operation data. The operation data represents a result inferred by the dehydration control model with the status data as input, and specifically may include a probability value for increasing the RPM from the current RPM and a probability value for decreasing the RPM. The washing machine performs control in the direction indicated by the larger probability value of the two probability values ​​(i.e., increasing or decreasing). If the difference between the two probability values ​​is large, the reliability of the inference result is expected to be high. However, if the two probability values ​​are the same or similar, the reliability of the inference result may not be expected to be high.

[0123] The washing machine may be configured to determine whether the rotation speed of the drum or motor has reached a rotation target, i.e., a command RPM, within a preset time (S4160). If the rotation speed of the drum or motor has reached the rotation target within the preset time, the data collection and tagging procedure (S410) may be terminated. To this end, the washing machine may be configured to initiate a timer whenever spin control is initiated or initialized.

[0124] If the rotation speed of the drum fails to reach the rotation target within a preset time, the washing machine may be configured to determine that the spin control attempt has failed. If the rotation speed of the drum or motor fails to reach the rotation target within the preset time, the washing machine may be configured to perform tagging on data collected since the start or initialization of spin control (S4170). However, unlike FIG. 5, the washing machine may be configured to perform tagging only on uncertainty data among the collected data.

[0125] Uncertainty data corresponds to data that differs significantly from the distribution of data used to train the existing model (baseline).

[0126] As an example, uncertainty data may include motion data in which the difference between the probability value for increasing the RPM from the current RPM and the probability value for decreasing the RPM among the collected motion data is within a threshold value.

[0127] As another embodiment, the uncertainty data may include a set of motion data in which the difference between the probability value for increasing the RPM from the current RPM and the probability value for decreasing the RPM among the collected motion data is within a threshold value more than a preset number. That is, the washing machine may be configured to tag all motion data among the data collected since the start or initialization of the spin-drying control, or to tag N motion data in which the difference between the two probability values ​​is within a threshold value, when there are N or more motion data (N is a natural number) among the data collected since the start or initialization of the spin-drying control.

[0128] The tagged collection data can be stored locally on the washing machine.

[0129] Additionally, if the washing machine is equipped with an image sensor to recognize the type of laundry inside the drum or can recognize the type of laundry inside the drum through other means, it may be configured to add the recognized laundry type information to the tagged collected data.

[0130] The washing machine may be configured to determine whether the number of spin-drying control attempts exceeds a preset number (S4180). If the number of spin-drying control attempts exceeds the preset number, the washing machine may be configured to determine that the spin-drying control has failed. Accordingly, the method illustrated in FIG. 6 may be terminated. If the number of spin-drying control attempts does not exceed the preset number, the washing machine may be configured to initialize the spin-drying control. Accordingly, the washing machine may be configured to return to the spin-drying cycle at S4110.

[0131] When the data collection and tagging method illustrated in FIG. 6 is performed, tagged collected data can be stored locally in the washing machine. The difference between FIG. 6 and FIG. 5 lies in the data being tagged. Unlike FIG. 5, FIG. 6 performs tagging on data with high uncertainty. However, in the method of FIG. 6, tagging may be performed not only on uncertainty data, but also on state data associated with the uncertainty data (i.e., operation data) (i.e., state data that forms the basis of the operation data). In other words, tagging may be performed on state data and operation data of one step.

[0132] Of course, in some cases, collected data may not be tagged. Tagged collected data can be used as training data for future updates to the dehydration control model.

[0133]

[0134] Data collection and tagging will be described with reference to Figure 7.

[0135] The data collection and tagging method of Fig. 7 combines the data collection and tagging methods of Figs. 5 and 6. After the data collection and tagging corresponding to Fig. 5 are sequentially performed, the data collection and tagging corresponding to Fig. 6 can be performed.

[0136] The washing machine can initiate spin-drying control or initialize spin-drying control (S4110). The distinction between initiating spin-drying control and initializing spin-drying control is made because, even if the spin-drying cycle has not begun, if the drum or tub of the washing machine cannot rotate at the commanded RPM due to vibration, spin-drying control is performed again from the initial commanded RPM (e.g., 0).

[0137] The washing machine may be configured to perform dehydration control using a dehydration control model and collect dehydration control data such as status data (S4120). The dehydration control data collected by the washing machine may include not only status data but also operation data.

[0138] The washing machine may be configured to determine whether vibration measurements obtained from a vibration sensor or similar device exceed a preset threshold (S4130). If the vibration measurements do not exceed the threshold, the data collection and tagging process may be terminated.

[0139] If the vibration measurement exceeds the above threshold, the washing machine may be configured to determine that the spin control attempt has failed. If the vibration measurement exceeds the above threshold, the washing machine may be configured to perform tagging on data collected since the start or initialization of spin control (S4140).

[0140] As an example, tagging may be performed on state data and motion data (i.e., collected data) collected after the start or initialization of dehydration control.

[0141] As another example, tagging may be performed on collected data for a preset or predetermined period of time after the start or initialization of dehydration control. For example, tagging may be performed on collected data up to a preset period of time prior to the point at which a vibration measurement is determined to exceed the threshold.

[0142] The tagged collected data may be stored locally on the washing machine. Furthermore, if the washing machine is equipped with an image sensor capable of recognizing the type of laundry inside the drum or using other means, the washing machine may be configured to add the recognized laundry type information to the tagged collected data.

[0143] The washing machine may be configured to determine whether the number of spin-drying control attempts exceeds a preset number (S4150). If the number of spin-drying control attempts exceeds the preset number, the washing machine may be configured to determine that the spin-drying control has failed. Accordingly, the method illustrated in FIG. 7 may be terminated. If the number of spin-drying control attempts does not exceed the preset number, the washing machine may be configured to initialize the spin-drying control. Accordingly, the washing machine may be configured to return to the spin-drying cycle at S4110.

[0144] The washing machine may be configured to determine whether the rotation speed of the drum or motor has reached a rotation target, i.e., a command RPM, within a preset time (S4160). If the rotation speed of the drum or motor has reached the rotation target within the preset time, the data collection and tagging process may be terminated. To this end, the washing machine may be configured to initiate a timer whenever spin control is initiated or initialized.

[0145] If the rotation speed of the drum or motor fails to reach the rotation target within a preset time, the washing machine may be configured to determine that the spin control attempt has failed. If the rotation speed of the drum or motor fails to reach the rotation target within a preset time, the washing machine may be configured to perform tagging on uncertainty data (S4170).

[0146] As an example, uncertainty data may include motion data in which the difference between the probability value for increasing the RPM from the current RPM and the probability value for decreasing the RPM among the collected motion data is within a threshold value.

[0147] As another example, the uncertainty data may include a set of motion data in which the difference between the probability value for increasing the RPM from the current RPM and the probability value for decreasing the RPM among the collected motion data is within a threshold value, and is greater than a preset number of motion data.

[0148] The tagged collection data can be stored locally on the washing machine.

[0149] The washing machine may be configured to determine whether the number of spin-drying control attempts exceeds a preset number (S4180). If the number of spin-drying control attempts exceeds the preset number, the washing machine may be configured to determine that the spin-drying control has failed. Accordingly, the method illustrated in FIG. 7 may be terminated. If the number of spin-drying control attempts does not exceed the preset number, the washing machine may be configured to initialize the spin-drying control. Accordingly, the washing machine may be configured to return to the spin-drying cycle at S4110.

[0150] According to the procedures of FIGS. 5 to 7, if it is determined that the dehydration control has failed, the dehydration process may be terminated. In addition, according to the procedures of FIGS. 6 to 7, if the rotation speed of the drum or motor according to the dehydration control reaches the rotation target value, or if the vibration measurement according to the dehydration control does not exceed the reference value and the rotation speed of the drum or motor reaches the rotation target value, the dehydration process may be determined to be successful, and the dehydration process may be terminated.

[0151] According to the procedures of FIGS. 5 to 7, tagged data associated with the dehydration process can be stored locally in the washing machine. The tagged data is classified as learning-required data for the dehydration control model, and needs to be transmitted or reported to a server device for updating the dehydration control model.

[0152] Below, we will describe the process of reporting stored tagged data to the server device side.

[0153]

[0154] Figure 8 illustrates a flowchart for a data transmission method for dehydration control according to the present invention.

[0155] The washing machine can periodically or aperiodically check whether the tagged data exceeds a threshold amount (S4210).

[0156] When the tagged data exceeds a threshold amount, the washing machine may be configured to transmit the tagged data to a server device and / or to delete the tagged data locally in the washing machine (S4220).

[0157] Here, the tagged data transmitted corresponds to the data required for learning.

[0158] The server device can be configured to retrain the learning-required data, update the dehydration control model accordingly, and distribute the updated dehydration control model to the washing machine(s).

[0159]

[0160] Figure 9 illustrates a flowchart for the mutual operation between a washing machine or a dehydration control device and a server according to the present invention.

[0161] The washing machine (1) or the dehydration control device (1000) can check whether the tagged data exceeds a critical amount due to a failed dehydration attempt during the dehydration control process (S91).

[0162] The washing machine (1) or the dehydration control device (1000) can transmit tagged data, i.e., data requiring learning, to the server device (2000) (S92)

[0163] The server device (2000) can perform retraining using the received learning-required data (S93). The server device (2000) can update the dehydration control model by performing fine-tuning on the existing model (baseline) using the received learning-required data.

[0164] Meanwhile, retraining using the received learning-required data of the server device (2000) may not be performed every time learning-required data is received. For example, the server device (2000) may be configured to perform retraining when the received learning-required data exceeds a preset amount.

[0165] The server device (2000) can obtain an updated dehydration control model through relearning (S94).

[0166] Then, the server device (2000) can transmit or distribute the updated dehydration control model to the washing machine (1) or the dehydration control device (1000) (S95).

[0167] The washing machine (1) or the dehydration control device (1000) can locally store the received updated dehydration control model (S96). The washing machine (1) or the dehydration control device (1000) can be configured to perform dehydration control using the updated dehydration control model.

[0168] Meanwhile, although what is illustrated in FIG. 9 is one washing machine (1) or dehydration control device (1000), the server device (2000) may be configured to receive learning-required data from a plurality of washing machines (1) or dehydration control devices (1000).

[0169]

[0170] Figure 10 is a graph showing dehydration performance according to an updated dehydration control model obtained by relearning data collected for dehydration control according to the present invention.

[0171] The y-axis of the graph in Figure 10 represents the dehydration success rate, and the x-axis represents laundry type information. As the dehydration control model is updated compared to the existing model (baseline), the dehydration success rate improves for most laundry.

[0172]

[0173] FIG. 11 illustrates a block diagram of a device for data collection and tagging or transmission for dehydration control according to the present invention.

[0174] The dehydration control device (1000) can be configured to collect learning-required data for dehydration control and report it to the server device (2000).

[0175] The dehydration control device (1000) may be configured to include a memory (1010) configured to collect learning-required data while performing dehydration control and to store code for reporting the collected learning-required data to the server.

[0176] The dehydration control device (1000) may be configured to include a processor (1020) configured to execute code stored in a memory (1010) for performing operations to collect the above learning-needed data and report the collected learning-needed data.

[0177] The processor (1020) may be configured to collect dehydration control data according to the performance of dehydration control following the initiation of a dehydration process. The dehydration control data may include status data including measurement sensor values ​​related to the dehydration control and operation data including probability values ​​for rotation increase and decrease commands for the dehydration control.

[0178] Additionally, the processor (1020) may be configured to select learning-required data from the collected dehydration control data. The processor (1020) may include transmitting the learning-required data to the server. The learning-required data may be used to update a dehydration control model for the dehydration control.

[0179] The dehydration control device (1000) may be configured to include a transmitter / receiver (1030) for transmitting learning-required data or receiving an updated dehydration control model.

[0180] The above learning-required data may include, among the collected dehydration control data, dehydration control data collected for a dehydration process in which a vibration measurement value exceeds a reference value, or, among the collected dehydration control data, dehydration control data in which a difference between a probability value for a first command for the dehydration control and a probability value for a second command for the dehydration control is within a threshold value. Here, the probability value may be obtained through a dehydration control model for the dehydration control.

[0181] The processor (1020) may be configured to perform tagging on dehydration control data collected for a dehydration process in which a vibration measurement value exceeds a reference value, or to perform tagging on specific dehydration control data among the dehydration control data collected for a dehydration process in which the rotation of the drum does not reach a rotation target value within a preset time. The learning-required data may include the tagged dehydration control data.

[0182] For a dehydration process in which the vibration measurement does not exceed the reference value and the rotation of the drum reaches the rotation target value within a preset time, data requiring learning may not be selected.

[0183] The above specific dehydration control data may include dehydration control data in which the difference between the probability value for the first command for the dehydration control and the probability value for the second command for the dehydration control is within a threshold value in the collected dehydration control data.

[0184] The processor (1020) may be configured to transmit the learning-needed data to the server device when the learning-needed data reaches a preset amount.

[0185] The processor (1020) may be configured to delete the collected dehydration control data or the learning-needed data after the learning-needed data is transmitted to the server device.

[0186] In the above, the dehydration control device (1000) of the present invention has been described with reference to FIG. 11. Even if not described with reference to FIG. 11, the dehydration control device (1000) of the present invention may perform the operation of the present invention according to FIGS. 1 to 10 described above.

[0187]

[0188] Figure 12 illustrates a block diagram of a device for updating a dehydration control model for dehydration control according to the present invention.

[0189] The server device (2000) can be configured to relearn a dehydration control model using learning-required data for dehydration control and transmit it to the dehydration control device (1000).

[0190] The server device (2000) may be configured to include a memory (2010) configured to store a code for relearning a dehydration control model using learning-required data received from the dehydration control device (1000) and transmitting the updated dehydration control model through relearning to the dehydration control device (1000).

[0191] The server device (2000) may be configured to include a processor (2020) configured to retrain a dehydration control model using learning-required data and to execute code stored in a memory (2010) for transmitting an updated dehydration control model through retraining.

[0192] Additionally, the server device (2000) may be configured to include a transceiver (2030) for receiving learning-required data or transmitting an updated dehydration control model.

[0193] In the above, the server device (2000) of the present invention has been described with reference to FIG. 12. Even if not described with reference to FIG. 11, the server device (2000) of the present invention may perform the operation according to the present invention according to FIG. 4 or FIG. 9 described above.

[0194]

[0195] In addition, as another aspect of the present invention, the operation of the proposal or invention described above may be implemented, performed or executed by a “computer” (a comprehensive concept including a system on chip (SoC) or a (micro) processor, etc.), or may be provided as a code or a computer-readable storage medium storing or including the code or a computer program product, and the scope of the present invention may be extended to the code or the computer-readable storage medium storing or including the code or the computer program product.

[0196]

[0197] Additionally, the present invention can be implemented by configuring two of the devices (1 or 1000, 2000) described above as one system.

[0198] For example, the dehydration control device (1000) and the server device (2000) may operate as a single system. In this case, the dehydration control device (1000) may be referred to as a client device.

[0199]

[0200] The detailed description of the preferred embodiments of the present invention disclosed above has been provided to enable those skilled in the art to implement and practice the present invention. While the above description has been made with reference to preferred embodiments of the present invention, those skilled in the art will appreciate that various modifications and variations of the present invention, as defined by the following claims, are possible. Accordingly, the present invention is not intended to be limited to the embodiments disclosed herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A device configured to collect learning-required data for dehydration control and report it to a server. A memory configured to store code for collecting learning-required data while performing dehydration control and reporting the collected learning-required data to the server; and A processor configured to execute the code for collecting the above learning-needed data and performing an operation for reporting the collected learning-needed data, The above actions are: Collect dehydration control data according to the performance of dehydration control following the start of the dehydration process, Select the data that needs learning from the above collected dehydration control data, and A device comprising transmitting the above learning-required data to the server.

2. In the first paragraph, the learning necessary data is In the above collected dehydration control data, dehydration control data collected for the dehydration process in which the vibration measurement value exceeds the reference value, or A device comprising dehydration control data in which the difference between the probability value for the first command for dehydration control and the probability value for the second command for dehydration control is within a threshold value, from the collected dehydration control data.

3. In the first paragraph, the dehydration control data is A device comprising status data including measurement sensor values ​​related to the dehydration control and operation data including probability values ​​for rotation increase and decrease commands for the dehydration control.

4. In either paragraph 2 or paragraph 3, The above probability value is obtained through a dehydration control model for the above dehydration control.

5. In the first paragraph, the operation is Tagging the collected dehydration control data for dehydration operations where the vibration measurement exceeds the reference value, or A device comprising tagging specific dehydration control data among dehydration control data collected for a dehydration process in which the rotation of the drum does not reach a rotation target value within a preset time.

6. In paragraph 5, A device wherein the above learning required data includes the tagged dehydration control data.

7. In paragraph 6, the specific dehydration control data is A device comprising dehydration control data in which the difference between the probability value for the first command for dehydration control and the probability value for the second command for dehydration control is within a threshold value, from the collected dehydration control data.

8. In the first paragraph, the operation is A device comprising transmitting the learning need data to the server when the learning need data reaches a preset amount.

9. In paragraph 1, A device in which learning-required data is not selected for a dehydration process in which the vibration measurement does not exceed the reference value and the rotation of the drum reaches the rotation target value within a preset time.

10. In the first paragraph, the operation is A device comprising, after the above learning-required data is transmitted to the server, deleting the collected dehydration control data or the learning-required data.

11. In the first paragraph, the learning-required data is used to update the dehydration control model for the dehydration control, the device.

12. A method configured to collect learning-required data for dehydration control and report it to a server, A step of collecting dehydration control data according to the performance of dehydration control following the initiation of the dehydration process; A step of selecting data requiring learning from the collected dehydration control data; and A method comprising the step of transmitting the above learning-required data to the server.

13. In paragraph 12, the learning-required data is In the above collected dehydration control data, dehydration control data collected for the dehydration process in which the vibration measurement value exceeds the reference value, or A method comprising, in the collected dehydration control data, dehydration control data in which the difference between the probability value for the first command for the dehydration control and the probability value for the second command for the dehydration control is within a threshold value.

14. In paragraph 13, The above probability value is obtained through a dehydration control model for the above dehydration control.

15. In paragraph 12, Tagging is performed on the dehydration control data collected for the dehydration process in which the vibration measurement exceeds the reference value, or A method comprising the step of performing tagging on specific dehydration control data among the dehydration control data collected for a dehydration process in which the rotation of the drum does not reach the rotation target value within a preset time.

16. In paragraph 15, A method wherein the above learning required data includes the tagged dehydration control data.

17. In paragraph 15, the specific dehydration control data is A method comprising, in the collected dehydration control data, dehydration control data in which the difference between the probability value for the first command for the dehydration control and the probability value for the second command for the dehydration control is within a threshold value.

18. In the 12th paragraph, the dehydration control data is A method comprising state data including measurement sensor values ​​related to the dehydration control and operation data including probability values ​​for rotation increase and decrease commands for the dehydration control.

19. A computer-readable medium storing code configured to execute a method according to any one of claims 12 to 18 by a computer or processor.

20. A system configured to collect and report learning-required data for dehydration control, and to relearn the learning-required data to update the dehydration control model for the dehydration control. A client device configured to collect and report learning needs data; and Including a server device configured to relearn the above learning-required data and update the dehydration control model, The client device has a memory configured to store code for collecting learning-required data while performing dehydration control and reporting the collected learning-required data to the server; and A system comprising a processor configured to execute the code for collecting the learning-needed data and performing an operation for reporting the collected learning-needed data.

Images