A control method and system for a high-precision brushless DC motor

Abstract

The present invention provides a control method and system for a high-precision brushless DC motor, which relates to the field of motor control technology. The method includes determining information about multiple pieces of laundry based on an internal video of a washing machine's trial run speed; clustering the information about the multiple pieces of laundry based on a K value of a clustering algorithm to obtain K clusters; processing the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; controlling a high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtaining an internal video of the washing machine for a first washing time; determining subsequent working time and power of the high-precision brushless DC motor based on the internal video of the washing machine for the first washing time; and controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor. The method can efficiently and accurately determine a target washing strategy for the washing machine.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and in particular to a control method and system for a high-precision brushless DC motor. Background Art

[0002] Washing machines have become an essential appliance for daily laundry. Currently, traditional washing machines mostly use fixed wash modes, such as standard, gentle, and heavy wash, with pre-set wash strategies. In practice, users face a wide variety of laundry simultaneously, ranging from delicate silk products to easily deformed wool items to relatively tough synthetic fiber products. These items vary greatly in material properties and soiling levels. For example, when faced with a wide range of laundry, selecting the gentle mode may not achieve the desired effect on heavily soiled items. Using the heavy mode can damage delicate items, potentially leading to fiber breakage and surface wear. This fixed mode design forces users to rely on subjective judgment to distinguish between items and select a wash mode, which is time-consuming and labor-intensive, and makes it difficult to achieve optimal washing results.

[0003] Therefore, how to efficiently and accurately determine the target washing strategy of the washing machine is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem solved by the present invention is how to efficiently and accurately determine the target washing strategy of a washing machine.

[0005] According to a first aspect, the present invention provides a control method for a high-precision brushless DC motor, comprising: obtaining an internal video of a washing machine's trial run speed; determining information about multiple pieces of laundry based on the internal video of the washing machine's trial run speed; determining a K value of a clustering algorithm based on the information about the multiple pieces of laundry; clustering the information about the multiple pieces of laundry based on the K value of the clustering algorithm to obtain K clusters; processing the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; controlling a high-precision brushless DC motor to wash based on the washing time and power of the cluster with the shortest washing time, and obtaining an internal video of the washing machine for a first washing time; notifying a user to take out the laundry in the cluster with the shortest washing time after the first washing time is reached, and determining a subsequent working time and power of the high-precision brushless DC motor based on the internal video of the washing machine for the first washing time; and controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor.

[0006] In a possible implementation, determining the subsequent working time and power of the high-precision brushless DC motor based on the video inside the washing machine during the first washing time includes:

[0007] determining remaining laundry information based on the video inside the washing machine during the first washing time;

[0008] generating a plurality of washing strategies and a virtual generated video of each washing strategy based on the remaining laundry information and the video inside the washing machine during the first washing time, wherein each washing strategy includes a subsequent working time and power of a high-precision brushless DC motor;

[0009] Constructing a knowledge graph, the knowledge graph includes multiple washing strategy nodes and multiple edges between the multiple washing strategy nodes. The node feature of each washing strategy node includes a virtual generated video of each washing strategy. The edges between the nodes represent the power difference and time difference between the washing strategies.

[0010] The knowledge graph is processed based on a graph neural network to determine a target washing strategy.

[0011] In a possible implementation, the speed analysis model is a long-short term neural network model.

[0012] In a possible implementation, the clustering algorithm is a K-means clustering algorithm.

[0013] According to a second aspect, the present invention provides a control system for a high-precision brushless DC motor, comprising:

[0014] An acquisition module for acquiring an internal video of the washing machine's test run speed;

[0015] an information determination module for determining information of a plurality of laundry items based on an internal video of the trial operation speed of the washing machine;

[0016] A K value determination module, configured to determine a K value of a clustering algorithm based on the information of the plurality of laundry items;

[0017] A clustering module, configured to cluster the information of the plurality of laundry items based on a K value of the clustering algorithm to obtain K clusters;

[0018] An analysis module, configured to process the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time;

[0019] a first control module, configured to control the high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtain a video of the interior of the washing machine during the first washing time;

[0020] a strategy determination module configured to notify a user to remove laundry from a cluster with the shortest washing time when a first washing time is reached, and to determine a subsequent operating time and power of the high-precision brushless DC motor based on a video of the interior of the washing machine during the first washing time;

[0021] The second control module is configured to control the high-precision brushless DC motor based on subsequent working time and power of the high-precision brushless DC motor.

[0022] In a possible implementation, the policy determination module is further configured to:

[0023] determining remaining laundry information based on the video inside the washing machine during the first washing time;

[0024] generating a plurality of washing strategies and a virtual generated video of each washing strategy based on the remaining laundry information and the video inside the washing machine during the first washing time, wherein each washing strategy includes a subsequent working time and power of a high-precision brushless DC motor;

[0025] Constructing a knowledge graph, the knowledge graph includes multiple washing strategy nodes and multiple edges between the multiple washing strategy nodes. The node feature of each washing strategy node includes a virtual generated video of each washing strategy. The edges between the nodes represent the power difference and time difference between the washing strategies.

[0026] The knowledge graph is processed based on a graph neural network to determine a target washing strategy.

[0027] In a possible implementation, the speed analysis model is a long-short term neural network model.

[0028] In a possible implementation, the clustering algorithm is a K-means clustering algorithm.

[0029] According to a third aspect, an embodiment of the present invention provides an electronic device, comprising: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method as described above, the method comprising: obtaining an internal video of a washing machine at a trial run speed; determining information about multiple pieces of laundry based on the internal video of the washing machine at a trial run speed; determining a K value of a clustering algorithm based on the information about multiple pieces of laundry; clustering the information about multiple pieces of laundry based on the K value of the clustering algorithm to obtain K clusters; processing the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; controlling a high-precision brushless DC motor to wash based on the washing time and power of the cluster with the shortest washing time, and obtaining an internal video of the washing machine for a first washing time; notifying a user to take out the laundry in the cluster with the shortest washing time after the first washing time is reached, and determining a subsequent working time and power of the high-precision brushless DC motor based on the internal video of the washing machine for the first washing time; and controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor.

[0030] According to a fourth aspect, the present embodiment provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned method for controlling a high-precision brushless DC motor, the method comprising: obtaining an internal video of a washing machine's trial run speed; determining information about a plurality of laundry items based on the internal video of the washing machine's trial run speed; determining a K value of a clustering algorithm based on the information about the plurality of laundry items; clustering the information about the plurality of laundry items based on the K value of the clustering algorithm to obtain K clusters; processing the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; controlling the high-precision brushless DC motor to wash based on the washing time and power of the cluster with the shortest washing time, and obtaining an internal video of the washing machine for a first washing time; notifying a user to take out the laundry in the cluster with the shortest washing time after the first washing time is reached, and determining a subsequent working time and power of the high-precision brushless DC motor based on the internal video of the washing machine for the first washing time; and controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor.

[0031] The present invention provides a control method and system for a high-precision brushless DC motor, the method comprising: obtaining an internal video of a washing machine's trial operation speed; determining information about multiple pieces of laundry based on the internal video of the washing machine's trial operation speed; determining a K value of a clustering algorithm based on the information about the multiple pieces of laundry; clustering the information about the multiple pieces of laundry based on the K value of the clustering algorithm to obtain K clusters; processing the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; controlling a high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtaining an internal video of the washing machine for a first washing time; notifying a user to remove the laundry in the cluster with the shortest washing time after the first washing time is reached, and determining subsequent working time and power of the high-precision brushless DC motor based on the internal video of the washing machine for the first washing time; and controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor. This method can efficiently and accurately determine a target washing strategy for the washing machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic flow chart of a method for controlling a high-precision brushless DC motor provided by an embodiment of the present invention;

[0033] Figure 2 A schematic diagram of a flow chart for determining the subsequent operating time and power of a high-precision brushless DC motor provided by an embodiment of the present invention;

[0034] Figure 3 A schematic diagram of a control system for a high-precision brushless DC motor provided by an embodiment of the present invention; DETAILED DESCRIPTION

[0035] The present invention will be further described in detail below by means of specific embodiments in conjunction with the accompanying drawings. Similar elements in different embodiments are numbered with associated similar elements. In the following embodiments, many detailed descriptions are provided to enable the present invention to be better understood. However, those skilled in the art will readily appreciate that some of the features may be omitted under different circumstances, or may be replaced by other elements, materials, or methods. In some cases, some operations related to the present invention are not shown or described in the specification. This is to avoid the core of the present invention being overwhelmed by excessive descriptions, and for those skilled in the art, it is not necessary to describe these related operations in detail. They can fully understand the related operations based on the description in the specification and the general technical knowledge in the art.

[0036] In an embodiment of the present invention, there is provided Figure 1 A control method for a high-precision brushless DC motor is shown, and the control method for the high-precision brushless DC motor includes steps S1 to S8:

[0037] Step S1, obtaining an internal video of the washing machine's test run speed.

[0038] The trial run speed is the speed at which the washing machine runs in order to detect the condition of the clothes in the washing machine before officially starting washing. The trial run speed of the washing machine is lower than the maximum speed during normal washing.

[0039] Internal video is captured by a high-speed camera installed at a specific location inside the washing machine. The high-speed camera captures details of extremely fast-changing images at a high frame rate, recording the real-time conditions inside the washing machine at test speed in a continuous sequence of images.

[0040] Step S2: determining information of a plurality of laundry items based on the internal video of the washing machine during a trial operation speed.

[0041] In some embodiments, information about multiple pieces of laundry can be determined based on an internal video of the washing machine's test run speed using a Transformer model, where the input of the Transformer model is the internal video of the washing machine's test run speed, and the output of the Transformer model is information about multiple pieces of laundry.

[0042] The Transformer model is a deep learning model based on the attention mechanism. Through its multi-head attention mechanism, the Transformer model can capture the dependencies between elements at different positions in a sequence and has powerful feature extraction and sequence modeling capabilities.

[0043] The multi-laundry information is a comprehensive collection of features related to the various laundry items in the washing machine, output by the Transformer model. This information includes laundry type, cleanliness, stain type, stain area, stain depth, value, and weight.

[0044] The types of laundry include but are not limited to laundry made of cotton, silk, chemical fiber and other materials.

[0045] Cleanliness is assessed based on the type of stain, area of ​​stain, and depth of contamination.

[0046] The degree of value is measured according to the rarity of the material, brand value, and the value of special craftsmanship such as hand embroidery and special printing and dyeing.

[0047] The stain type includes information such as oil stains and pigment stains.

[0048] The internal video of a washing machine running at a test speed fully captures the appearance, stains, and movement of the laundry. The Transformer model can infer the material from the texture and grain displayed in the video, and determine the cleanliness of the laundry by identifying features such as the location and color of stains. Furthermore, laundry of varying weights and materials exhibits distinct movement within the washing machine, providing clues for the Transformer model to determine the weight and quantity of the laundry. The Transformer model utilizes a multi-head attention mechanism to simultaneously focus on different regions and moments in the video of the washing machine running at a test speed, analyzing changes and correlations between consecutive video frames without missing key details. This allows the model to accurately transform complex video information into information about multiple laundry items, enabling precise judgment of characteristics such as laundry type, quantity, and cleanliness.

[0049] Step S3: determining a K value of a clustering algorithm based on the information of the plurality of laundry items.

[0050] In some embodiments, the K value of the clustering algorithm can be determined based on the information of the multiple pieces of laundry using an information processing model, where the information processing model is a deep neural network model, the input of the information processing model is the information of the multiple pieces of laundry, and the output of the information processing model is the K value of the clustering algorithm.

[0051] Deep neural network models include deep neural networks (DNNs), which are powerful machine learning models. DNNs consist of an input layer, multiple hidden layers, and an output layer. Their nonlinear mapping capabilities enable them to handle complex data and tasks. They also continuously adjust weights through the backpropagation algorithm, optimizing the model's inference and computational performance.

[0052] The clustering algorithm is the K-means clustering algorithm, which is an unsupervised learning algorithm that can divide samples in a dataset into K different clusters. The K-means clustering algorithm calculates the similarity between samples and groups samples with high similarity into the same cluster.

[0053] The K value is a key parameter in the K-means clustering algorithm, output by the information processing model. It represents the number of clusters. An appropriate K value ensures high similarity among samples within each cluster, while maintaining significant differences between samples in different clusters. The complexity of multi-piece laundry information makes determining an appropriate K value particularly crucial. If the K value is too small, different types of data will be incorrectly clustered together, resulting in a loss of data diversity. If the K value is too large, the amount of data in each cluster will be too small, resulting in overly fragmented clustering and difficulty in identifying underlying patterns in the data.

[0054] Information about multiple laundry items encompasses multiple characteristics, including material, quantity, weight, and value. These characteristics are interrelated. For example, silk laundry is often more valuable, requires more stringent washing methods, and cannot be washed for extended periods of time. These interconnections form a complex pattern, providing a basis for determining the K value. By comprehensively analyzing these interconnections, the deep neural network model can group laundry items with similar characteristics and washing requirements into the same category.

[0055] Deep neural networks leverage their powerful nonlinear fitting and feature learning capabilities. By connecting neurons and adjusting weights, they deeply mine the complex features within multiple laundry items, establishing a correlation between feature combinations and the number of clusters. Deep neural networks continuously optimize their parameters to better fit patterns in the data. Given multiple laundry items, deep neural networks can perform in-depth analysis and uncover potential classification patterns, ultimately outputting an appropriate K value for the clustering algorithm to achieve appropriate classification of the laundry items.

[0056] Step S4: clustering the information of the plurality of laundry items based on the K value of the clustering algorithm to obtain K clusters.

[0057] In some embodiments, K clusters and the cluster centers of each of the K clusters can be obtained by the following steps: randomly selecting K laundry information data points as initial cluster centers; then measuring the similarity between each laundry information data point and the K cluster centers using Euclidean distance, and assigning each data point to the cluster corresponding to the cluster center with the highest similarity; then recalculating the mean of all laundry information data points in each cluster to update the positions of the K cluster centers; repeating the above process of calculating similarity, assigning data points, and updating cluster centers until the cluster centers no longer change significantly or a pre-set upper limit on the number of iterations is reached, ultimately obtaining K clusters.

[0058] Each cluster is a collection of data samples with similar characteristics. In some embodiments, a cluster is a category group formed by dividing the information of multiple laundry items according to the K value. The laundry information within the same cluster has high similarity in characteristics such as material and value, and the characteristics of different clusters vary significantly.

[0059] Step S5: Process the K clusters based on the speed analysis model to determine the washing time and power of the cluster with the shortest washing time.

[0060] The speed analysis model is a long-short term neural network model, the input of the speed analysis model is K clusters, and the output of the speed analysis model is the washing time and power of a cluster with the shortest washing time.

[0061] Long-term and short-term neural network models include long short-term memory networks (LSTMs). LSTMs are a special type of recurrent neural network (RNN) composed of an input gate, a forget gate, and an output gate. The input gate controls the input of new information, the forget gate determines whether to retain or discard previously remembered information, and the output gate determines the output based on the current state and input. These gating mechanisms enable LSTMs to selectively remember or forget past information and make decisions based on current input and historical memory.

[0062] The cluster with the shortest wash time is the collection of categories with the shortest wash times. The laundry within this cluster possesses the highest value attributes compared to laundry in other clusters. Because the fibers and chemical properties of laundry in this cluster are more sensitive and fragile, prolonged washing can easily lead to irreversible damage from mechanical friction, chemical erosion, and temperature fluctuations, such as luster loss in silk and fiber breakage in cashmere. Therefore, to protect the physical structure and appearance of this type of laundry, it is assigned a shorter wash time, thus being designated as the cluster with the shortest wash time.

[0063] Each cluster contains a variety of characteristic information about the laundry. The long-short-term neural network, through a gating mechanism, selectively retains or forgets this information and captures the dynamic changes of these features over time. The differences in characteristics between clusters, such as material, weight, and soiling, provide the model with rich comparative information. By learning these differential features, the model can more accurately identify the unique characteristic patterns of the cluster with the shortest wash time. Based on this, by combining the learned patterns from historical data, it can accurately determine the appropriate wash time and power for this cluster, ensuring optimal washing results while protecting these high-value, fragile laundry.

[0064] In some embodiments, the speed analysis model includes an information extraction layer, a washing parameter estimation layer, and a comparative decision layer. Each of these layers comprises a long-term and short-term neural network structure. The information extraction layer takes as input K clusters, and outputs the comprehensive feature information for each cluster and the difference between the cluster with the shortest wash time and each other cluster. The washing parameter estimation layer takes as input the comprehensive feature information for each cluster, and outputs the estimated wash time range and estimated wash power range for each cluster. The comparative decision layer takes as input the difference between the cluster with the shortest wash time and each other cluster, the estimated wash time range and estimated wash power range for each cluster, and outputs the wash time and power for the cluster with the shortest wash time.

[0065] The comprehensive characteristic information of each cluster is a quantitative description of the overall characteristics of all the laundry in the same cluster. The characteristic information integrates key information such as the material type, overall dirtiness, total weight, fiber structure and common chemical properties of the laundry in the cluster.

[0066] The difference information between the cluster with the shortest wash time and each other cluster is the degree of difference in material information, soiling level, weight, fiber structure, and chemical properties between the cluster with the shortest wash time and each other. Material differences can be expressed as a similarity percentage, soiling level differences as a multiple, weight differences as a specific difference, and fiber structure and chemical property differences as a comprehensive score.

[0067] Different layers perform distinct functions. The information extraction layer focuses on integrating the common features of each of the K clusters and identifying key feature differences between the cluster with the shortest wash time and the others. The washing parameter estimation layer uses this information to make preliminary estimates of each cluster's wash parameters, calculating the appropriate wash power and wash time range for each cluster. The comparative decision layer combines this difference information with the estimated parameters to accurately determine the wash time and power for the target cluster. This layered processing breaks down the complex wash strategy decision-making task to improve the targeted nature of information processing, thereby enhancing model processing efficiency and decision accuracy.

[0068] Step S6: controlling the high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtaining a video inside the washing machine during the first washing time.

[0069] The video inside the washing machine during the first washing time is a video inside the washing machine captured by a high-speed camera at a specific position inside the washing machine at a high frame rate during a time period when a cluster of laundry is washed for the shortest time.

[0070] Step S7: When the first washing time is reached, the user is notified to take the laundry in the cluster with the shortest washing time, and the subsequent working time and power of the high-precision brushless DC motor are determined based on the video inside the washing machine during the first washing time.

[0071] In some embodiments, Figure 2 A schematic diagram of a flow chart for determining the subsequent operating time and power of a high-precision brushless DC motor provided by an embodiment of the present invention, wherein determining the subsequent operating time and power of the high-precision brushless DC motor includes steps S21 to S24:

[0072] Step S21 : determining information about remaining laundry based on the video inside the washing machine during the first washing time.

[0073] In some embodiments, the remaining laundry information is determined using a laundry determination model based on the video inside the washing machine at the first washing time. The laundry determination model is a Transformer model. The input of the laundry determination model is the video inside the washing machine at the first washing time, and the output of the laundry determination model is the remaining laundry information.

[0074] Remaining laundry information is information about the laundry remaining in the washing machine after the first wash cycle. This information includes laundry type, cleanliness level, type and location of remaining stains, and differences in remaining cleanliness between laundry materials. This remaining laundry information can be used to more accurately control the washing machine's subsequent operating time and power.

[0075] The video of the washing machine's interior during the first wash cycle captures the dynamic changes in the laundry's state during the wash cycle, including details such as residual stains. The Transformer model, leveraging its multi-head attention mechanism, can capture key information from different frames and regions of the video. Combining its feature extraction and sequence modeling capabilities, the model can accurately identify remaining stain features, changes in the laundry's state, and other information from the video of the washing machine's interior during the first wash cycle, thereby determining the remaining laundry.

[0076] Step S22: generating a plurality of washing strategies and a virtual generated video of each washing strategy based on the remaining laundry information and the internal video of the washing machine during the first washing time. Each washing strategy includes the subsequent working time and power of the high-precision brushless DC motor.

[0077] In some embodiments, a policy generation model can be used to generate multiple washing strategies and a virtual generated video for each washing strategy based on the remaining laundry information and the video inside the washing machine at the first washing time. The policy generation model is a generative adversarial network. The input of the policy generation model is the remaining laundry information and the video inside the washing machine at the first washing time. The output of the policy generation model is multiple washing strategies and a virtual generated video for each washing strategy.

[0078] A generative adversarial network (GAN) is a deep learning model composed of a generator and a discriminator. The generator generates new content based on input data, while the discriminator determines the authenticity of the content. Through adversarial training, the two are continuously optimized to improve the quality of generation.

[0079] Multiple washing strategies are a series of washing plans formulated for the remaining laundry in the washing machine. Specifically, they include the setting of parameters such as the subsequent working time and power of the high-precision brushless DC motor to achieve the best washing effect with high efficiency, energy saving and protection of the laundry.

[0080] Each washing strategy's virtual generated video is generated using a generative adversarial network to simulate the washing process and results inside the washing machine under each washing strategy. This virtual generated video allows for a more intuitive analysis of the potential washing results of each strategy, enabling the selection of the optimal washing strategy.

[0081] The remaining load information and the first wash cycle video of the washing machine contain rich details. The remaining load information details the current state of the laundry, including the type and location of any remaining stains, and the differences in cleanliness between different materials. These details are the core basis for developing washing strategies. The first wash cycle video of the washing machine provides a visual representation of the changes in the laundry's state during the wash cycle, complementing the remaining load information.

[0082] Generative adversarial networks (GANs) possess powerful feature learning and data generation capabilities. Their generator learns underlying patterns and regularities in the input data and extracts key features from the remaining laundry information and videos, transforming them into a variety of possible washing strategies and corresponding virtual videos. The discriminator, based on actual washing experience and logic, evaluates the generator's output to determine its rationality and effectiveness. Through continuous adversarial training, the generator gradually optimizes its generated washing strategies and virtual videos, ensuring they not only accurately match the actual remaining laundry but also meet the washing requirements of different materials and levels of soiling.

[0083] Step S23, constructing a knowledge graph, the knowledge graph includes multiple washing strategy nodes and multiple edges between the multiple washing strategy nodes, the node feature of each washing strategy node includes a virtual generated video of each washing strategy, and the edges between the nodes represent the power difference and time difference between the washing strategies.

[0084] A knowledge graph is a data structure consisting of nodes and edges. Nodes represent entities, node features are attribute information related to the nodes, and edges connect different nodes in the knowledge graph.

[0085] The knowledge graph presents complex washing strategies and their relationships in an intuitive and structured manner. Each node represents a specific washing strategy, and the node features include a virtual generated video of each washing strategy. Edges represent the power and time differences between washing strategies.

[0086] Step S24: Process the knowledge graph based on a graph neural network to determine a target washing strategy.

[0087] Graph Neural Network (GNN) is a deep learning model that runs directly on knowledge graph data. It transfers information and aggregates features between nodes, and learns the feature representations of nodes and edges in the graph, thereby mining the potential patterns and relationships of the knowledge graph.

[0088] The target washing strategy uses a graph neural network to process the knowledge graph and select the most suitable washing plan for the remaining laundry. The target washing strategy includes parameter settings for the subsequent operating time and power of the high-precision brushless DC motor. This strategy can achieve energy savings and protect the laundry while ensuring washing results.

[0089] The knowledge graph clearly presents multiple washing strategies and the relationships between them. The nodes carry the specific information of each washing strategy and are the basis for constructing the knowledge graph and graph neural network for analysis.

[0090] Node features enrich the information of washing strategy nodes and allow graph neural networks to more comprehensively understand the characteristics and expected effects of each washing strategy, thereby facilitating the evaluation and comparison of washing strategies.

[0091] Edges can reflect the degree of difference between different washing strategies. The power difference and time difference represented by the edges can help graph neural networks analyze the similarities and differences between washing strategies.

[0092] Graph neural networks excel at processing knowledge graphs. They update node feature representations by transferring information between adjacent nodes and performing weighted aggregation of node features based on factors like edge weights. This allows them to mine potential patterns and relationships within knowledge graphs. When presented with knowledge graphs, graph neural networks can fully leverage the characteristics of washing strategy nodes and the relationship information embedded in their edges to uncover the potential value of different washing strategies in meeting remaining laundry needs. By learning and analyzing knowledge graphs, graph neural networks can select the target washing strategy that best meets actual needs from multiple washing strategies, enabling them to make washing strategy decisions.

[0093] Step S8 , controlling the high-precision brushless DC motor based on the subsequent working time and power of the high-precision brushless DC motor.

[0094] When the target washing strategy is determined, the high-precision brushless DC motor is controlled based on the subsequent working time and power of the high-precision brushless DC motor in the target washing strategy to complete washing of the remaining laundry.

[0095] Based on the same inventive concept, Figure 3A schematic diagram of a control system for a high-precision brushless DC motor provided in an embodiment of the present invention, wherein the control system for the high-precision brushless DC motor includes:

[0096] An acquisition module 31 is used to acquire an internal video of the test running speed of the washing machine;

[0097] An information determination module 32 is configured to determine information about a plurality of laundry items based on an internal video of the washing machine's test run speed;

[0098] A K value determination module 33 is used to determine the K value of the clustering algorithm based on the information of the plurality of laundry items;

[0099] A clustering module 34 is configured to cluster the laundry information into K clusters based on a K value of the clustering algorithm;

[0100] An analysis module 35 is configured to process the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time;

[0101] A first control module 36 is configured to control the high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtain a video of the interior of the washing machine during the first washing time;

[0102] a strategy determination module 37 for notifying the user to remove the laundry in the cluster with the shortest washing time when the first washing time is reached, and determining the subsequent operating time and power of the high-precision brushless DC motor based on the video of the interior of the washing machine during the first washing time;

[0103] The second control module 38 is configured to control the high-precision brushless DC motor based on subsequent working time and power of the high-precision brushless DC motor.

[0104] In addition, unless expressly stated in the claims, the order of the processing elements and sequences, the use of alphanumeric characters, or the use of other names described in this specification are not intended to limit the order of the processes and methods of this specification. Although the above disclosure discusses some of the invention embodiments currently considered useful through various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the spirit and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only by software solutions, such as installing the described system on an existing server or mobile device.

[0105] Similarly, it should be noted that, in order to simplify the presentation of this specification and thus facilitate understanding of one or more embodiments of the invention, the foregoing descriptions of the embodiments of this specification sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this disclosure method does not imply that the subject matter of this specification requires more features than those recited in the claims. In fact, an embodiment may have fewer features than all of the features of a single disclosed embodiment.

[0106] Finally, it should be understood that the embodiments described in this specification are intended only to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly described and illustrated in this specification.

Claims

1. A control method for a high-precision brushless DC motor, characterized in that: include: Get an inside video of your washing machine running at speed; Determining information about a plurality of laundry items based on the internal video of the washing machine's test run speed, wherein determining information about a plurality of laundry items based on the internal video of the washing machine's test run speed comprises: determining information about a plurality of laundry items based on the internal video of the washing machine's test run speed using a Transformer model; determining a K value of a clustering algorithm based on the information of the plurality of laundry items; Clustering the plurality of laundry information based on the K value of the clustering algorithm to obtain K clusters; Based on the speed analysis model, the K clusters are processed to determine the washing time and power of the cluster with the shortest washing time; controlling a high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtaining a video inside the washing machine during the first washing time; When the first washing time is reached, the user is notified to take out the laundry in the cluster with the shortest washing time, and the subsequent operating time and power of the high-precision brushless DC motor are determined based on the video inside the washing machine during the first washing time. The determining of the subsequent operating time and power of the high-precision brushless DC motor based on the video inside the washing machine during the first washing time includes: determining remaining laundry information based on the video inside the washing machine during the first washing time; generating a plurality of washing strategies and a virtual generated video of each washing strategy based on the remaining laundry information and the video inside the washing machine during the first washing time, wherein each washing strategy includes a subsequent working time and power of a high-precision brushless DC motor; Constructing a knowledge graph, the knowledge graph includes multiple washing strategy nodes and multiple edges between the multiple washing strategy nodes. The node feature of each washing strategy node includes a virtual generated video of each washing strategy. The edges between the nodes represent the power difference and time difference between the washing strategies. Processing the knowledge graph based on a graph neural network to determine a target washing strategy; The high-precision brushless DC motor is controlled based on the subsequent working time and power of the high-precision brushless DC motor.

2. The control method of the high-precision brushless DC motor according to claim 1, wherein: The speed analysis model is a long-term and short-term neural network model.

3. The control method of the high-precision brushless DC motor according to claim 1, wherein: The clustering algorithm is the K-means clustering algorithm.

4. A high-precision brushless DC motor control system, characterized in that: include: An acquisition module for acquiring an internal video of the washing machine's test run speed; an information determination module, configured to determine information about a plurality of laundry items based on an internal video of the washing machine's test run speed, wherein determining information about the plurality of laundry items based on the internal video of the washing machine's test run speed comprises: determining information about the plurality of laundry items based on the internal video of the washing machine's test run speed using a Transformer model; A K value determination module, configured to determine a K value of a clustering algorithm based on the information of the plurality of laundry items; A clustering module, configured to cluster the information of the plurality of laundry items based on a K value of the clustering algorithm to obtain K clusters; An analysis module, configured to process the K clusters based on a speed analysis model to determine the washing time and power of a cluster with the shortest washing time; a first control module, configured to control the high-precision brushless DC motor to perform washing based on the washing time and power of the cluster with the shortest washing time, and obtain a video of the interior of the washing machine during the first washing time; A strategy determination module is configured to notify a user to remove laundry from the cluster with the shortest washing time when a first washing time is reached, and to determine a subsequent operating time and power of the high-precision brushless DC motor based on the video of the interior of the washing machine during the first washing time. The strategy determination module is further configured to: determining remaining laundry information based on the video inside the washing machine during the first washing time; generating a plurality of washing strategies and a virtual generated video of each washing strategy based on the remaining laundry information and the video inside the washing machine during the first washing time, wherein each washing strategy includes a subsequent working time and power of a high-precision brushless DC motor; Constructing a knowledge graph, the knowledge graph includes multiple washing strategy nodes and multiple edges between the multiple washing strategy nodes. The node feature of each washing strategy node includes a virtual generated video of each washing strategy. The edges between the nodes represent the power difference and time difference between the washing strategies. Processing the knowledge graph based on a graph neural network to determine a target washing strategy; The second control module is configured to control the high-precision brushless DC motor based on subsequent working time and power of the high-precision brushless DC motor.

5. The control system of the high-precision brushless DC motor according to claim 4, characterized in that: The speed analysis model is a long-term and short-term neural network model.

6. The control system of the high-precision brushless DC motor according to claim 4, characterized in that: The clustering algorithm is the K-means clustering algorithm.

7. An electronic device, characterized in that: include: processor; Memory; and a computer program; wherein the computer program is stored in the memory and is configured to be executed by the processor to implement the control method of the high-precision brushless DC motor according to any one of claims 1 to 3.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the control method of the high-precision brushless DC motor according to any one of claims 1 to 3 is implemented.

Images