Refrigerator control method and device, refrigerator and medium

Abstract

The embodiment of the invention provides a refrigerator control method and device, a refrigerator and a medium. The method comprises the steps that behavior information corresponding to taking behaviors of all users for articles in the refrigerator in a family environment is determined; the behavior information comprises at least one of information of a user initiating a taking behavior, article information of a taken article, occurrence time of the taking behavior and environment information when the taking behavior occurs; according to the behavior information, determining behavior semantics and behavior popularity of the taking behavior; according to the behavior semantics and behavior popularity of the taking behavior, fresh-keeping adjustment parameters of a target area related to the taking behavior in the refrigerator are determined; and according to the fresh-keeping adjustment parameters, the target area of the refrigerator is adjusted, so that the operation state can be dynamically adjusted according to the behavior semantics and the behavior popularity, the dynamic adjustment capacity of different areas of the refrigerator is optimized, and the fresh-keeping effect and the energy consumption problem between the different areas are balanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of smart home, and in particular, to a refrigerator control method and device, a refrigerator and a medium. BACKGROUND

[0002] With the rapid development of smart home, the refrigerator, as an indispensable electrical appliance in the family, its function has been expanded from simple refrigeration and freezing to the field of intelligent management. However, in the actual use process, the existing refrigerator usually adopts fixed temperature and humidity settings during operation. This single preservation mode may result in insufficient preservation effect in some areas, while energy waste may exist in other areas. SUMMARY

[0003] In view of the above problems, the present application embodiments are proposed to provide a refrigerator control method, device, equipment and medium which can overcome the problem that some areas cannot provide sufficient preservation effect, while other areas may exist energy waste, or at least partially solve the above problems.

[0004] To solve the above problems, the present application embodiments disclose a refrigerator control method, which comprises:

[0005] Determine the behavior information corresponding to the taking behavior of each user in the home environment to the items in the refrigerator; the behavior information comprises at least one of the user information initiating the taking behavior, the item information taking the items, the occurrence time of the taking behavior and the environmental information at the time of the taking behavior;

[0006] According to the behavior information, determine the behavior semantics and behavior heat of the taking behavior;

[0007] According to the behavior semantics and behavior heat of the taking behavior, determine the preservation adjustment parameter of the target area related to the taking behavior in the refrigerator;

[0008] According to the preservation adjustment parameter, adjust the preservation environment of the target area of the refrigerator.

[0009] Optionally, the behavior information comprises the item information taking the items and the occurrence time of the taking behavior, and the determination of the behavior heat of the taking behavior according to the behavior information comprises:

[0010] According to the item information of the taking items and the occurrence time of the taking behavior, determine the behavior label of the taking behavior; the behavior label is used to represent the category and occurrence time of the taking behavior;

[0011] According to the behavior label, determine the behavior heat of the taking behavior.

[0012] Optionally, the determining the behavior heat of the taking behavior according to the behavior label comprises:

[0013] obtaining occurrence time of at least one historical taking behavior corresponding to the behavior label and a behavior heat decay coefficient corresponding to the behavior label;

[0014] determining the behavior heat of the taking behavior according to the occurrence time of the taking behavior, the occurrence time of the historical taking behavior and the heat decay coefficient.

[0015] Optionally, the determining the behavior heat of the taking behavior according to the occurrence time of the taking behavior, the occurrence time of the historical taking behavior and the heat decay coefficient comprises:

[0016] determining the behavior heat of the taking behavior according to the occurrence time of the taking behavior, the occurrence time of the historical taking behavior and the heat decay coefficient according to the following formula:

[0017]

[0018] wherein H represents the behavior heat, λ represents the behavior heat decay coefficient, and Δt represents a time difference between the occurrence time of the at least one taking behavior and a current time.

[0019] Optionally, the item information of the taken item comprises a type of the taken item and a quantity of the taken item, and the determining the behavior semantics of the taking behavior according to the behavior information comprises:

[0020] determining a semantic feature vector of the taking behavior according to the shelf life of the taken item, the type of the taken item and the quantity of the taken item;

[0021] determining the behavior semantics of the taking behavior according to the semantic feature vector.

[0022] Optionally, the determining the preservation adjustment parameter of the target region related to the taking behavior in the refrigerator according to the behavior semantics and the behavior heat of the taking behavior comprises:

[0023] determining a user demand influence factor according to the behavior semantics and the behavior heat;

[0024] determining an energy consumption influence factor according to preset refrigerator power data;

[0025] determining a preservation adjustment mapping coefficient of the target region related to the taking behavior in the refrigerator according to the user demand influence factor and the energy consumption influence factor.

[0026] Optionally, the determining the preservation adjustment mapping coefficient of the target region related to the taking behavior in the refrigerator according to the shelf life influence factor, the user demand influence factor and the energy consumption influence factor comprises:

[0027] The preservation adjustment mapping coefficient of the target region related to the taking behavior in the refrigerator is determined according to the user demand influence factor and the energy consumption influence factor according to the following formula:

[0028] P=w 1* E+w 2* U, wherein E represents the energy consumption influence factor, U represents the user demand influence factor, P represents the mapping coefficient, w1 represents the energy consumption influence factor weight, and w2 represents the user demand influence factor weight;

[0029] The preservation adjustment parameter of the target region related to the taking behavior in the refrigerator is determined according to the mapping coefficient.

[0030] Optionally, the behavior information comprises the item information of the taken item, and the determining the behavior information corresponding to the taking behavior of each user in the household environment to the item in the refrigerator comprises:

[0031] The action image of the user using the refrigerator and the weight change information of the target region inside the refrigerator are acquired;

[0032] In a case where it is determined that the user performs the taking behavior according to the weight change information, the item information of the item taken by the user is determined according to the action image.

[0033] In another aspect, the present application further provides a refrigerator control device, which comprises:

[0034] A behavior information acquisition module is configured to determine behavior information corresponding to the taking behavior of each user in the household environment to the item in the refrigerator, wherein the behavior information comprises at least one of the user information of the taking behavior initiator, the item information of the taken item, the occurrence time of the taking behavior and the environmental information at the occurrence time of the taking behavior;

[0035] A semantic heat determination module is configured to determine the behavior semantic and the behavior heat of the taking behavior according to the behavior information;

[0036] A parameter determination module is configured to determine the preservation adjustment parameter of the target region related to the taking behavior in the refrigerator according to the behavior semantic and the behavior heat of the taking behavior;

[0037] An adjustment module is configured to adjust the preservation environment of the target region of the refrigerator according to the preservation adjustment parameter.

[0038] Optionally, the behavior information comprises article information of the taken article and a time of occurrence of the taking behavior, and the semantic hotness determination module comprises:

[0039] a behavior label determination sub-module configured to determine a behavior label of the taking behavior according to the article information of the taken article and the time of occurrence of the taking behavior, the behavior label being used to represent a category and the time of occurrence of the taking behavior;

[0040] a behavior hotness determination sub-module configured to determine a behavior hotness of the taking behavior according to the behavior label.

[0041] Optionally, the behavior hotness determination sub-module comprises:

[0042] a decay coefficient determination unit configured to obtain a time of occurrence of at least one historical taking behavior corresponding to the behavior label and a behavior hotness decay coefficient corresponding to the behavior label;

[0043] a behavior hotness calculation unit configured to determine the behavior hotness of the taking behavior according to the time of occurrence of the taking behavior, the time of occurrence of the historical taking behavior and the hotness decay coefficient.

[0044] Optionally, the behavior hotness calculation unit comprises:

[0045] a first formula calculation unit configured to determine the behavior hotness of the taking behavior according to the time of occurrence of the taking behavior, the time of occurrence of the historical taking behavior and the hotness decay coefficient according to the following formula:

[0046]

[0047] wherein H represents the behavior hotness, λ represents the behavior hotness decay coefficient, and Δt represents a time difference between the time of occurrence of the at least one taking behavior and a current time.

[0048] Optionally, the article information of the taken article comprises a category of the taken article and a quantity of the taken article, and the semantic hotness determination module comprises:

[0049] a semantic determination sub-module configured to determine a semantic feature vector of the taking behavior according to a shelf life of the taken article, the category of the taken article and the quantity of the taken article;

[0050] a behavior semantic determination sub-module configured to determine a behavior semantic of the taking behavior according to the semantic feature vector.

[0051] Optionally, the parameter determination module comprises:

[0052] a first impact factor determining submodule configured to determine a user demand impact factor according to the behavior semantics and the behavior heat;

[0053] a second impact factor determining submodule configured to determine an energy consumption impact factor according to preset refrigerator power data;

[0054] a mapping coefficient determining submodule configured to determine a preservation adjustment mapping coefficient of a target area related to the taking behavior in the refrigerator according to the user demand impact factor and the energy consumption impact factor.

[0055] Optionally, the mapping coefficient determining submodule comprises:

[0056] a second formula calculating unit configured to determine the preservation adjustment mapping coefficient of the target area related to the taking behavior in the refrigerator according to the user demand impact factor and the energy consumption impact factor according to the following formula:

[0057] P=w 1* E+w 2* U, wherein E represents the energy consumption impact factor, U represents the user demand impact factor, P represents the mapping coefficient, w1 represents the energy consumption impact factor weight, and w2 represents the user demand impact factor weight;

[0058] a preservation parameter determining unit configured to determine a preservation adjustment parameter of the target area related to the taking behavior in the refrigerator according to the mapping coefficient.

[0059] Optionally, the behavior information comprises article information of an article taken by the user, and the behavior information acquiring module comprises:

[0060] an image and weight information acquiring submodule configured to acquire action images of the user using the refrigerator and weight change information of a target area inside the refrigerator;

[0061] an article information determining submodule configured to, in a case where it is determined that the user performs a taking behavior according to the weight change information, determine article information of an article taken by the user according to the action images.

[0062] Correspondingly, an embodiment of the present application discloses a refrigerator, comprising a processor, a memory, and a computer program stored in the memory and capable of running on the processor, and the computer program is executed by the processor to implement each step of the refrigerator control method embodiment.

[0063] Correspondingly, an embodiment of the present application discloses a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement each step of the refrigerator control method embodiment.

[0064] The embodiment of the present application comprises the following advantages: by determining behavior information corresponding to the taking behavior of each user in the home environment on the items in the refrigerator; the behavior information comprises at least one of user information initiating the taking behavior, item information taking the item, occurrence time of the taking behavior and environmental information at the time of the taking behavior; according to the behavior information, determining behavior semantics and behavior heat of the taking behavior; according to the behavior semantics and behavior heat of the taking behavior, determining a preservation adjustment parameter of a target area in the refrigerator related to the taking behavior, the use frequency of different areas in the refrigerator can be determined according to the behavior heat, and the preservation strategy of each area of the refrigerator is further formulated more targeted and flexibly in combination with the behavior semantics; according to the preservation adjustment parameter, adjusting the target area of the refrigerator, thereby realizing the transformation of the refrigerator from a single function device to an intelligent management device, enabling it to dynamically adjust the running state according to the behavior semantics and behavior heat, optimizing the dynamic adjustment capability of different areas of the refrigerator, thereby balancing the preservation effect and energy consumption problems between different areas. BRIEF DESCRIPTION OF DRAWINGS

[0065] Figure 1 is a step flow chart of an embodiment of a refrigerator control method of the present application;

[0066] Figure 2 is a step flow chart of another embodiment of a refrigerator control method of the present application;

[0067] Figure 3 is a structure block diagram of an embodiment of a refrigerator control device of the present application. DETAILED DESCRIPTION

[0068] In order to make the above-mentioned objects, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0069] One of the core ideas of the embodiment of the present application is to analyze the heat and semantic information of the user taking behavior, determine the underlying meaning and corresponding heat of the user taking behavior, and perform real-time control of the refrigerator according to the underlying meaning and corresponding heat, so as to optimize the preservation performance of the refrigerator.

[0070] Referring to Figure 1 , a step flow chart of an embodiment of a refrigerator control method of the present application is shown, which can specifically comprise the following steps:

[0071] Step 101, determining behavior information corresponding to the taking behavior of each user in the home environment on the items in the refrigerator; the behavior information comprises at least one of user information initiating the taking behavior, item information taking the item, occurrence time of the taking behavior and environmental information at the time of the taking behavior;

[0072] The core function of a refrigerator is to provide a suitable preservation environment for stored items, but different items have different requirements for temperature, humidity, and other conditions. By analyzing user behavior with the refrigerator, the preservation strategy for each area of the refrigerator can be adjusted more accurately, thereby improving user experience and the preservation effect of items. After understanding the types of items that the user frequently accesses and their storage locations, the refrigerator can dynamically adjust the temperature and humidity in different areas based on these habits. For example, if the user often stores vegetables in a certain area, that area can automatically switch to a high-humidity mode. By recording the user's access frequency and item information, the refrigerator can predict that certain items may be forgotten or stored for a long time, thereby reminding the user to use them in time or adjusting the preservation conditions in that area. Different users may have different dietary preferences and storage habits, such as users who prefer cold drinks may need lower refrigeration temperatures. By analyzing behavior data, the refrigerator can provide personalized preservation solutions for each user.

[0073] In general, user behavior with the refrigerator mainly includes the following categories: storage behavior, representing the user's action of placing items into the refrigerator; retrieval behavior, representing the user's action of taking items out of the refrigerator; neglect behavior, representing the user's behavior of not touching certain items for a long time; adjustment behavior, representing the user's behavior of actively adjusting the refrigerator settings; and abnormal behavior, representing behavior that does not conform to regular usage habits.

[0074] For example, by installing weight sensors, temperature sensors, humidity sensors, etc. inside the refrigerator, the effect of determining user behavior can be achieved by monitoring the changes in items in real time, for example, when the weight of a certain area decreases, it can be determined that the user has taken the item.

[0075] In another example, a camera and image recognition technology can be used to detect changes in items inside the refrigerator to determine user behavior, in addition to which face recognition, fingerprint recognition, or mobile app binding can be used to determine the specific user who initiated the behavior, further determining user behavior.

[0076] For a smart refrigerator system, it generally includes the refrigerator body and its built-in hardware components such as sensors, cameras, display screens, etc. These hardware components can be responsible for collecting user behavior data and executing corresponding preservation strategy adjustments. The smart chip and operating system built into the refrigerator can be combined with a cloud server to determine the preservation strategy based on user behavior. At the same time, users can interact with the refrigerator software system through smart terminals such as smartphones or tablets.

[0077] In the embodiments of the present application, each user in the home environment refers to all members who may contact or use the refrigerator in the home scene, including but not limited to family members, visitors, housekeeping service personnel, etc. The taking behavior of the refrigerator contents refers to the action or operation of the user taking out a certain item from the refrigerator. This behavior usually occurs frequently in daily life, such as taking out food, drinks or other stored items. The behavior information refers to a specific set of data related to the taking behavior, which is used to describe and record the characteristics of this behavior. The behavior information at least includes the following information:

[0078] The user information initiating the taking behavior refers to the identity information of the specific user performing the taking action. The user can be identified in various ways, such as face recognition, fingerprint recognition, voice recognition, wearable device binding, etc. The user information helps to distinguish the habits of different family members;

[0079] The item information of the taken item refers to the detailed information of the taken item, which includes but is not limited to: item name, item category, item weight, item storage location, item shelf life, etc. The record of item information can be used to track inventory changes, predict replenishment needs, and analyze user consumption preferences.

[0080] The occurrence time of the taking behavior refers to the specific time point when the user completes the taking action. The time information can be refined to year, month, day, hour, minute, or even second to accurately record the time of the behavior. Through the analysis of time data, the user's usage patterns can be found, such as some users like to take milk in the morning and beer at night.

[0081] The environmental information at the time of the taking behavior refers to the surrounding environmental state when the user takes the item. The environmental information can include: physical environment, such as indoor temperature, humidity, light intensity, etc.; social environment, such as whether there are other family members present, whether it is meal time, etc.; refrigerator state, such as the internal temperature of the refrigerator, the number of remaining items, the door opening and closing frequency, etc.; Environmental information can help understand the motivation and influencing factors behind the behavior, such as high temperature weather may lead to an increase in the frequency of taking cold drinks.

[0082] By comprehensively collecting the above information, a detailed database about the user's taking behavior is established. These data can serve as the basis for subsequent development of intelligent refrigerator functions, such as the implementation of personalized recommendations, automatic replenishment reminders, energy optimization, etc. At the same time, this fine-grained data collection also helps to improve user experience, making the refrigerator more intelligent and humanized.

[0083] Step 102, determining the behavior semantics and behavior heat of the taking behavior according to the behavior information;

[0084] The behavior semantics refers to the meaningful patterns or rules extracted from the user's taking behavior, reflecting the user's behavior purpose, preference and habit. For example, the behavior semantics of the taking behavior can be determined by identifying the type of the taken item; for example, frequently taking beverages may indicate that the user likes cold drinks; the behavior semantics of the taking behavior can be determined in combination with the occurrence time of the taking behavior. For example, taking milk in the morning may be for breakfast, and taking beer in the evening may be for relaxation; the behavior semantics of the taking behavior can be determined by considering environmental information. For example, frequently taking cold drinks in high-temperature weather may be related to the cooling demand;

[0085] The behavior heat refers to the frequency or intensity of a certain behavior, which is used to measure the importance and priority of the behavior. For example, the behavior heat can be determined by calculating the number of times an item or a region is taken. For example, if a region is visited multiple times a day, its behavior heat is high; the behavior heat can be determined by analyzing the distribution of the behavior in a day or a week, for example, some items may be frequently taken only in a certain time period; in addition, the behavior heat calculated according to the behavior distribution in a preset event period or the number of times an item is taken can be weighted according to different dimensions of information, such as user identity, item importance, etc., to obtain the final behavior heat. For example, the behavior of a housewife taking food may have higher heat than the behavior of a visitor taking beverages.

[0086] In step 103, the preservation adjustment parameter of the target region related to the taking behavior in the refrigerator is determined according to the behavior semantics and the behavior heat of the taking behavior.

[0087] The target region refers to the internal space of the refrigerator directly related to the user's taking behavior, such as the upper layer of the refrigeration chamber, a drawer of the freezer, etc. The related content about the target region can be determined according to the item storage location in the item information of the taken item, and the granularity of the region division can be set according to the business scenario.

[0088] The preservation adjustment parameter refers to the specific settings that need to be adjusted in order to optimize the preservation effect of the target area, including temperature, humidity, air flow speed, etc. If the user frequently takes out vegetable items, the humidity of the target area can be adjusted to a high humidity mode of 80%-90% to prolong the preservation time of vegetables; if the user frequently takes out beverages, the temperature of the target area can be lowered to near freezing point, such as set to 2-4℃, to maintain the cool taste of the beverage; for areas with high behavior heat, the preservation parameters can be adjusted first. For example, if a certain area is frequently accessed every day, it is necessary to ensure that the temperature and humidity of the area are always in the best state. For areas with low behavior heat, the preservation requirements can be appropriately relaxed to save energy. In addition, the behavior semantics and behavior heat can also be combined to develop dynamic adjustment strategies. For example, for high-frequency taking and high-preservation-required items such as fresh meat, the temperature of the target area can be set to a lower value, and the air circulation can be increased

[0089] Step 104, according to the preservation adjustment parameter, adjusting the preservation environment of the target area of the refrigerator.

[0090] The preservation adjustment parameters of the refrigerator, such as temperature, humidity, air flow speed, etc., can be adjusted by the refrigerator's temperature control system after obtaining the preservation adjustment parameters to adjust the corresponding preservation environment of the target area of the refrigerator; the temperature control system can accurately control the temperature by adjusting the refrigeration power of the target area. For example, reduce the temperature of a certain layer of the refrigeration chamber from 5℃ to 3℃; the humidity level of the target area can also be adjusted by a humidifying or dehumidifying device. For example, increase the humidity of the vegetable storage area from 70% to 85%; and optimize the air circulation of the target area by changing the fan speed. For example, increase the cold air flow of a certain area to quickly cool down. The refrigerator system will continuously collect new behavior data and dynamically adjust the preservation parameters of the target area according to the latest behavior semantics and behavior heat. For example, if the user suddenly starts to frequently take out a new item, the refrigerator will quickly adjust the settings of the relevant area to adapt to this change.

[0091] The behavior information corresponding to the taking behavior of each user in the home environment to the items in the refrigerator is determined; the behavior information includes at least one of user information initiating the taking behavior, item information of the taken item, occurrence time of the taking behavior, and environmental information at the time of the taking behavior; the behavior semantics and the behavior heat of the taking behavior are determined according to the behavior information; the preservation adjustment parameter of the target area in the refrigerator related to the taking behavior is determined according to the behavior semantics and the behavior heat of the taking behavior, the use frequency of different areas in the refrigerator can be determined according to the behavior heat, and the preservation strategy of each area of the refrigerator is more targeted and flexible in combination with the behavior semantics; the target area of the refrigerator is adjusted according to the preservation adjustment parameter, thereby realizing the transformation of the refrigerator from a single function device to an intelligent management device, enabling the refrigerator to dynamically adjust the running state according to the behavior semantics and the behavior heat, optimizing the dynamic adjustment capability of different areas of the refrigerator, and balancing the preservation effect and energy consumption problems between different areas.

[0092] Referring to Figure 2 , a step flow chart of another refrigerator control method embodiment of the application is shown, which can specifically include the following steps:

[0093] Step 201, determining behavior information corresponding to the taking behavior of each user in the home environment to the items in the refrigerator; the behavior information includes at least one of user information initiating the taking behavior, item information of the taken item, occurrence time of the taking behavior, and environmental information at the time of the taking behavior;

[0094] In an embodiment, the behavior information includes item information of the taken item, and step 201 can include the following sub-steps:

[0095] Sub-step S11, acquiring action images of the user using the refrigerator and weight change information of the target area inside the refrigerator;

[0096] A camera can be installed inside or outside the refrigerator, which will capture the action images of the user in front of the refrigerator and inside the refrigerator when the user opens the refrigerator door to operate. These images can show the hand movements of the user, such as information about which area to take the item from, the shape of the taken item, etc.

[0097] For the weight change information, a weight sensor can be installed in each area inside the refrigerator, such as different layers, drawers, etc. When the user takes the item, the weight of the corresponding area will change. By monitoring this weight change, it can be preliminarily judged whether the item is taken and the approximate weight range of the taken item.

[0098] Sub-step S12, in the case where it is determined according to the weight change information that the user has performed the taking behavior, determining the item information of the taken item of the user according to the action images.

[0099] If the weight sensor detects a weight decrease in a certain area, and the decrease is consistent with the characteristics of an item being taken, such as a reduced weight within a reasonable range of item weights, it can be determined that the user has performed a taking action; after determining that the user's behavior towards the refrigerator belongs to a taking action, the specific information of the taken item is further determined using the action image. For example, by analyzing the appearance characteristics of the item in the action image, such as color, shape, text or pattern on the packaging, etc., combined with a pre-established item image database, the specific item information is identified. For example, if the image shows that a red, cylindrical object with a Coca-Cola logo is taken, it can be determined that the taken item is Coca-Cola. For image data acquisition, an embedded lightweight object recognition model can be used to identify the item category that the user is interacting with, such as existing technologies such as YOLO model.

[0100] In addition to the specific category of the taken item, the value of the taken item can also be obtained. This value judgment can be achieved by combining user-defined rules with an embedded knowledge base. When the item salmon is identified through the image, the system will query the value of salmon in this knowledge base; for special items such as medicines that directly affect the safety of the user's life, the highest value will be assigned in the knowledge base. In addition, the user can interact with the refrigerator to manually mark the value of a certain item. For example, the user can mark a bowl of leftovers as having to be eaten tonight, and the system will temporarily increase the value of the item. In addition, the system can dynamically learn to adjust the value of the item corresponding to the item placed in the refrigerator through an embedded deep learning module. By analyzing image data, the user's behavior and the information of the taken item can be more flexible and accurate, providing a data basis for subsequent analysis.

[0101] Step 202, determining the behavior label of the taking behavior according to the item information of the taken item and the occurrence time of the taking behavior; the behavior label is used to represent the category and occurrence time of the taking behavior;

[0102] The category and occurrence time of the taking behavior are combined to form a behavior label; for example, taking beverages-15:00 or taking vegetables-18:00. This behavior label helps to further analyze the user's behavior habits, such as finding that the user tends to take a certain type of item during a certain time period, thereby providing a basis for optimizing the preservation needs of the refrigerator.

[0103] Exemplarily, the behavior label of the taking behavior can be determined according to an adaptive resonance theory (ART), which is a neural network model for pattern recognition and classification. It has online learning ability and can dynamically adjust the classification result in the process of continuously receiving new data while maintaining the memory of existing knowledge. This makes the ART model very suitable for dealing with the dynamic change of user behavior patterns in the home environment.

[0104] One of the ways of using the ART model to determine the behavior label can be:

[0105] After collecting multi-dimensional data from the refrigerator system and preprocessing, an effective input feature vector X is constructed. The target detection algorithm or action recognition algorithm such as YOLO can be used to extract key action sequences from video streams. The output key information includes the category of the object and the hand trajectory information. The current time is recorded synchronously, and the user identity number of the user initiating the behavior is determined through face recognition, fingerprint recognition, etc. The time stamp can be used to analyze the time distribution of the behavior, and the user identity can help to distinguish the habits of different users.

[0106] For different types of feature data, such as continuous numerical features like weight change information and timestamp data, and discrete feature data like action classification output from images, the input vector X finally constructed can be [image features, weight change information, timestamp, user identity number]

[0107] Generally, the ART model can include three layers, input layer, matching layer and output layer; in the embodiment of the present application, the setting and role of the input layer are the same as those of the general ART model, which will not be described here. For the similarity threshold in the matching layer of the ART network, in order to cope with the dynamic changes in the home environment such as the addition of new users and the change of user habits, a dynamic similarity threshold adjustment mechanism is introduced.

[0108] Exemplarily, one of the adjustment mechanisms can be to set a reference similarity threshold base Generally, it will be set when the user first uses the refrigerator system to control the refrigerator or the control system of the refrigerator is reset. When the system detects a new user, the system will set base- a, to achieve fast learning of user learning, and achieve the effect of accurate classification of taking behavior. Wherein, a is a learning sensitivity factor. This learning sensitivity factor can be flexibly set according to business needs. After reducing the value of p, the behavior of the new user is more likely to be identified as a new category, thereby quickly establishing a behavior model for the new user; after learning several stable behavior patterns of the user, or after a preset time period, the similarity threshold can be gradually restored to p base ;

[0109] When the system does not create any new behavior category for a long period of time, such as continuously for a month, it indicates that the current behavior pattern tends to be stable, and at this time the similarity threshold is set to p base + β, wherein β is a "category merging factor" (for example, β = 0.05). After increasing the value of p, some very similar behavior categories are merged to prevent over-fragmentation of behavior categories.

[0110] In addition, the user is allowed to intervene through the App or the interactive interface, for example, when a major change in life habits is about to occur, such as starting to exercise or the birth of a new baby, the user can set the system to relearn the taking behavior classification of the user. At this time, the system will set the value of p to a lower initial value such as 0.6 p base , and clear the old behavior heat data. At this time, the system will enter a fast learning period and rebuild the behavior model. After classifying the taking behavior, the specific behavior label can be generated according to the output result of the ART network.

[0111] For example, the behavior label can be frequent beverage taking, first vegetable storage, milk taking at breakfast time, snack taking at midnight, 4-day non-use in the cold storage area, excessive use in the freezer area, etc. The specific label setting logic will be set according to the system business needs.

[0112] Through the optimization design of the ART model and its dynamic similarity threshold p, the dynamic change problem of user behavior patterns in the home environment can be effectively solved. Precise behavior classification and label setting are achieved, and data support is provided for refrigerator intelligence.

[0113] Step 203, determining the behavior heat of the taking behavior according to the behavior label;

[0114] The behavior heat is a quantitative indicator to measure the activity level of user behavior. It not only considers the frequency of taking behavior, but also combines the time distribution characteristics and heat decay law of taking behavior. By calculating the behavior heat, the importance of a behavior label can be evaluated, and a basis is provided for subsequent adjustment of the refrigerator preservation strategy.

[0115] In one embodiment, step 203 can include the following sub-steps:

[0116] Sub-step S21, obtain the occurrence time of at least one historical taking behavior corresponding to the behavior label and the behavior heat decay coefficient corresponding to the behavior label;

[0117] The occurrence time of the historical taking behavior refers to the time point of each taking behavior related to a specific behavior label. For example, for the behavior label of taking milk during breakfast time, there may be multiple occurrence times of taking behavior, such as 2025-07-14 07:15:30, 2025-07-13 07:10:15, etc.

[0118] The behavior heat decay coefficient (λ) is a parameter for measuring the degree of influence of taking behavior on behavior heat over time. Its value will be between 0 and 1. If λ is larger, it means that the influence of taking behavior a long time ago on current behavior heat is still larger. If λ is smaller, it means that the influence of taking behavior a long time ago on current behavior heat rapidly weakens.

[0119] Sub-step S22, according to the occurrence time of the taking behavior, the occurrence time of the historical taking behavior and the heat decay coefficient, determine the behavior heat of the taking behavior.

[0120] The more remote the behavior, the smaller its contribution to the behavior heat. In order to reflect the rule that heat gradually weakens over time, the contributions of all related behaviors can be weighted and summed to obtain the final behavior heat value. This comprehensive calculation method ensures that the behavior heat can dynamically reflect the change trend of user behavior.

[0121] In an embodiment, sub-step S22 can include the following sub-steps:

[0122] Sub-step S221, according to the occurrence time of the taking behavior, the occurrence time of the historical taking behavior and the heat decay coefficient, determine the behavior heat of the taking behavior according to the following formula:

[0123]

[0124] Wherein, H represents the behavior heat, λ represents the behavior heat decay coefficient, and Δt represents the time difference between the occurrence time of at least one taking behavior and the current time.

[0125] Generally, recent behaviors will have higher weights: therefore, in the formula, the heat decay coefficient λ is introduced to use the mathematical rule of the exponential function to naturally achieve the effect that the closer to the current time the taking behavior is, the greater its contribution to the behavior heat. In this way, the activity level of taking behavior under a certain behavior label can be dynamically reflected, considering factors such as the recency and frequency of behavior, thereby providing a basis for subsequent operations such as adjusting the fresh-keeping strategy of the refrigerator area based on behavior heat;

[0126] In addition to the above method, the frequency of behavior execution can be obtained according to a statistical algorithm, and the frequency is taken as the behavior heat. For example, when the statistical base is small, for example, less than a preset threshold, a linear counting method can be used for statistics, and when the statistical base is greater than the threshold, HyperLogLog, HyperLogLog++, or the like can be used for statistics. When the HyperLogLog technology is used, the system will additionally use the number of zero values (zero_count) in the bucket to make a secondary correction to the estimation result in the case of a smaller base, such as greater than a preset threshold but less than the use threshold of HyperLogLog. The specific correction formula can be: Since the correction method, HyperLogLog, and HyperLogLog++ are all existing bases, they will not be described here.

[0127] Step 204, determining the behavior semantics of the taking behavior according to the behavior information;

[0128] In an embodiment, the item information of the taken item includes the category of the taken item and the quantity of the taken item, and step 204 can include the following sub-steps:

[0129] Sub-step S31, determining the semantic feature vector of the taking behavior according to the shelf life of the taken item, the category of the taken item, and the quantity of the taken item;

[0130] The shelf life of the taken item refers to the length of time that the item can maintain its quality under certain storage conditions. For example, milk may have a shelf life of 7 days, and some fruits may only have a few days. The shelf life information is important for understanding the taking behavior. If the user frequently takes items that are about to expire, it may indicate that the user is cleaning up food that is about to spoil; if the taken item has a long shelf life, it may be for long-term storage. The quantity of the taken item indicates how much of that item is involved in a taking behavior. For example, taking two bottles of cola or five bottles of cola. The quantity can reflect the user's demand intensity or consumption scale. Taking a large amount of a certain item may mean a family gathering or a high consumption rate of that item. The semantic feature vector is a multidimensional vector, and each dimension corresponds to an attribute of the taken item, such as shelf life, category, quantity, etc. By numerically combining these attributes, a vector is formed.

[0131] For example, it can be represented by a three-dimensional vector, the first dimension represents the shelf life, the implementation can be represented by the remaining shelf life days, such as 5 days for 5, the second dimension represents the category of the taken goods, the implementation can be represented by the code, such as 1 for beverage, 2 for food, 3 for fresh food, etc., and the third dimension represents the quantity of the taken goods, which can be directly represented by the taken quantity, such as 3 bottles for 3. Then if the taken goods are beverages with a remaining shelf life of 5 days and a quantity of 3 bottles, the semantic feature vector may be [5, 1, 3].

[0132] Sub-step S32, determining the behavior semantics of the taking behavior according to the semantic feature vector.

[0133] When there is a semantic feature vector, the meaning behind the taking behavior, i.e. the behavior semantics, can be accurately inferred from the vector. This inference can be made through pre-set rules or machine learning models. If it is a rule, for example, it is stipulated that if the shelf life dimension value is small, the category dimension is fresh food, and the quantity is large, the behavior semantics may be to clean up the fresh food that is about to deteriorate; if it is a machine learning model, the semantic feature vector is input into the trained model, and the model will output the corresponding behavior semantics according to the data patterns learned before.

[0134] It should be noted that one semantic feature vector may correspond to multiple behavior semantics. For example, in the case of taking beverages with short shelf life and small quantity, there may be different semantics in different situations such as different time periods and different user identities, such as trying new flavors or trying new flavors. So when determining the behavior semantics, more context information such as the time mentioned above and the user identity should be combined for comprehensive judgment to improve the accuracy of the behavior semantics determination.

[0135] Illustratively, one way to infer the behavior semantics of the taking behavior based on the semantic feature vector can be to map the semantic feature vector to a specific behavior semantics through pre-defined rules. If the shelf life < 2 days and the category = fresh food and the quantity > 5, the behavior semantics is to clean up the fresh food that is about to deteriorate; if the shelf life > 7 days and the category = beverage and the quantity = 1, the behavior semantics is to temporarily quench thirst.

[0136] In another example, the way to infer the behavior semantics of the taking behavior based on the semantic feature vector can also be a machine learning-based method, which trains a classification model such as decision tree, random forest, support vector machine, etc. using historical data, and inputs the semantic feature vector into the model to predict the behavior semantics.

[0137] In another example, the behavior semantic manner of inferring the taking behavior based on the semantic feature vector can also be a deep learning based method, using neural networks such as multilayer perceptron, convolutional neural network, etc. to model the semantic feature vector, suitable for high-dimensional features and complex patterns.

[0138] In addition, the dimension of the semantic feature vector can be further supplemented, and in addition to the above three-dimensional vector, context information such as time, user identity, and environmental information can be combined to comprehensively judge the semantics of the taking behavior by combining the machine learning or deep learning methods mentioned above.

[0139] For example: Suppose a family member A takes a bottle of milk (with a shelf life of 3 days, with 1 day remaining) from the refrigerator on July 15, 2025 at 7:30 am. At the same time, the system records that the weather is hot (35°C) on that day, and family member A's habit is to drink milk every morning as part of breakfast (the habit of family behavior can be output as a behavior label by the ART model, or by analyzing the user's historical data, which is not described here) ;

[0140] At this time, the data that constitutes the semantic feature vector can be: shelf life: 1 day (about to expire), type: beverage (milk), quantity: 1 bottle, time: 7:30 am (breakfast time), user identity: family member A, environmental information: hot weather (35°C), user habit: drink milk every morning. For ease of description, the behavior semantic analysis at this time adopts the first method mentioned above, that is, the semantic analysis is performed by pre-defined rules;

[0141] According to the pre-set rule shelf life < 2 days and type = beverage and quantity = 1, the behavior semantics is initially inferred as "cleaning up the beverage about to deteriorate. But this inference ignores the context information and may not be accurate. Further analysis combined with context information is time distribution: 7:30 am is family member A's breakfast time, which is consistent with his habit of drinking milk every morning; user identity: family member A usually drinks milk in the morning, indicating that this is a daily dietary behavior; environmental information: hot weather may cause the consumption of milk to speed up, but this does not affect the user's normal demand for milk; user habit: family member A has the habit of drinking milk every morning, further verifying the regularity of the behavior. Based on the above analysis, the behavior semantics should be "drink milk for breakfast", not "clean up the beverage about to deteriorate".

[0142] By combining semantic feature vectors and contextual information, the behavior semantics of user taking behavior can be more accurately determined. This method not only considers the attributes of the item itself, but also integrates time, user identity, environmental information and other multi-dimensional factors, thereby better reflecting the real intention of user behavior. This comprehensive analysis method is particularly suitable for complex and variable user behavior patterns in a home environment.

[0143] In step 205, according to the behavior semantics and the behavior heat of the taking behavior, the preservation adjustment parameter of the target area related to the taking behavior in the refrigerator is determined.

[0144] In an embodiment, step 205 can include the following sub-steps:

[0145] In sub-step S41, the user demand influence factor is determined according to the behavior semantics and the behavior heat.

[0146] The behavior semantics S and the behavior heat H can be combined to determine the user demand influence factor. For example, different behavior semantics weights W s and behavior heat weights W h are set, and then the user demand influence factor is determined according to the formula W s ×S+W h ×H. As mentioned above, the vector representation of the behavior semantics is [shelf life of the taken item, type of the taken item, quantity of the taken item], and the quantified representation of the behavior semantics can be calculated by weighted summation according to the data in each dimension of the vector representation. For example, for the shelf life dimension of the taken item, the shorter the remaining shelf life, the higher the weight. For example, the weight is set to 0.8 when the remaining shelf life is less than 1 day, and the weight is set to 0.2 when the remaining shelf life is more than 7 days.

[0147] For the type dimension of the taken item, different types of items have different influences on user demand. For example, the weight of fresh food items can be set to 0.6, and the weight of beverage items can be set to 0.3. For medicine items, the weight can be set to 0.8.

[0148] For the quantity dimension of the taken item, the more the quantity, the greater the demand. A linear growth method can be used, for example, the weight increases by 0.1 for every 1 unit increase in quantity. According to actual needs, the quantified representation of the behavior semantics obtained by preliminary calculation can be normalized to the same dimension as the behavior heat value to facilitate subsequent calculation. The semantic weight W s and the behavior heat weight W h can be determined flexibly according to business needs or user identity; for example, if a user who frequently uses the refrigerator pays more attention to healthy diet, the behavior semantics weight W sIf set to 0.7, the weight of behavior popularity W is reduced. h For example, set it to 0.3. If the user of the refrigerator is a user who prefers personalized services, the behavior heat weight W can be increased. h If set to 0.6, the behavioral semantic weight W is reduced. s For example, set it to 0.4.

[0149] The behavior popularity H has already been calculated through the previous steps and is usually a value between 0 and 1. If further adjustment is needed, it can be weighted or non-linearly transformed according to business requirements. For example, for highly popular behaviors, the popularity value can be squared to amplify its impact.

[0150] Sub-step S42: Determine the energy consumption influencing factors based on the preset refrigerator power data;

[0151] A refrigerator's power output determines its energy consumption at different preservation intensities. When adjusting the preservation strategy for a target area of ​​the refrigerator, energy consumption needs to be considered to achieve a balance between energy saving and preservation effectiveness.

[0152] Assume the refrigerator has multiple power settings, each corresponding to a different energy consumption rate. The energy consumption E per unit time at each power setting can be measured or estimated in advance. i Then, based on the current operating status of the refrigerator and the adjustment needs of the target area, select the appropriate power level and calculate the energy consumption impact factor E as follows: E = (E max -E min ) / (E current -E min )

[0153] Among them, E current It is the energy consumption per unit time at the currently selected power level, E min and E max These represent the minimum and maximum energy consumption per unit time across all power levels. The energy consumption impact factor calculated in this way ranges from 0 to 1, with a smaller value indicating less energy consumption.

[0154] Sub-step S43: Based on the user demand influence factor and the energy consumption influence factor, determine the preservation adjustment mapping coefficient of the target area in the refrigerator related to the retrieval behavior.

[0155] In order to determine the preservation adjustment mapping coefficient of the target area, the shelf life influence factor, the user demand influence factor and the energy consumption influence factor need to be comprehensively considered. The preservation adjustment mapping coefficient reflects the degree of preservation adjustment of the target area of the refrigerator after comprehensively considering various factors. The greater the coefficient value is, the higher the preservation adjustment intensity required for the area, such as reducing the temperature, increasing the humidity, etc. The smaller the coefficient value is, the preservation requirement can be appropriately relaxed to save energy. In this way, the preservation environment of the target area of the refrigerator can be more targeted controlled, the user experience can be improved, and the resources can be saved.

[0156] In an embodiment, the sub-step S44 can include the following sub-steps:

[0157] In sub-step S441, according to the user demand influence factor and the energy consumption influence factor, the preservation adjustment mapping coefficient of the target area related to the taking behavior in the refrigerator is determined according to the following formula:

[0158] P=w 1* E+w 2* U, wherein E represents the energy consumption influence factor, U represents the user demand influence factor, P represents the mapping coefficient, w1 represents the energy consumption influence factor weight, and w2 represents the user demand influence factor weight.

[0159] According to the mapping coefficient, the preservation adjustment parameter of the target area related to the taking behavior in the refrigerator is determined.

[0160] The formula is used to comprehensively calculate the preservation adjustment mapping coefficient P of the target area, which reflects the degree of preservation adjustment required for the target area of the refrigerator on the basis of considering the user demand and the energy consumption.

[0161] w1 is the weight of the energy consumption influence factor, and w2 is the weight of the user demand influence factor.

[0162] The allocation of the weights can be flexibly adjusted according to business requirements or user preferences. For example, if energy saving is more important, the value of w1 can be increased, such as being set to 0.7, and the value of w2 can be reduced, such as being set to 0.3.

[0163] If the user experience is more important, the value of w2 can be increased, such as being set to 0.6, and the value of w1 can be reduced, such as being set to 0.4. The two weights satisfy w1+w2=1, ensuring the proportional relationship between them.

[0164] The mapping coefficient P is an intermediate variable for guiding the specific setting of the subsequent target area preservation adjustment parameter. The preservation adjustment parameters of the refrigerator, such as temperature, humidity, air flow speed, etc., usually have a physical range. For example, the temperature range is [1℃, 10℃], and the humidity range is [30%, 90%]. At this time, if P is close to 1, such as P = 0.86, the temperature is adjusted to be close to the low temperature end, such as 2℃, and the humidity is adjusted to be close to the high humidity end, such as 80%. If P is close to 0, such as P = 0.2, the temperature is adjusted to be close to the high temperature end, such as 8℃, and the humidity is adjusted to be close to the low humidity end, such as 40%. The system can adjust the preservation strategy of the target area in real time according to the user behavior and the running state of the refrigerator, and can dynamically optimize the preservation strategy of the target area of the refrigerator according to the mapping coefficient, so as to realize the double improvement of user satisfaction and energy efficiency.

[0165] Step 206, adjusting the preservation environment of the target area of the refrigerator according to the preservation adjustment parameter.

[0166] By determining the behavior information corresponding to the taking behavior of each user in the home environment to the items in the refrigerator; the behavior information includes at least one of the user information of the taking behavior, the item information of the taken item, the occurrence time of the taking behavior and the environmental information at the time of the taking behavior; according to the behavior information, the behavior semantics and the behavior heat of the taking behavior are determined; according to the behavior semantics and the behavior heat of the taking behavior, the preservation adjustment parameter of the target area related to the taking behavior in the refrigerator is determined, the use frequency of different areas in the refrigerator can be determined according to the behavior heat, and the preservation strategy of each area of the refrigerator is further formulated more targeted and flexibly in combination with the behavior semantics; the target area of the refrigerator is adjusted according to the preservation adjustment parameter, so as to realize the transformation of the refrigerator from a single function device to an intelligent management device, so that it can dynamically adjust the running state according to the behavior semantics and the behavior heat, optimize the dynamic adjustment capability of different areas of the refrigerator, and balance the preservation effect and energy consumption problem between different areas.

[0167] It should be noted that, for the method embodiment, in order to simply describe, it is expressed as a series of action combinations, but those skilled in the art should know that the embodiment of the present application is not limited by the described action sequence, because according to the embodiment of the present application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the involved actions are not necessarily required by the embodiment of the present application.

[0168] Referring to Figure 3 , a structural block diagram of an embodiment of a refrigerator control device of the present application is shown, which can specifically include the following modules:

[0169] The behavior information acquisition module 301 is configured to determine behavior information corresponding to a taking behavior of an article in the refrigerator by each user in the family environment; the behavior information includes at least one of user information initiating the taking behavior, article information of the taken article, a time of occurrence of the taking behavior, and environment information at the time of occurrence of the taking behavior;

[0170] The semantic heat determination module 302 is configured to determine a behavior semantic and a behavior heat of the taking behavior according to the behavior information.

[0171] The parameter determination module 303 is configured to determine a preservation adjustment parameter of a target region of the refrigerator related to the taking behavior according to the behavior semantic and the behavior heat of the taking behavior.

[0172] The adjustment module 304 is configured to adjust a preservation environment of the target region of the refrigerator according to the preservation adjustment parameter.

[0173] In an embodiment, the behavior information includes the article information of the taken article and the time of occurrence of the taking behavior, and the semantic heat determination module includes:

[0174] The behavior label determination sub-module is configured to determine a behavior label of the taking behavior according to the article information of the taken article and the time of occurrence of the taking behavior; the behavior label is used to represent a category and a time of occurrence of the taking behavior.

[0175] The behavior heat determination sub-module is configured to determine a behavior heat of the taking behavior according to the behavior label.

[0176] In an embodiment, the behavior heat determination sub-module includes:

[0177] The decay coefficient determination unit is configured to acquire a time of occurrence of at least one historical taking behavior corresponding to the behavior label and a behavior heat decay coefficient corresponding to the behavior label.

[0178] The behavior heat calculation unit is configured to determine the behavior heat of the taking behavior according to the time of occurrence of the taking behavior, the time of occurrence of the historical taking behavior, and the heat decay coefficient.

[0179] In an embodiment, the behavior heat calculation unit includes:

[0180] The first formula calculation unit is configured to determine the behavior heat of the taking behavior according to the time of occurrence of the taking behavior, the time of occurrence of the historical taking behavior, and the heat decay coefficient according to the following formula:

[0181]

[0182] Wherein, H represents behavior heat, λ represents the behavior heat decay coefficient, and Δt represents the time difference between the occurrence time of the at least one taking behavior and the current time.

[0183] In an embodiment, the item information of the taken item includes the category of the taken item and the quantity of the taken item, and the semantic heat determination module includes:

[0184] a semantic determination sub-module configured to determine a semantic feature vector of the taking behavior according to the shelf life of the taken item, the category of the taken item, and the quantity of the taken item.

[0185] a behavior semantic determination sub-module configured to determine a behavior semantic of the taking behavior according to the semantic feature vector.

[0186] In an embodiment, the parameter determination module includes:

[0187] a first influence factor determination sub-module configured to determine a user demand influence factor according to the behavior semantic and the behavior heat;

[0188] a second influence factor determination sub-module configured to determine an energy consumption influence factor according to preset refrigerator power data;

[0189] a mapping coefficient determination sub-module configured to determine a preservation adjustment mapping coefficient of a target region related to the taking behavior in the refrigerator according to the user demand influence factor and the energy consumption influence factor.

[0190] In an embodiment, the mapping coefficient determination sub-module includes:

[0191] a second formula calculation unit configured to determine the preservation adjustment mapping coefficient of the target region related to the taking behavior in the refrigerator according to the user demand influence factor and the energy consumption influence factor according to the following formula:

[0192] P=w 1* E+w 2* U, wherein E represents the energy consumption influence factor, U represents the user demand influence factor, P represents the mapping coefficient, w1 represents the energy consumption influence factor weight, and w2 represents the user demand influence factor weight.

[0193] a preservation parameter determination unit configured to determine a preservation adjustment parameter of the target region related to the taking behavior in the refrigerator according to the mapping coefficient.

[0194] In an embodiment, the behavior information includes item information of a taken item, and the behavior information acquisition module includes:

[0195] An image and weight information acquisition submodule is configured to acquire an action image of the user using the refrigerator and weight change information of the target region inside the refrigerator.

[0196] An article information determination submodule is configured to, in a case where it is determined that the user performs a taking action according to the weight change information, determine article information of the article taken by the user according to the action image.

[0197] By determining the behavior information corresponding to the taking action of the article in the refrigerator by each user in the home environment, the behavior information includes at least one of the user information of the user initiating the taking action, the article information of the article taken, the occurrence time of the taking action and the environmental information at the time of the taking action, the behavior semantics and the behavior heat of the taking action are determined according to the behavior information, the preservation adjustment parameter of the target region in the refrigerator related to the taking action is determined according to the behavior semantics and the behavior heat of the taking action, the use frequency of different regions in the refrigerator can be determined according to the behavior heat, and the preservation strategy of each region in the refrigerator is more targeted and flexible by further combining the behavior semantics. According to the preservation adjustment parameter, the target region of the refrigerator is adjusted, so that the refrigerator is changed from a single function device to an intelligent management device, which can dynamically adjust the running state according to the behavior semantics and the behavior heat, optimize the dynamic adjustment capability of different regions in the refrigerator, and balance the preservation effect and energy consumption problems between different regions.

[0198] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the related parts refer to the part of the method embodiment.

[0199] The refrigerator embodiment also provides a refrigerator, which comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, the computer program is executed by the processor to realize each process of the refrigerator control method embodiment, and the same technical effects can be achieved, and details are not repeated here.

[0200] The refrigerator embodiment also provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program is executed by the processor to realize each process of the refrigerator control method embodiment, and the same technical effects can be achieved, and details are not repeated here.

[0201] Each embodiment in the specification is described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to.

[0202] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, apparatus, or computer program product. Accordingly, embodiments of the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, and the like) embodying computer program instructions.

[0203] Embodiments of the present application are described herein with reference to the Figure 1 one or more functions specified in a flow or multiple flows and / or blocks. Figure 1 means for performing one or more functions specified in a flow or multiple flows and / or blocks.

[0204] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the Figure 1 one or more functions specified in a flow or multiple flows and / or blocks. Figure 1 means for performing one or more functions specified in a flow or multiple flows and / or blocks.

[0205] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the Figure 1 one or more functions specified in a flow or multiple flows and / or blocks. Figure 1 means for performing one or more functions specified in a flow or multiple flows and / or blocks.

[0206] While preferred embodiments of the present application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the foregoing description. Therefore, the appended claims are intended to encompass within their scope all such variations and modifications as are included within the scope of the present application.

[0207] Finally, it is to be understood that the phraseology or terminology such as "first" and "second" etc. used herein is merely intended to differentiate one entity or operation from another entity or operation, without necessarily requiring or implying any actual such relationship or order between such entities or operations. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0208] The above provides a detailed description of the refrigerator control method, device, refrigerator and medium provided by the application. The principles and implementation modes of the application are described by applying specific examples. The above example is only used to help understand the method and core idea of the application. Meanwhile, for those skilled in the art, the specific implementation modes and application ranges can be changed according to the idea of the application. In conclusion, the content of the specification should not be understood as a limitation of the application.

Claims

1. A refrigerator control method, characterized in that, The method includes: Determine the behavioral information corresponding to each user's actions of taking items from the refrigerator in the home environment; the behavioral information includes at least one of the following: user information initiating the taking action, item information of the item taken, time of the taking action, and environmental information at the time of the taking action; Based on the behavioral information, determine the behavioral semantics and behavioral popularity of the taking behavior; Based on the semantics and popularity of the taking action, determine the preservation adjustment parameters of the target area in the refrigerator related to the taking action; The preservation environment of the target area of ​​the refrigerator is adjusted according to the preservation adjustment parameters.

2. The refrigerator control method according to claim 1, characterized in that, The behavioral information includes item information of the item being taken and the time when the taking action occurred. Determining the behavioral intensity of the taking action based on the behavioral information includes: Based on the item information of the item being taken and the time when the taking action occurred, a behavior tag for the taking action is determined; the behavior tag is used to indicate the type and time of the taking action. Based on the behavior tags, determine the popularity of the taking behavior.

3. A refrigerator control method according to claim 2, characterized in that, The step of determining the behavioral popularity of the taking behavior based on the behavioral tag includes: Obtain the occurrence time of at least one historical retrieval behavior corresponding to the behavior tag, and the behavior heat decay coefficient corresponding to the behavior tag; The behavior intensity of the taking action is determined based on the occurrence time of the taking action, the occurrence time of the historical taking actions, and the intensity decay coefficient.

4. A refrigerator control method according to claim 3, characterized in that, The determination of the behavioral heat of the taking action based on the occurrence time of the taking action, the occurrence time of the historical taking actions, and the heat decay coefficient includes: Based on the occurrence time of the taking action, the occurrence time of the historical taking actions, and the heat decay coefficient, the heat of the taking action is determined according to the following formula: Wherein, H represents the behavior heat, λ represents the behavior heat decay coefficient, and Δt represents the time difference between the occurrence time of at least one take-up action and the current time.

5. A refrigerator control method according to claim 1, characterized in that, The item information for taking items includes the type and quantity of the items taken. Determining the behavioral semantics of the taking action based on the behavioral information includes: The semantic feature vector of the taking action is determined based on the shelf life of the item, the type of item, and the quantity of the item. The semantics of the taking action are determined based on the semantic feature vector.

6. A refrigerator control method according to claim 5, characterized in that, The step of determining the preservation adjustment parameters of the target area in the refrigerator related to the taking action based on the behavioral semantics and behavioral heat of the taking action includes: Based on the aforementioned behavioral semantics and behavioral popularity, determine the factors influencing user demand; Based on the preset refrigerator power data, determine the energy consumption influencing factors; Based on the user demand influencing factor and the energy consumption influencing factor, determine the preservation adjustment mapping coefficient of the target area in the refrigerator related to the retrieval behavior.

7. A refrigerator control method according to claim 6, characterized in that, The step of determining the preservation adjustment mapping coefficient of the target area in the refrigerator related to the retrieval behavior based on the shelf life influencing factor, the user demand influencing factor, and the energy consumption influencing factor includes: Based on the user demand influencing factor and the energy consumption influencing factor, the preservation adjustment mapping coefficient of the target area in the refrigerator related to the retrieval behavior is determined according to the following formula: P = w 1* E+w 2* U, where E represents the energy consumption impact factor, U represents the user demand impact factor, P represents the mapping coefficient, w1 represents the weight of the energy consumption impact factor, and w2 represents the weight of the user demand impact factor. The preservation adjustment parameters of the target area in the refrigerator related to the retrieval behavior are determined based on the mapping coefficient.

8. A refrigerator control method according to claim 1, characterized in that, The behavioral information includes item information related to taking items. The step of determining the behavioral information corresponding to each user's actions of taking items from the refrigerator in the home environment includes: Acquire motion images of the user using the refrigerator and weight change information of the target area inside the refrigerator; If it is determined that the user has performed a picking action based on the weight change information, the item information of the item picked up by the user is determined based on the action image.

9. A refrigerator control device, characterized in that, The device includes: The behavior information acquisition module is used to determine the behavior information corresponding to the actions of each user in the home environment in taking items from the refrigerator; the behavior information includes at least one of the following: user information that initiates the taking action, item information that is taken, time of the taking action, and environmental information at the time of the taking action; The semantic popularity determination module is used to determine the behavioral semantics and behavioral popularity of the taking behavior based on the behavioral information. The parameter determination module is used to determine the preservation adjustment parameters of the target area in the refrigerator related to the taking behavior based on the behavioral semantics and behavioral heat of the taking behavior; An adjustment module is used to adjust the preservation environment of the target area of ​​the refrigerator according to the preservation adjustment parameters.

10. A refrigerator, characterized in that, include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the refrigerator control method as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the refrigerator control method as described in any one of claims 1-8.

Images