Intelligent interaction control method and system suitable for aged kitchen equipment and medium

Abstract

The invention discloses an intelligent interaction control method and system for aged kitchen equipment and a medium, and the method comprises the steps: obtaining the behavior characteristics of an aged user in the auditory sense, visual sense, tactile sense and cognition aspects, and generating a multi-mode behavior characteristic; extracting operation time features and operation program features by using historical operation data of different kitchen devices, setting interaction priorities of the kitchen devices in different time periods, and determining control preference data of the elderly users; and according to the multi-modal behavior characteristics, the interaction priorities of the kitchen devices in different time periods and the control preference data, constructing a user portrait, determining an optimal interaction mode through the user portrait, and according to the optimal interaction mode, generating an interaction control guidance workflow of the different kitchen devices according with the control preferences, thereby realizing interaction prompt for the elderly users. Reasonable interaction prompts are set through user control preferences, so that elderly users are ensured to skillfully operate the intelligent kitchen equipment, and the operation efficiency of the elderly groups on the intelligent kitchen equipment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent kitchen equipment, and more specifically, to an intelligent interactive control method, system and medium for aging-friendly kitchen equipment. Background Art

[0002] In the context of the rapid development of the intelligent era and new technologies, the "digital divide" problem caused by the use of smart products by elderly users has attracted more and more attention from all walks of life. As one of the most frequently used spaces for the elderly in their home care, the importance of the kitchen is self-evident. Today's products are increasingly integrated with intelligent control, and smart kitchen equipment has brought a lot of convenience to people, but with smart kitchen equipment comes the application of intelligent interactive control systems. Today's elderly people have certain obstacles in using interactive interfaces and the Internet. As interfaces replace buttons, these "conveniences" have become troubles for the elderly, thus causing the "elimination of aging" phenomenon of smart home appliances.

[0003] At present, most of the popular smart kitchen equipment on the market serves young groups, ignoring the special needs of vulnerable groups. Therefore, there are still many problems in the aging-friendly design of kitchen equipment, such as inconvenient design, mismatched interaction, unsmooth operation, etc., which makes it difficult for the elderly to live independently. Smart kitchen equipment cannot be controlled according to the specific usage status of the elderly, which easily increases the control operation time of the elderly and increases the difficulty of intelligent control. At the same time, the control process cannot accurately determine whether the control efficiency meets the needs of the elderly, so that unreasonable control causes the elderly's action execution efficiency to decrease. Therefore, in the interactive control of smart kitchen equipment, how to analyze the needs of elderly users and reduce the memory burden of control steps is an urgent problem to be solved. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an intelligent interactive control method, system and medium for aging-friendly kitchen equipment, with the purpose of matching the interactive control of intelligent kitchen equipment according to the needs of users, ensuring that elderly users can operate intelligent kitchen equipment proficiently through reasonable interactive prompts, and solving the problem of difficulty in operating intelligent kitchen equipment for the elderly.

[0005] A first aspect of the present invention provides an intelligent interactive control method for aging-friendly kitchen equipment, comprising: Evaluate the behavioral characteristics of elderly users in terms of vision, hearing, touch and cognition, and integrate the behavioral characteristics of different aspects to generate multimodal behavioral characteristics of elderly users; Extracting operation time characteristics and operation program characteristics using historical operation data of different kitchen appliances, determining the interaction priority of kitchen appliances in different time periods according to the operation time characteristics, and determining the control preference data of elderly users according to the operation program characteristics of different kitchen appliances; Building a user profile based on the multimodal behavior characteristics, the interaction priorities of kitchen equipment at different time periods, and the control preference data, and determining the optimal interaction mode through the user profile; Based on the optimal interaction mode, the interactive control guidance workflow of different kitchen equipment that meets the control preferences of elderly users is obtained, the key control steps in the interactive control guidance workflow are extracted, and the interactive prompt information of the key control steps is generated based on the interactive control guidance workflow, and multi-channel interactive prompts are provided to elderly users.

[0006] In this solution, the behavioral characteristics of elderly users in hearing, vision, touch and cognition are evaluated, and the behavioral characteristics of different aspects are integrated to generate multimodal behavioral characteristics of elderly users, specifically: Obtaining program operation data of different types of kitchen equipment, extracting icon keyword data, guidance prompt sound data, touch vibration data and icon position data using the program operation data, and establishing behavioral characteristic evaluation items in terms of hearing, vision, touch and cognition based on the extracted multi-dimensional data; Sending the behavior characteristic evaluation project to a preset terminal device, and using the preset terminal device to obtain the visual behavior characteristics, auditory behavior characteristics, tactile behavior characteristics and cognitive behavior characteristics of the elderly user; The benchmark value of each modal behavior feature is set according to the mean of the evaluation results of different modal behavior features of young users, and the metric function is constructed using the Euclidean distance. The Euclidean distance between the visual behavior features, auditory behavior features, tactile behavior features and cognitive behavior features of the elderly users and the corresponding benchmark values ​​is obtained through the metric function. Normalization processing is performed according to the Euclidean distance to generate feature scores of different modal behavior characteristics of elderly users, the feature scores are matched with corresponding modal behavior characteristics, and all modal behavior characteristics of elderly users are integrated to generate multimodal behavior characteristics.

[0007] In this solution, the interaction priority of kitchen equipment in different time periods is determined according to the operating time characteristics, and the control preference data of elderly users is determined according to the operating program characteristics of different kitchen equipment, specifically: Extracting historical operation data of different kitchen appliances in the elderly user's kitchen scene within a preset time step, wherein the historical operation data is divided into natural days, and the operation time data of different kitchen appliances within a natural day are extracted to determine the operation time characteristics of different kitchen appliances; Soft clustering is performed using the operating time characteristics of different kitchen appliances, and clustering data sets of each time period are obtained according to the clustering results. The data volume of different categories of kitchen appliances in the clustering data sets is extracted, and the interaction priority of kitchen appliances relative to elderly users in different time periods is determined based on the data volume. Acquire the operation program features of different kitchen appliances according to the historical operation data, acquire the operation program feature sequences of different kitchen appliances through time series integration, and repeatedly divide the operation program feature sequences of different kitchen appliances using different time length division standards; For any kitchen equipment, the running program feature subsequences of different lengths obtained by division are embedded into low-dimensional vectors, and a multi-head attention mechanism is introduced to weight each low-dimensional vector. The weighted low-dimensional vectors are connected to form a feature vector containing all low-dimensional vector features. The feature vector is introduced into a multi-layer perceptron, and a double-layer multi-layer perceptron structure is used to determine the control preference data of the elderly user for the kitchen equipment.

[0008] In this solution, a user profile is constructed based on the multimodal behavior characteristics, kitchen equipment interaction priorities at different time periods, and control preference data, specifically: Calculate user similarity based on multimodal behavior characteristics of elderly users, interaction priorities of kitchen equipment at different time periods, and control preference data, select users who meet the preset similarity criteria as similar users of elderly users, and construct social graph data based on elderly users and similar users; Constructing interaction graph data based on kitchen equipment in the kitchen scene of elderly users, constructing an elderly user graph using the interaction graph data of the social graph data set, taking elderly users and kitchen equipment as nodes in the elderly user graph, setting directed edges based on whether there is a relationship between the nodes, adding multimodal behavior characteristics of elderly users as additional attributes of the elderly user nodes, and adding kitchen equipment interaction priorities and control preference data at different time periods as additional attributes of the kitchen equipment nodes; Use graph neural network to learn the graph data of elderly user graph, use neighborhood random sampling to select the neighborhood nodes of elderly user nodes, introduce multi-head attention mechanism to weight each neighborhood node, and use the weighted neighborhood node pairs for aggregation update; Obtain the elderly user node after aggregation and update, obtain the updated embedding representation of the elderly user node, and after multiple aggregations, concatenate the updated embedding representation of the elderly user node with the original embedding representation to output the user portrait of the elderly user.

[0009] In this solution, the optimal interaction mode is determined by the user portrait, specifically: Obtaining examples of kitchen equipment aging-friendly interactive design, performing graphic processing on the examples of kitchen equipment aging-friendly interactive design to generate graph structure data, encoding the graph structure data, and obtaining user portraits in each example; Different user portrait groups are obtained by clustering, and a user evaluation matrix for interaction modalities is obtained within the user portrait groups, where the elements in the evaluation matrix represent the user's evaluation of the interaction modality; Determine the corresponding user portrait group using the user portrait of the elderly user, perform similarity calculation within the user portrait group to select a preset number of similar users, and select the interaction mode with the highest evaluation based on the evaluation matrix of the similar users; According to the selected interaction mode, the interaction prompts of different kitchen equipment are simulated, the feedback of elderly users on the interaction prompts of different interaction modes is obtained through the terminal device, and the best interaction mode is selected according to the feedback.

[0010] In this scheme, based on the optimal interaction mode, the interactive control guidance workflow of different kitchen equipment that meets the control preferences of elderly users is obtained, the key control steps in the interactive control guidance workflow are extracted, and the interactive prompt information of the key control steps is generated based on the interactive control guidance workflow, specifically: Determine corresponding function program information based on the control preference data of the elderly user, read the operation process of the function program information of different kitchen equipment, and construct an interactive control guidance workflow through the optimal interaction mode; When an elderly user enters the kitchen scene, an operation interaction prompt is generated according to the kitchen equipment interaction priority in different time periods, and it is determined whether the kitchen equipment operated by the elderly user meets the kitchen equipment interaction priority. The frequency of non-satisfaction is counted. When the non-satisfaction frequency is greater than a preset threshold, the kitchen equipment interaction priority of the current time period is reset; Obtain the kitchen equipment operated by the elderly user, determine the corresponding interactive control guidance workflow according to the operation information of the elderly user, obtain the average reaction time of the elderly user for each operation control step in the interactive control guidance workflow, and mark it as a critical control step when the average reaction time is greater than a preset time threshold; Generate interactive prompt information of different interactive modes based on the key control steps, perform multi-channel prompts on elderly users, statistically analyze the average reaction time of elderly users to the key control steps, and cancel the interactive prompt information of different interactive modes when the average reaction time is less than . When the elderly user's reaction time to the operation control step is less than the average reaction time, the corresponding guidance in the interactive control guidance workflow is removed, and the interactive control guidance workflow is updated according to the subsequent feedback from the elderly user.

[0011] A second aspect of the present invention provides an intelligent interactive control system for aging-friendly kitchen equipment, comprising a user multimodal behavior feature evaluation unit, an equipment operation data analysis unit, a user portrait construction unit, an interactive control guidance determination unit, and a control feedback unit; The user multimodal behavior characteristic evaluation unit evaluates the behavior characteristics of the elderly user in terms of hearing, vision, touch and cognition, and integrates the behavior characteristics in different aspects to generate the multimodal behavior characteristics of the elderly user; The device operation data analysis unit extracts operation time characteristics and operation program characteristics using historical operation data of different kitchen devices, determines the interaction priority of kitchen devices in different time periods according to the operation time characteristics, and determines the control preference data of the elderly user according to the operation program characteristics of different kitchen devices; The user portrait construction unit constructs a user portrait according to multimodal behavior characteristics, kitchen equipment interaction priorities at different time periods, and control preference data; The interactive control guidance determination unit determines the best interactive mode and the interactive control guidance workflow that meets the control preference of different kitchen equipment through the user portrait, extracts the key control steps of the interactive control guidance workflow, generates interactive prompt information of the key control steps based on the best interactive mode, and performs multi-channel interactive prompts; The control feedback unit obtains the elderly user's selection feedback on the interactive control guidance workflow, as well as the reaction time to each operation control step of the kitchen equipment under the interactive control guidance workflow, and analyzes and updates the interactive prompt information and the interactive control guidance workflow.

[0012] A third aspect of the present invention provides a computer-readable storage medium, which includes a smart interactive control method program for aging-friendly kitchen equipment. When the smart interactive control method program for aging-friendly kitchen equipment is executed by a processor, the steps of the smart interactive control method for aging-friendly kitchen equipment are implemented.

[0013] Compared with the prior art, the present invention has the following beneficial effects: The present invention selects the most suitable interaction prompt mode for elderly users by analyzing the multimodal behavior characteristics of elderly users, ensuring that the interaction prompt can provide multi-channel interaction prompts for elderly users to the greatest extent, so as to meet their actual needs and reduce the memory burden of traditional equipment control steps.

[0014] The present invention ensures that elderly users are proficient in operating smart kitchen equipment by setting reasonable interaction prompts through user control preferences, extracts key control steps according to the elderly users' proficiency in operating the equipment, sets key interaction prompts in the key control steps, reduces the time elderly users spend on interactive interface selection, and improves the operating efficiency of smart kitchen equipment for the elderly group. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0016] Figure 1 A flow chart showing an intelligent interactive control method for aging-friendly kitchen equipment is shown; Figure 2 A flowchart of an embodiment for obtaining kitchen equipment interaction priority and control preference data of elderly users in different time periods is shown; Figure 3 A flow chart showing an embodiment of determining the optimal interaction mode for an elderly user through a user portrait; Figure 4 A block diagram of an intelligent interactive control system for aging-friendly kitchen equipment is shown. DETAILED DESCRIPTION

[0017] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0018] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0019] Figure 1 A flow chart of an intelligent interactive control method for aging-friendly kitchen equipment is shown.

[0020] like Figure 1 As shown, this embodiment provides an intelligent interactive control method for aging-friendly kitchen equipment, including: S102, evaluating the behavioral characteristics of the elderly user in terms of vision, hearing, touch and cognition, and integrating the behavioral characteristics in different aspects to generate multimodal behavioral characteristics of the elderly user; S104, extracting operation time characteristics and operation program characteristics by using historical operation data of different kitchen appliances, determining the interaction priority of kitchen appliances in different time periods according to the operation time characteristics, and determining the control preference data of the elderly user according to the operation program characteristics of different kitchen appliances; S106, constructing a user portrait according to the multimodal behavior characteristics, the interaction priorities of the kitchen equipment at different time periods, and the control preference data, and determining the optimal interaction mode through the user portrait; S108, based on the optimal interaction mode, obtain the interactive control guidance workflow of different kitchen equipment that meets the control preferences of elderly users, extract the key control steps in the interactive control guidance workflow, generate interactive prompt information of the key control steps based on the interactive control guidance workflow, and provide multi-channel interactive prompts to elderly users.

[0021] It should be noted that the human multi-channel perception system mainly involves the perception of the three sensory channels of vision, hearing and touch and the feedback of the cognitive channel. The sensory channel is mainly used to perceive information, while the cognitive channel is mainly used to process the perceived information and perform actions. In the process of multimodal behavioral feature evaluation of elderly users, program operation data of different categories of kitchen equipment are obtained, such as the operation behavior corresponding to the cooking function program of smart rice cookers. The program operation data is used to extract icon keyword data, guidance prompt sound data, touch vibration data and icon position data, and behavioral feature evaluation items in hearing, vision, touch and cognition are established based on the extracted multidimensional data. The behavior characteristic evaluation project is sent to a preset terminal device, and the visual behavior characteristics, auditory behavior characteristics, tactile behavior characteristics and cognitive behavior characteristics of the elderly user are obtained by using the preset terminal device; preferably, the elderly user's vision and vision adjustment ability, light brightness and discrimination ability, flicker value, contrast and color recognition ability are measured according to the kitchen equipment button and icon keyword data, shape data, color data, and size data to obtain the elderly user's visual behavior characteristics, and by setting kitchen equipment buttons and chart keywords of different sizes, the elderly user's feedback on charts of different sizes is obtained, and the corresponding focus adjustment distance curve is generated to obtain the corresponding vision and vision adjustment ability; when distinguishing small objects, a certain level of illumination is usually required, so the illumination required by the elderly users for distinction is obtained to obtain their distinction ability and light brightness recognition ability; the flicker value is the frequency that can be felt by people when the stimulus light is quickly lit or extinguished. The changes in the stimulation perception of elderly users are most obvious. Therefore, the normal frequency will also appear as continuous light for the elderly with declining vision. The flicker value is used to judge the elderly users' ability to capture flickering light; due to the yellowing of the color of the lens of elderly users, their vision tends to be yellowish. Set buttons and icon keywords of different colors to obtain the elderly users' ability to recognize different colors.

[0022] The hearing and sound discrimination ability of elderly users are tested through the guidance prompt sound data to obtain the auditory behavior characteristics of elderly users. The natural aging of the auditory system makes presbycusis particularly common. The thickening of the walls of the inner ear blood vessels and the shrinking of the lumen lead to functional disorders in the transmission of sound waves from the inner ear to the brain, making it difficult for the elderly to hear the sounds of the outside world. By setting guidance prompt sounds of different decibels, the hearing stimulation threshold curve of elderly users is obtained, the hearing ability of elderly users is evaluated, the feedback of elderly users on stimulation is obtained, and it is determined whether there is mishearing or misunderstanding. The judgment results of different levels of hearing stimulation show that elderly users often have a decreased hearing accuracy, which leads to an increase in the error rate and slow speed of sound information processing. The sound discrimination ability of elderly users is obtained by playing guidance information of different contents.

[0023] The vibration perception threshold and reaction threshold of elderly users are measured by touch vibration data, the tactile behavior characteristics of elderly users are obtained, and the vibration sensory feedback of elderly users is obtained through different vibration intensities. The short-term memory capacity of elderly users is obtained through icon position data, and the memory of the positions of kitchen equipment buttons or icons is evaluated based on the position data of buttons or icons. The cognitive behavior characteristics of elderly users are obtained through the memory coherence of the positions of different buttons or icons. The changes in icon positions are used to judge the capture degree of icon positions by elderly users, and the cognitive behavior characteristics of elderly users are obtained.

[0024] The evaluation results of several young users on the behavioral characteristics evaluation items in auditory, visual, tactile and cognitive aspects are obtained, the benchmark value of each modal behavioral feature is set according to the mean value of the evaluation results of different modal behavioral features of young users, the metric function is constructed using Euclidean distance, and the Euclidean distance between the visual behavioral features, auditory behavioral features, tactile behavioral features and cognitive behavioral features of elderly users and the corresponding benchmark values ​​is obtained through the metric function; the Euclidean distance is normalized according to the Euclidean distance to generate feature scores of different modal behavioral features of elderly users, the higher the feature score, the more normal the corresponding behavioral feature is, the feature score is matched with the corresponding modal behavioral feature, the feature score can be used to distinguish the weak behavioral features of elderly users, and the basis is provided for determining the best interaction mode for elderly users in the future, and the splicing operation is used to integrate all the modal behavioral features after the feature scores are matched to generate multimodal behavioral features of elderly users.

[0025] Figure 2 A flow chart of an embodiment for obtaining kitchen equipment interaction priority and control preference data of elderly users in different time periods is shown.

[0026] According to an embodiment of the present invention, the interaction priority of the kitchen equipment in different time periods is determined according to the operating time characteristics, and the control preference data of the elderly user is determined according to the operating program characteristics of different kitchen equipment, specifically: S202, extracting historical operation data of different kitchen appliances in the elderly user's kitchen scene within a preset time step, wherein the historical operation data is divided into natural days, and the operation time data of different kitchen appliances within a natural day are extracted to determine the operation time characteristics of different kitchen appliances; S204, performing soft clustering using the operating time characteristics of different kitchen appliances, obtaining clustering data sets for each time period according to the clustering results, extracting the data volume of different categories of kitchen appliances in the clustering data sets, and determining the interaction priority of the kitchen appliances relative to the elderly users in different time periods based on the data volume; S206, obtaining operation program features of different kitchen appliances according to the historical operation data, obtaining operation program feature sequences of different kitchen appliances through time series integration, and repeatedly dividing the operation program feature sequences of different kitchen appliances using different time length division standards; S208, for any kitchen equipment, embed the divided running program feature subsequences of different lengths into low-dimensional vectors, introduce a multi-head attention mechanism to weight each low-dimensional vector, and connect the weighted low-dimensional vectors to form a feature vector containing all low-dimensional vector features; S210, importing the feature vector into a multi-layer perceptron, and using a double-layer multi-layer perceptron structure to determine the elderly user's control preference data for kitchen equipment.

[0027] It should be noted that the operation status of kitchen equipment in different time periods is aggregated and analyzed based on the historical operation data of different kitchen equipment, and the usage and operation status of kitchen equipment is distinguished by the amount of data in different clusters. The higher the amount of data, the higher the usage rate of the kitchen equipment in the current time period. Based on the operation status, the kitchen equipment with the highest operation frequency is determined and used as the preferred kitchen equipment for the corresponding time period. Due to the degradation of the memory ability of elderly users, they often forget what to do next. In order to avoid this situation in the kitchen scene, when the IoT device detects that the elderly user enters the kitchen scene, the elderly user is reminded to interact according to the interaction priority of the kitchen equipment in the current time period.

[0028] For different kitchen equipment, elderly users usually have their own usage habits. The commonly used functional programs of elderly users are obtained according to the historical operation data of different kitchen equipment. The operation program features are obtained through the operation data corresponding to the functional programs. The operation program feature sequences of different kitchen equipment are obtained after time series integration. The operation program feature sequences are repeatedly divided according to the preset division length ratio, such as 1 / 5, 1 / 3, 1 / 2, and 1. The operation program feature sequences of four different lengths are obtained. The operation program feature sequences of different lengths contain the short-term and long-term preferences of elderly users. The four different lengths of the operation program feature sequences are first embedded into the low-dimensional vector space using the embedding layer, and the multi-head attention mechanism is introduced to calculate the obtained low-dimensional vector using the scaled dot product attention. Different linear transformations are performed in a single attention head to obtain the importance of different low-dimensional vectors. The parameters can be learned and calculated to ensure that relevant features are learned from different subspaces. The attention results output by all attention heads are input into the connection layer for merging and splicing to form a feature vector containing all features. After the MLP multi-layer perceptron, the control preference data of elderly users for different kitchen equipment is obtained. A multilayer perceptron is used to generate the usage representation of kitchen equipment by elderly users based on the feature sequences of the operating programs of different kitchen equipment. Multiple fully connected layers are used, and each fully connected layer performs feature transformation through linear transformation and nonlinear activation function. The short-term and long-term preferences of elderly users are integrated to determine the control preference data of elderly users for kitchen equipment.

[0029] It should be noted that user similarity is calculated based on the multimodal behavioral characteristics of elderly users, interaction priorities of kitchen equipment in different time periods, and control preference data, and users who meet the preset similarity standards are selected as similar users of the elderly users. Social graph data is constructed based on the elderly users and similar users; interaction graph data is constructed based on the kitchen equipment in the kitchen scenes of elderly users, and the elderly users and kitchen equipment are combined through the constructed interaction graph data, so that the user portrait can learn the strong correlation between the elderly users and the influence of different degrees of interaction between the elderly users and kitchen equipment during the training process.

[0030] The social graph data and interaction graph data are used to construct an elderly user graph, and the representations of the social graph data and interaction graph data are converted into triple forms to construct an elderly user graph, in which the elderly user is the head node, the relationship is the interaction attribute or the similarity attribute, and the tail node is the similar user or kitchen equipment. In the elderly user graph, the elderly user and the kitchen equipment are used as nodes, and directed edges are set according to whether there is an association between the nodes. If the elderly user A and the elderly user B are similar users, there is an edge structure between the elderly user A node and the elderly user B node. If the elderly user A interacts with both kitchen equipment C and D, there is an edge structure between the elderly user A node and the kitchen equipment C and D nodes. The multimodal behavior characteristics of the elderly user are added as additional attributes of the elderly user node, and the interaction priority and control preference data of the kitchen equipment in different time periods are added as additional attributes of the kitchen equipment node.

[0031] The graph neural network is used to learn the graph data of the elderly user graph. In the process of aggregation, the neighborhood nodes of the elderly user nodes are selected by random sampling of the neighborhood to reduce the magnitude of the calculated data during aggregation. A preset number of neighborhood nodes are randomly selected for aggregation. Different neighborhood nodes are randomly selected again in each iteration. The nodes ensure that most of the neighboring nodes participate in the neighborhood aggregation, and the random sampling of the neighborhood can solve the overfitting problem of the graph neural network to a certain extent. The multi-head attention mechanism is introduced to weight each neighborhood node, and the weighted neighborhood nodes are used to update the aggregation. When the information of adjacent users or kitchen equipment is aggregated, the weights of the kitchen equipment and similar users with preferences are set according to the preference control data and similarity of the elderly users, and the corresponding neighbor nodes are emphasized, and the information of the above neighbor nodes is retained. The elderly user nodes after aggregation update are obtained, and nonlinear activation functions are used for calculation to obtain the updated embedding representation of the elderly user nodes. After multiple aggregations, the updated embedding representation of the elderly user nodes is spliced ​​with the original embedding representation to strengthen the weight of the elderly user's own attributes in the user portrait, and finally the user portrait of the elderly user is output through the output layer of the graph neural network.

[0032] Figure 3 A flow chart of an embodiment for determining the optimal interaction mode for elderly users through user portraits is shown.

[0033] According to an embodiment of the present invention, the optimal interaction mode is determined by the user portrait, specifically: S302, obtaining an example of kitchen equipment aging-friendly interactive design, performing graph processing on the example of kitchen equipment aging-friendly interactive design to generate graph structure data, encoding the graph structure data, and obtaining a user portrait in each example; S304, obtaining different user portrait groups through clustering, and obtaining a user evaluation matrix for interaction modalities in the user portrait groups, where the elements in the evaluation matrix represent the user's evaluation of the interaction modality; S306, using the user portrait of the elderly user to determine a corresponding user portrait group, performing similarity calculation within the user portrait group to select a preset number of similar users, and selecting the interaction mode with the highest evaluation according to the evaluation matrix of the similar users; S308, simulating interaction prompts of different kitchen appliances according to the selected interaction mode, obtaining feedback from the elderly user on the interaction prompts of different interaction modes through the terminal device, and selecting the best interaction mode according to the feedback.

[0034] It should be noted that, an example of kitchen equipment aging-friendly interaction design is obtained, and aging-friendly measures set for elderly users in the example are analyzed and extracted, such as setting button tactile modal feedback and auditory modal guidance for elderly users with decreased vision, etc., and the association between the elderly users and the aging-friendly interaction modality is generated according to the aging-friendly measures set for the elderly users in the example, and the corresponding graph processing is performed using the association, and each instance corresponds to a unique graph structure, and the graph structure is associated and spliced ​​through the similarity association between the nodes to complete the graph processing of the kitchen equipment aging-friendly interaction design example. According to the graph processing, the kitchen equipment aging-friendly interaction design example is constructed into a corresponding graph structure, which contains user nodes and interaction modality nodes. Based on the attribute information of the corresponding user portrait of the elderly user, the portrait model trained by the graph neural network is used to embed the user nodes and interaction modality nodes in the graph structure, and the user portrait in each instance is obtained. The interaction modality nodes are sorted by clustering and recall algorithms, wherein the user's evaluation of the interaction modality in the example is obtained according to the user operation fluency obtained by feedback, and the higher the operation fluency, the better the corresponding evaluation. The interaction mode is selected based on the highest evaluation of similar users, and the operation interaction prompt is virtually generated on the preset terminal device. Feedback is obtained to obtain the interaction fluency of the elderly user and the target capture degree to determine the optimal interaction mode. The optimal interaction mode can be one or more, determined based on the feedback from the elderly user.

[0035] It should be noted that the corresponding function program information is determined based on the control preference data of the elderly user, and the operation flow of the function program information of different kitchen equipment is read, such as rotating the lid to open the lid, taking out the inner pot, putting the inner pot back, screwing the lid back, turning on the power, pressing the start button, pressing the function button, selecting the subsequent subdivided function, setting the time, and starting the function. The operation flow is set to the corresponding interactive guidance through the optimal interactive mode. For example, voice broadcast, icon process guidance, selection error vibration feedback, etc. are used to construct an interactive control guidance workflow. Preferably, based on the function program operation flow of the elderly user's control preference data, the elderly user's dietary habits are obtained, and the optimal operation flow of the corresponding function program is obtained based on the dietary habits and a large number of historical cooking examples. The deviation between the optimal operation flow and the elderly user's habitual operation flow is obtained. When the deviation is greater than the preset deviation threshold, the modification coefficient is set to make a slight adjustment to the user's habitual operation flow, and the interactive control guidance workflow is generated according to the adjusted operation flow, and the elderly user's feedback information on cooking food is obtained. The modification coefficient is reset according to the feedback information to improve the taste of the cooked food.

[0036] When an elderly user enters the kitchen scene, an operation interaction prompt is generated according to the interaction priority of kitchen equipment in different time periods, and it is judged whether the kitchen equipment operated by the elderly user meets the kitchen equipment interaction priority, and the frequency of non-satisfaction is counted. When the frequency of non-satisfaction is greater than a preset threshold, the kitchen equipment interaction priority of the current time period is reset; the kitchen equipment operated by the elderly user is obtained, and the demand information is estimated according to the operation information of the elderly user. The corresponding interaction control guidance workflow is determined according to the selection feedback of the elderly user on the estimated demand information, and the average reaction time of the elderly user to each operation control step in the interaction control guidance workflow is obtained. When the average reaction time is greater than the preset time threshold, it proves that the elderly user does not understand the interaction guidance, and it is marked as critical. control steps; based on the key control steps, interactive prompt information of different interactive modes is generated, multi-channel prompts are given to the elderly users, and the average reaction time of the elderly users to the key control steps is statistically analyzed. When the average reaction time is less than, the interactive prompt information of different interactive modes is revoked; when the reaction time of the elderly user to the operation control step is less than the average reaction time, it proves that the elderly user is sufficiently proficient in operating a certain step of the functional program, and the corresponding guidance in the interactive control guidance workflow is removed, and the interactive control guidance workflow is updated according to the subsequent feedback from the elderly users. By dynamically updating the interactive control guidance workflow, the memory ability of the elderly users can be trained to a certain extent, and the acceptance of the elderly users to smart kitchen equipment can be improved.

[0037] A second embodiment of the present invention provides an intelligent interactive control system for aging-friendly kitchen equipment, including a user multimodal behavior feature evaluation unit, an equipment operation data analysis unit, a user portrait construction unit, an interactive control guide determination unit, and a control feedback unit; The user multimodal behavior characteristic evaluation unit evaluates the behavior characteristics of the elderly user in terms of hearing, vision, touch and cognition, and integrates the behavior characteristics in different aspects to generate the multimodal behavior characteristics of the elderly user; The device operation data analysis unit extracts operation time characteristics and operation program characteristics using historical operation data of different kitchen devices, determines the interaction priority of kitchen devices in different time periods according to the operation time characteristics, and determines the control preference data of the elderly user according to the operation program characteristics of different kitchen devices; The user portrait construction unit constructs a user portrait according to multimodal behavior characteristics, kitchen equipment interaction priorities at different time periods, and control preference data; The interactive control guidance determination unit determines the best interactive mode and the interactive control guidance workflow that meets the control preference of different kitchen equipment through the user portrait, extracts the key control steps of the interactive control guidance workflow, generates interactive prompt information of the key control steps based on the best interactive mode, and performs multi-channel interactive prompts; The control feedback unit obtains the elderly user's selection feedback on the interactive control guidance workflow, as well as the reaction time to each operation control step of the kitchen equipment under the interactive control guidance workflow, and analyzes and updates the interactive prompt information and the interactive control guidance workflow.

[0038] The third embodiment of the present invention provides a computer-readable storage medium, which includes an intelligent interactive control method program for aging-friendly kitchen equipment. When the intelligent interactive control method program for aging-friendly kitchen equipment is executed by a processor, the steps of the intelligent interactive control method for aging-friendly kitchen equipment are implemented.

[0039] In the several embodiments provided in the present application, it should be understood that the disclosed method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the units can be electrical, mechanical or other forms.

[0040] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0041] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0042] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0043] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An intelligent interactive control method for aging-friendly kitchen equipment, characterized in that: The following steps are involved: Evaluate the behavioral characteristics of elderly users in terms of vision, hearing, touch and cognition, and integrate the behavioral characteristics of different aspects to generate multimodal behavioral characteristics of elderly users; Extracting operation time characteristics and operation program characteristics using historical operation data of different kitchen appliances, determining the interaction priority of kitchen appliances in different time periods according to the operation time characteristics, and determining the control preference data of elderly users according to the operation program characteristics of different kitchen appliances; Building a user profile based on the multimodal behavior characteristics, the interaction priorities of kitchen equipment at different time periods, and the control preference data, and determining the optimal interaction mode through the user profile; Based on the optimal interaction mode, the interactive control guidance workflow of different kitchen equipment that meets the control preferences of elderly users is obtained, the key control steps in the interactive control guidance workflow are extracted, and the interactive prompt information of the key control steps is generated based on the interactive control guidance workflow, and multi-channel interactive prompts are provided to elderly users.

2. The intelligent interactive control method for aging-friendly kitchen equipment according to claim 1, characterized in that: Evaluate the behavioral characteristics of elderly users in hearing, vision, touch and cognition, and integrate the behavioral characteristics of different aspects to generate multimodal behavioral characteristics of elderly users, specifically: Obtaining program operation data of different types of kitchen equipment, extracting icon keyword data, guidance prompt sound data, touch vibration data and icon position data using the program operation data, and establishing behavioral characteristic evaluation items in terms of hearing, vision, touch and cognition based on the extracted multi-dimensional data; Sending the behavior characteristic evaluation project to a preset terminal device, and using the preset terminal device to obtain the visual behavior characteristics, auditory behavior characteristics, tactile behavior characteristics and cognitive behavior characteristics of the elderly user; The benchmark value of each modal behavior feature is set according to the mean of the evaluation results of different modal behavior features of young users, and the metric function is constructed using the Euclidean distance. The Euclidean distance between the visual behavior features, auditory behavior features, tactile behavior features and cognitive behavior features of the elderly users and the corresponding benchmark values ​​is obtained through the metric function. Normalization processing is performed according to the Euclidean distance to generate feature scores of different modal behavior characteristics of elderly users, the feature scores are matched with corresponding modal behavior characteristics, and all modal behavior characteristics of elderly users are integrated to generate multimodal behavior characteristics.

3. The intelligent interactive control method for aging-friendly kitchen equipment according to claim 1, characterized in that: The interaction priorities of the kitchen devices in different time periods are determined according to the operating time characteristics, and the control preference data of the elderly users are determined according to the operating program characteristics of different kitchen devices, specifically: Extracting historical operation data of different kitchen appliances in the elderly user's kitchen scene within a preset time step, wherein the historical operation data is divided into natural days, and the operation time data of different kitchen appliances within a natural day are extracted to determine the operation time characteristics of different kitchen appliances; Soft clustering is performed using the operating time characteristics of different kitchen appliances, and clustering data sets of each time period are obtained according to the clustering results. The data volume of different categories of kitchen appliances in the clustering data sets is extracted, and the interaction priority of kitchen appliances relative to elderly users in different time periods is determined based on the data volume. Acquire the operation program features of different kitchen appliances according to the historical operation data, acquire the operation program feature sequences of different kitchen appliances through time series integration, and repeatedly divide the operation program feature sequences of different kitchen appliances using different time length division standards; For any kitchen equipment, the running program feature subsequences of different lengths obtained by division are embedded into low-dimensional vectors, and a multi-head attention mechanism is introduced to weight each low-dimensional vector. The weighted low-dimensional vectors are connected to form a feature vector containing all low-dimensional vector features. The feature vector is introduced into a multi-layer perceptron, and a double-layer multi-layer perceptron structure is used to determine the control preference data of the elderly user for the kitchen equipment.

4. The intelligent interactive control method for aging-friendly kitchen equipment according to claim 1, characterized in that: A user profile is constructed based on the multimodal behavior characteristics, the interaction priorities of kitchen equipment at different time periods, and the control preference data, specifically: Calculate user similarity based on multimodal behavior characteristics of elderly users, interaction priorities of kitchen equipment at different time periods, and control preference data, select users who meet the preset similarity criteria as similar users of elderly users, and construct social graph data based on elderly users and similar users; Constructing interaction graph data based on kitchen equipment in the kitchen scene of elderly users, constructing an elderly user graph using the interaction graph data of the social graph data set, taking elderly users and kitchen equipment as nodes in the elderly user graph, setting directed edges based on whether there is a relationship between nodes, adding multimodal behavior characteristics of elderly users as additional attributes of elderly user nodes, and adding kitchen equipment interaction priorities and control preference data in different time periods as additional attributes of kitchen equipment nodes; Use graph neural network to learn the graph data of elderly user graph, use neighborhood random sampling to select the neighborhood nodes of elderly user nodes, introduce multi-head attention mechanism to weight each neighborhood node, and use the weighted neighborhood node pairs for aggregation update; Obtain the elderly user node after aggregation and update, obtain the updated embedding representation of the elderly user node, and after multiple aggregations, concatenate the updated embedding representation of the elderly user node with the original embedding representation to output the user portrait of the elderly user.

5. The intelligent interactive control method for aging-friendly kitchen equipment according to claim 1, characterized in that: The best interaction mode is determined through the user portrait, specifically: Obtaining examples of kitchen equipment aging-friendly interactive design, performing graphic processing on the examples of kitchen equipment aging-friendly interactive design to generate graph structure data, encoding the graph structure data, and obtaining user portraits in each example; Obtain different user portrait groups through clustering, and obtain a user evaluation matrix for interaction modalities within the user portrait groups, where the elements in the evaluation matrix represent the user's evaluation of the interaction modality; Determine the corresponding user portrait group using the user portrait of the elderly user, perform similarity calculation within the user portrait group to select a preset number of similar users, and select the interaction mode with the highest evaluation based on the evaluation matrix of the similar users; According to the selected interaction mode, the interaction prompts of different kitchen equipment are simulated, the feedback of elderly users on the interaction prompts of different interaction modes is obtained through the terminal device, and the best interaction mode is selected according to the feedback.

6. The intelligent interactive control method for aging-friendly kitchen equipment according to claim 1, characterized in that: Based on the optimal interaction mode, the interactive control guidance workflow of different kitchen equipment that meets the control preferences of elderly users is obtained, the key control steps in the interactive control guidance workflow are extracted, and the interactive prompt information of the key control steps is generated based on the interactive control guidance workflow, specifically: Determine corresponding function program information based on the control preference data of the elderly user, read the operation process of the function program information of different kitchen equipment, and construct an interactive control guidance workflow through the optimal interaction mode; When an elderly user enters the kitchen scene, an operation interaction prompt is generated according to the kitchen equipment interaction priority in different time periods, and it is determined whether the kitchen equipment operated by the elderly user meets the kitchen equipment interaction priority. The frequency of non-satisfaction is counted. When the non-satisfaction frequency is greater than a preset threshold, the kitchen equipment interaction priority of the current time period is reset; Obtain the kitchen equipment operated by the elderly user, determine the corresponding interactive control guidance workflow according to the operation information of the elderly user, obtain the average reaction time of the elderly user for each operation control step in the interactive control guidance workflow, and mark it as a key control step when the average reaction time is greater than a preset time threshold; Generate interactive prompt information of different interactive modes based on the key control steps, perform multi-channel prompts on elderly users, statistically analyze the average reaction time of elderly users to the key control steps, and cancel the interactive prompt information of different interactive modes when the average reaction time is less than . When the elderly user's reaction time to the operation control step is less than the average reaction time, the corresponding guidance in the interactive control guidance workflow is removed, and the interactive control guidance workflow is updated according to the subsequent feedback from the elderly user.

7. An intelligent interactive control system for aging-friendly kitchen equipment, characterized in that: Implementing the intelligent interactive control method of the aging-friendly kitchen equipment as described in any one of claims 1 to 6, comprising a user multimodal behavior feature evaluation unit, an equipment operation data analysis unit, a user portrait construction unit, an interactive control guidance determination unit, and a control feedback unit; The user multimodal behavior characteristic evaluation unit evaluates the behavior characteristics of the elderly user in terms of hearing, vision, touch and cognition, and integrates the behavior characteristics in different aspects to generate the multimodal behavior characteristics of the elderly user; The device operation data analysis unit extracts operation time characteristics and operation program characteristics using historical operation data of different kitchen devices, determines the interaction priority of kitchen devices in different time periods according to the operation time characteristics, and determines the control preference data of the elderly user according to the operation program characteristics of different kitchen devices; The user portrait construction unit constructs a user portrait according to multimodal behavior characteristics, kitchen equipment interaction priorities at different time periods, and control preference data; The interactive control guidance determination unit determines the best interactive mode and the interactive control guidance workflow that meets the control preference of different kitchen equipment through the user portrait, extracts the key control steps of the interactive control guidance workflow, generates interactive prompt information of the key control steps based on the best interactive mode, and performs multi-channel interactive prompts; The control feedback unit obtains the elderly user's selection feedback on the interactive control guidance workflow, as well as the reaction time to each operation control step of the kitchen equipment under the interactive control guidance workflow, and analyzes and updates the interactive prompt information and the interactive control guidance workflow.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a smart interactive control method program for aging-friendly kitchen equipment. When the smart interactive control method program for aging-friendly kitchen equipment is executed by a processor, the steps of the smart interactive control method for aging-friendly kitchen equipment according to any one of claims 1 to 6 are implemented.

Images