Robot self-adaptive emotion feedback system based on multi-mode intelligent perception

Through a robot adaptive emotional feedback system combining multimodal intelligent perception and personality type framework, the problem of insufficient accuracy of emotion recognition in the existing technology is solved, accurate and personalized emotional feedback is achieved, and user interaction satisfaction is improved. Emotional experience.

CN118990481APending Publication Date: 2024-11-22LIAONING NEW DIMENSION HUMANOID ROBOT RESEARCH INSTITUTE CO LTD
View PDF 0 Cites 6 Cited by

Patent Information

Application Number
CN202411147633.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The existing emotional feedback system relies on a single or limited perceptual data source, resulting in insufficient accuracy of emotional recognition and feedback, unable to fully capture the user's true emotional state, lack of integration and in-depth analysis of multimodal data, affecting personalization and accuracy. Spend.

Method used

Adopting robot adaptive emotional feedback system based on multimodal intelligent perception, the multimodal perceptual data acquisition module (visual, auditory, tactile, physiological signal) combined with personality type framework is used to identify emotional categories and perform emotions in different virtual situations. Adjustment and change analysis, combined with natural language processing technology to explore deep emotional experience and adaptive assimilation of contradictions, and generate personalized emotional feedback.

Benefits of technology

Accurate emotional recognition and personalized emotional feedback are achieved, user interaction satisfaction and emotional experience are improved, and the continuity and adaptability of emotional support are enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118990481A_ABST
    Figure CN118990481A_ABST
Patent Text Reader

Abstract

The invention provides a robot adaptive emotion feedback system based on multi-modal intelligent perception, and relates to the technical field of intelligent feedback, and the system comprises a multi-modal perception data obtaining module which is used for obtaining multi-modal perception data of a target user; the emotion category recognition module is used for recognizing emotion categories; the virtual situation building module is used for building a virtual situation; the first emotion feedback set obtaining module is used for obtaining a first emotion feedback set; the second emotion feedback set obtaining module is used for obtaining a second emotion feedback set; the third emotion feedback set obtaining module is used for obtaining a third emotion feedback set; and the adaptive emotion feedback result acquisition module is used for acquiring an adaptive emotion feedback result. The technical problem of insufficient emotion recognition and feedback accuracy in the prior art can be solved, the technical targets of accurate emotion recognition and personalized emotion feedback are achieved, and the technical effect of improving user interaction satisfaction and emotion experience is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent feedback technology, and in particular to a robot adaptive emotion feedback system based on multimodal intelligent perception. Background Art

[0002] Emotional feedback systems are applied in the field of human-computer interaction. They improve the naturalness of interaction and user experience by identifying and responding to the user's emotional state.

[0003] At present, existing emotional feedback systems mainly rely on limited perceptual data sources, usually including visual, auditory or text analysis. There are significant deficiencies in the accuracy of emotion recognition and feedback, and it is difficult to fully capture the user's true emotional state. In addition, the existing technology lacks the integration and in-depth analysis of multimodal data, and cannot effectively handle the complex changes in user emotions, resulting in low personalization and accuracy of emotional feedback. Due to the lack of mining the user's long-term emotional history and deep emotional experience, the existing system also has obvious limitations in the sustainability and adaptability of emotional support.

[0004] In summary, the existing technology has technical problems in that the accuracy of emotion recognition and feedback is insufficient due to reliance on a single or limited perceptual data source, which further affects the personalization and accuracy of the emotion feedback system. Summary of the invention

[0005] The purpose of this application is to provide a robot adaptive emotion feedback system based on multimodal intelligent perception, so as to solve the technical problems in the prior art that the accuracy of emotion recognition and feedback is insufficient due to reliance on a single or limited perception data source, further affecting the personalization and accuracy of the emotion feedback system.

[0006] In view of the above problems, the present application provides a robot adaptive emotion feedback system based on multimodal intelligent perception.

[0007] The present application provides a robot adaptive emotion feedback system based on multimodal intelligent perception, including: a multimodal perception data acquisition module, used to acquire multimodal perception data of a target user through the multimodal perception data acquisition module, wherein the multimodal perception data acquisition module includes a visual perception device, an auditory perception device, a tactile perception device, and a physiological signal perception device; an emotion category recognition module, used to identify emotion categories based on the multimodal perception data, using an initial emotion recognition model corresponding to a personality type framework associated with pre-training, wherein the emotion categories include positive emotion indicators, neutral emotion indicators, and reverse emotion indicators; a virtual situation building module, used to build a first virtual situation based on the multimodal perception data and the positive emotion indicators in the emotion category; to build a second virtual situation based on the multimodal perception data and the neutral emotion indicators in the emotion category; and to build a third virtual situation based on the multimodal perception data and the reverse emotion indicators in the emotion category; a first emotion feedback set acquisition module, used to, in the first virtual situation, combine the personality type defined by the personality type framework of the target user, through The module comprises a module for obtaining a second emotional feedback set, which is used to perform emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in combination with the personality type defined by the personality type framework of the target user in the second virtual situation, and obtain the second emotional feedback set; a module for obtaining a third emotional feedback set, which is used to perform emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in combination with the personality type defined by the personality type framework of the target user in the third virtual situation, and obtain the third emotional feedback set; an adaptive emotional feedback result acquisition module, which is used to perform deep emotional experience mining on the user's language use, expression mode and emotional vocabulary based on the historical conversation content of the target user through natural language processing technology, and obtain deep emotional indicators, which reflect the user's deep and lasting emotional experience, and perform contradictory adaptive assimilation on the first emotional feedback set, the second emotional feedback set and the third emotional feedback set to obtain adaptive emotional feedback results.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] A multimodal perception data acquisition module is used to acquire the multimodal perception data of the target user through the multimodal perception data acquisition module, and the multimodal perception data acquisition module includes a visual perception device, an auditory perception device, a tactile perception device, and a physiological signal perception device; an emotion category recognition module is used to identify the emotion category based on the multimodal perception data and the initial emotion recognition model corresponding to the pre-training associated with the personality type framework, and the emotion category includes a positive emotion indicator, a neutral emotion indicator, and a reverse emotion indicator; a virtual situation construction module is used to build a first virtual situation based on the multimodal perception data and the positive emotion indicator in the emotion category; to build a second virtual situation based on the multimodal perception data and the neutral emotion indicator in the emotion category; to build a third virtual situation based on the multimodal perception data and the reverse emotion indicator in the emotion category; a first emotion feedback set acquisition module is used to, in the first virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the first emotion feedback A second emotion feedback set obtaining module is used to, in the second virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the second emotion feedback set; a third emotion feedback set obtaining module is used to, in the third virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the third emotion feedback set; an adaptive emotion feedback result obtaining module is used to, based on the historical conversation content of the target user, use natural language processing technology to perform deep emotion experience mining on the user's language use, expression mode and emotional vocabulary, obtain deep emotion indicators, and the deep emotion indicators reflect the user's deep and lasting emotional experience, and perform contradictory adaptive assimilation on the first emotion feedback set, the second emotion feedback set, and the third emotion feedback set to obtain adaptive emotion feedback results, thereby achieving the technical goals of accurate emotion recognition and personalized emotion feedback, and achieving the technical effect of improving user interaction satisfaction and emotional experience.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 This is a schematic diagram of the structure of the robot adaptive emotion feedback system based on multimodal intelligent perception in this application;

[0013] Figure 2 This is a flow chart of determining multiple virtual adaptation scenarios in a robot adaptive emotion feedback system based on multimodal intelligent perception in this application.

[0014] Description of reference numerals:

[0015] 11 multimodal perception data acquisition module, 12 emotion category recognition module, 13 virtual situation building module, 14 first emotion feedback set acquisition module, 15 second emotion feedback set acquisition module, 16 third emotion feedback set acquisition module, 17 adaptive emotion feedback result acquisition module. DETAILED DESCRIPTION

[0016] This application provides a robot adaptive emotion feedback system based on multimodal intelligent perception, which solves the technical problem in the prior art that the accuracy of emotion recognition and feedback is insufficient due to reliance on a single or limited perception data source, further affecting the personalization and accuracy of the emotion feedback system. The technical goal of accurate emotion recognition and personalized emotion feedback is achieved, and the technical effect of improving user interaction satisfaction and emotional experience is achieved.

[0017] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0018] For examples, please see the attached Figure 1 The present application provides a robot adaptive emotion feedback system based on multimodal intelligent perception, which specifically includes the following steps:

[0019] The multimodal perception data acquisition module 11 is used to acquire the multimodal perception data of the target user through the multimodal perception data acquisition module, and the multimodal perception data acquisition module includes a visual perception device, an auditory perception device, a tactile perception device, and a physiological signal perception device.

[0020] Specifically, the multimodal perception data acquisition module refers to the various perception devices built into the robot, including visual perception devices, auditory perception devices, tactile perception devices, and physiological signal perception devices. Visual perception devices are mainly used to capture the user's facial expressions and body movements, auditory perception devices analyze the emotional state by picking up the user's voice and tone, tactile perception devices determine the user's emotional response by detecting the user's contact and interaction methods, and physiological signal perception devices provide more in-depth emotional feedback data by monitoring physiological parameters such as heart rate and skin electrical response. The integration and analysis of perception data enables the robot to understand and respond to the user's emotional needs more accurately. The multimodal perception data acquisition module integrates the perception data of vision, hearing, touch and physiological signals to form a comprehensive and accurate emotion recognition system, which enables the robot to better understand and adapt to the user's emotional needs, thereby providing more intelligent and humane feedback.

[0021] The emotion category recognition module 12 is used to identify emotion categories based on the multimodal perception data using the initial emotion recognition model corresponding to the personality type framework association pre-training, and the emotion categories include positive emotion indicators, neutral emotion indicators, and negative emotion indicators.

[0022] Specifically, multimodal perception data includes visual, auditory, tactile and physiological signals. The user's emotional state can be reflected by comprehensive analysis of multimodal perception data. The personality type framework refers to the classification of different personality types of users by analyzing the user's behavior, habits and personality characteristics, which can help the emotion recognition model better understand and predict the user's emotional response. For example, the personality type framework includes the result types in the MBTI sixteen personality test. The pre-trained initial emotion recognition model is associated through the personality type framework to improve the accuracy and adaptability of emotion recognition. The pre-trained model has been trained on a large amount of emotion data and has a certain emotion recognition ability. By combining the personality type framework, the parameters of the model can be further refined and adjusted to make it more consistent with the emotional characteristics and reaction patterns of specific users, and the user's emotions can be more accurately identified and classified. The identified emotion categories include positive emotion indicators, neutral emotion indicators and negative emotion indicators. Positive emotion indicators refer to the user being in a positive emotional state, such as happiness, excitement, satisfaction, etc. Positive emotion indicators show the user's good psychological state. Neutral emotion indicators refer to the user's emotional state being relatively stable, without obvious positive or negative tendencies, such as calmness and calmness. The negative emotion indicator refers to the user being in a negative emotional state, such as sadness, anger, anxiety, etc. The negative emotion indicator shows the user's negative psychological state.

[0023] Through multimodal perception data and personality type framework, users' emotion categories can be effectively identified, providing targeted emotional feedback and interactive experience. Positive emotion indicators help the system encourage users to maintain positive emotions, neutral emotion indicators help the system provide neutral and stable emotional support, and negative emotion indicators help the system identify users' negative emotions and provide comfort and help in a timely manner. It can better understand users' emotional states and provide more personalized and humane emotional services.

[0024] The virtual situation building module 13 is used to build a first virtual situation based on the multimodal perception data and the positive emotion indicators in the emotion category; to build a second virtual situation based on the multimodal perception data and the neutral emotion indicators in the emotion category; and to build a third virtual situation based on the multimodal perception data and the negative emotion indicators in the emotion category.

[0025] Specifically, based on the multimodal perception data, a first virtual scenario can be built with positive emotion indicators in the emotion category. Positive emotion indicators reflect that the user is in a positive emotional state, such as happiness, excitement, or satisfaction. According to the user's multimodal perception data, such as pleasant facial expressions, cheerful tones, or stable physiological signals, a virtual scenario matching the emotion is designed, including pleasant activity scenes, cheerful music, or bright visual effects, to enhance and continue the user's positive emotional experience.

[0026] Based on the multimodal perception data and the neutral emotion indicators in the emotion category, a second virtual scenario is built. The neutral emotion indicator refers to the user's relatively stable emotional state, without obvious positive or negative tendencies. According to the user's multimodal perception data, such as calm facial expressions, calm tones or neutral physiological signals, a matching virtual scenario is designed, including a quiet environment, relaxing background music or soft visual effects, aiming to maintain the user's neutral emotional state and make him feel comfortable and relaxed.

[0027] Based on the multimodal perception data and the reverse emotion indicators in the emotion category, a third virtual scenario is built. The reverse emotion indicators reflect that the user is in a negative emotional state, such as sadness, anger, or anxiety. Based on the user's multimodal perception data, such as sad facial expressions, low voice tones, or unstable physiological signals, a virtual scenario is designed, including a soothing environment, comforting language, or warm visual effects, with the aim of helping users relieve negative emotions and provide support and comfort.

[0028] Through the design and application of scenarios, we can more effectively respond to users' emotional needs and provide personalized emotional support and interactive experience. Each virtual scenario is customized according to the user's multimodal perception data and emotional category, ensuring that users can get the most appropriate emotional feedback and experience in different emotional states.

[0029] The first emotion feedback set obtaining module 14 is used to obtain the first emotion feedback set by performing emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in the first virtual scenario in combination with the personality type defined by the personality type framework of the target user.

[0030] Specifically, in the first virtual situation that enhances the user's positive emotional experience, combined with the personality type defined by the personality type framework of the target user, the user's emotional change trend is captured by analyzing the user's multimodal perception data in the first virtual situation, and emotional regulation and emotional change analysis are performed. Emotional regulation refers to influencing and guiding the user's emotional direction by adjusting various elements of the first virtual situation, such as environmental atmosphere, interaction methods, and emotional feedback mechanisms. In a positive emotional state, a more pleasant situation setting is adopted to help users maintain a positive emotional state. Through meticulous emotional regulation, users are provided with an interactive experience that better fits their emotional needs.

[0031] Through the analysis of emotional changes, the user's emotional stability period, change nodes and potential emotional fluctuations are identified, thereby providing data support for subsequent emotional feedback and regulation. Combined with the emotional change analysis of the personality type framework, the emotional feedback is more personalized and accurate. Through the change trend analysis of multimodal perception data, the first emotional feedback set is obtained. The first emotional feedback set includes the user's emotional response, emotional change trajectory and emotional regulation suggestions in the first virtual situation, which is used to optimize the subsequent virtual situation and emotional support strategy.

[0032] The second emotion feedback set obtaining module 15 is used to obtain the second emotion feedback set by combining the personality type defined by the personality type framework of the target user in the second virtual situation and performing emotion regulation and emotion change analysis through the change trend of the multimodal perception data.

[0033] Specifically, in the second virtual situation that enhances the user's neutral emotional experience, combined with the personality type defined by the personality type framework of the target user, by analyzing the user's multimodal perception data in the second virtual situation, the user's emotional change trend is captured, and emotional regulation and emotional change analysis are performed. Emotional regulation refers to influencing and guiding the user's emotional direction by adjusting various elements of the second virtual situation, such as environmental atmosphere, interaction methods, and emotional feedback mechanisms. In a neutral emotional state, a calmer and more soothing situation setting is used to help users maintain a stable emotional state, or guide users to shift to positive emotions as needed. Through meticulous emotional regulation, users are provided with an interactive experience that better fits their emotional needs.

[0034] Through the analysis of emotional changes, the user's emotional stability period, change nodes and potential emotional fluctuations are identified, thereby providing data support for subsequent emotional feedback and regulation. Combined with the emotional change analysis of the personality type framework, the emotional feedback is more personalized and accurate. Through the change trend analysis of multimodal perception data, the second emotional feedback set is obtained. The second emotional feedback set includes the user's emotional response, emotional change trajectory and emotional regulation suggestions in the second virtual situation, which is used to optimize the subsequent virtual situation and emotional support strategy.

[0035] The third emotion feedback set obtaining module 16 is used to perform emotion regulation and emotion change analysis in the third virtual scenario in combination with the personality type defined by the personality type framework of the target user through the change trend of the multimodal perception data to obtain the third emotion feedback set.

[0036] Specifically, in the third virtual situation that enhances the user's reverse emotional experience, combined with the personality type defined by the personality type framework of the target user, by analyzing the user's multimodal perception data in the third virtual situation, the user's emotional change trend is captured, and emotional regulation and emotional change analysis are performed. Emotional regulation refers to influencing and guiding the user's emotional direction by adjusting various elements of the third virtual situation, such as environmental atmosphere, interaction methods, and emotional feedback mechanisms. In the reverse emotional state, a more comforting situation setting is adopted to help users maintain a warm emotional state, or guide users to change to positive emotions as needed. Through meticulous emotional regulation, users are provided with an interactive experience that better fits their emotional needs.

[0037] Through the analysis of emotional changes, the user's emotional stability period, change nodes and potential emotional fluctuations are identified, thereby providing data support for subsequent emotional feedback and regulation. The emotional change analysis combined with the personality type framework makes emotional feedback more personalized and accurate. Furthermore, there may be multiple emotional situations. The target user's emotional regulation is not achieved, but the target user's emotional changes are achieved, such as no longer being as sad as at the beginning after the passage of time, that is, being healed by time; the target user's emotional regulation is not achieved, and the target user's emotional changes are not achieved at the same time, such as not being healed by the passage of time, and then giving up on healing; the target user's emotional regulation is achieved, and the target user's emotional changes are achieved at the same time, such as regulating the emotions of the previous section through the emotional changes of the latter section; the target user's emotional regulation is achieved, but the target user's emotional changes are not achieved, etc.

[0038] The third emotional feedback set is obtained by analyzing the change trend of multimodal perception data. The third emotional feedback set includes the user's emotional response, emotional change trajectory and emotional regulation suggestions in the third virtual situation, so as to optimize the subsequent virtual situation and emotional support strategy.

[0039] The adaptive emotion feedback result acquisition module 17 is used to conduct deep emotion experience mining on the user's language usage, expression methods and emotional vocabulary based on the historical conversation content of the target user through natural language processing technology, obtain deep emotion indicators, and the deep emotion indicators reflect the user's deep and lasting emotional experience, and perform contradictory adaptive assimilation on the first emotion feedback set, the second emotion feedback set, and the third emotion feedback set to obtain adaptive emotion feedback results.

[0040] Specifically, based on the historical conversation content of the target user, deep emotional experience mining is carried out to obtain deep emotional indicators. The historical conversation content includes all conversation records of the user in the process of interacting with the robot, providing rich emotional information. Through natural language processing technology, the user's language usage, expression and emotional vocabulary are analyzed to identify the user's potential emotional state and deep emotional needs. Deep emotional experience mining can more comprehensively understand the user's emotional world and extract more refined and accurate deep emotional indicators. Deep emotions refer to deeper and more lasting emotional experiences, such as lasting sadness, long-term anxiety or deep happiness, which have a profound impact on the individual's mental health and happiness, and help provide more comprehensive emotional evaluation and support. For example, deep emotions include long-lasting emotions caused by childhood experiences or memories such as the death of parents in the original family.

[0041] After obtaining the deep emotion index, the first emotion feedback set, the second emotion feedback set and the third emotion feedback set are subjected to contradiction adaptive assimilation. The emotion feedback set is the user's emotional response and change data obtained in different virtual situations. Due to the contradictions and inconsistencies between different situations and emotion indicators, the feedback data is integrated through the adaptive assimilation algorithm. Contradiction adaptive assimilation is an intelligent processing method that can identify and resolve contradictions in the emotion feedback set, find consistency and commonality, and generate a comprehensive emotion feedback result. When performing contradiction adaptive assimilation, the user's deep emotion index and the data in each emotion feedback set are comprehensively considered to adjust and optimize the emotion feedback strategy. The adaptive assimilation process includes data fusion, conflict resolution and consistency correction, and then generates an adaptive emotion feedback result that not only meets the user's current emotional state but also takes into account the historical emotion trend, which can not only reflect the user's comprehensive emotional needs, but also provide a more reliable basis for emotional support. Based on the user's historical conversation content, deep emotional experience mining and adaptive assimilation of contradictions, adaptive emotional feedback results are obtained. The adaptive emotional feedback results contain the user's deep emotional needs and preferences, and can provide more accurate and personalized services in subsequent emotional support and interaction, which not only improves the user's emotional experience, but also enhances the system's emotional intelligence and user satisfaction.

[0042] The robot adaptive emotion feedback system based on multimodal intelligent perception achieves the technical goals of accurate emotion recognition and personalized emotion feedback, and achieves the technical effect of improving user interaction satisfaction and emotional experience.

[0043] Furthermore, the present application also includes:

[0044] Based on the changing trend of the multimodal perception data, the emotion change obstacle points are identified; the interest and attention information of the target user is collected, and with respect to the emotion change obstacle points, a plurality of virtual scenarios that attract the target user are created using a generative adversarial network, wherein the plurality of virtual scenarios include a first virtual scenario, a second virtual scenario, and a third virtual scenario; based on the plurality of virtual scenarios, the attention level of the target user and the progress of the emotion change process corresponding to the emotion change obstacle points are matched to determine a plurality of virtual adaptation scenarios.

[0045] Specifically, emotional change barriers refer to the difficulties or obstacles that users may encounter during the emotional transition process, which can be identified by analyzing the changing patterns of the user's multimodal perception data. Through comprehensive analysis of visual, auditory, tactile and physiological signals, the robot can accurately capture anomalies in the user's emotional changes and identify the links or moments where the user's emotional changes are stagnant or abnormal.

[0046] Then, by understanding the interests and concerns of the target users, we can better customize emotional feedback and situational experiences for them. Interest and concern information includes what things, activities or topics the user is interested in, which can be obtained through the user's daily interactions, browsing records and behavior data. After understanding the user's interest and concern information, we can design situational experiences for users in a more targeted manner to make them more in line with the user's preferences and needs.

[0047] Next, we use a generative adversarial network to create multiple virtual scenarios that attract target users, targeting the barriers to emotional change. The generative adversarial network is a deep learning algorithm that can generate high-quality and realistic virtual scenarios, including the first virtual scenario, the second virtual scenario, and the third virtual scenario, through adversarial training of the generative model and the discriminative model. Each scenario is designed based on the user's interest information and emotional change barriers. In this way, virtual scenarios can help users overcome obstacles in emotional change and promote a smooth transition of emotions.

[0048] Next, based on multiple virtual scenarios, the target user's attention and the progress of the emotion change process corresponding to the emotion change obstacle point are matched. Attention refers to the user's interest and participation in different virtual scenarios, and the progress of the emotion change process refers to the specific stage of the user's emotion change process. Through matching, the virtual scenario that best suits the user's current emotional state and needs can be screened out, which is called a virtual adaptation scenario. Virtual adaptation scenarios are the key to providing users with personalized emotional support and feedback, ensuring that each scenario can effectively help users deal with emotional changes.

[0049] Then, by accurately matching the user's interest information and emotional change process, multiple virtual adaptation scenarios are determined, which can provide support and guidance for the user's specific emotional needs, help the user overcome emotional change barriers, and promote the healthy development of emotions. Each adaptation scenario is carefully designed and optimized to ensure that users can get the best emotional experience and support in different emotional states.

[0050] By collecting user interest information, using generative adversarial networks to create multiple virtual scenarios, and matching the user's attention and emotional change process progress, multiple virtual adaptation scenarios are determined to provide accurate emotional support and interactive experience.

[0051] Further, if Figure 2 As shown, this application also includes:

[0052] Based on the physiological signal sensing device, the signal processing technology is used to analyze the physiological signal and extract the physiological signal features; based on the physiological signal features, the carrier form is associated with the attention of the target user to obtain a first associated feature, and the carrier form includes pictures, texts, and audios; based on the physiological signal features, the carrier form is associated with the progress of the emotion change process corresponding to the emotion change obstacle point to obtain a second associated feature; based on the carrier form, the multiple virtual situations are matched and filtered by comparing the first and second associated features to determine multiple virtual adaptation situations.

[0053] Specifically, physiological signal sensing devices can capture the user's heart rate, galvanic skin response, breathing rate and other physiological parameters. Through signal processing techniques such as filtering, spectrum analysis and feature extraction, meaningful features can be extracted from the original signal, which can reflect the user's emotional state and changes. For example, an increased heart rate may indicate excitement or anxiety, and an enhanced galvanic skin response may indicate tension or stress.

[0054] Then, the carrier forms include pictures, text and audio, and the carrier forms are different ways for users to receive information. By analyzing the user's attention to different carrier forms, it is determined in which form the user prefers to obtain information. For example, some users may prefer to obtain information through pictures, while other users may prefer text or audio. Among them, painters can use pictures of paintings to feedback emotions, writers can use the text of writing to feedback emotions, and singers can use the audio of singing to feedback emotions. Although pictures, text, and audio are carriers, different carriers are suitable for completing the process of emotional feedback. The first associated feature reflects the user's emotional feedback preference, so that virtual situations can be designed more targeted.

[0055] Next, the progress of the emotional change process reflects the specific stage of the user's emotional change process. Users may receive information differently at different stages. By associating the carrier form with the stage, it is possible to determine the more effective carrier form at different emotional stages. For example, in the early stage of emotional change, users may be more receptive to concise text information, while in the later stage of emotional change, users may prefer to get comfort and guidance through pictures or audio.

[0056] Next, based on the carrier form, the first correlation feature and the second correlation feature are compared to match and filter multiple virtual scenarios to determine multiple virtual adaptation scenarios. By comprehensively considering the user's attention and the progress of the emotional change process, the virtual scenario that best suits the user's current emotional state and information reception preference is selected, which can provide personalized and effective emotional support, help users better cope with emotional changes, and provide a more accurate interactive experience.

[0057] By analyzing physiological signals to extract features, associating the carrier form with the user's attention and emotional change process progress, matching and screening virtual situations, and determining multiple virtual adaptation situations, personalized and accurate emotional support and interactive experience can be provided.

[0058] Furthermore, the present application also includes:

[0059] Obtain user rating levels, and perform matrix decomposition with the first associated features as rows and the second associated features as columns to build a basic preference matrix for the target user; traverse the interest and attention information of the target user to build a first collaborative preference matrix, a second collaborative preference matrix, and up to an Mth collaborative preference matrix corresponding to the target user; perform collaborative filtering based on the basic preference matrix of the target user and the corresponding first collaborative preference matrix, second collaborative preference matrix, and up to an Mth collaborative preference matrix to determine multiple virtual adaptation scenarios.

[0060] Specifically, obtaining user ratings is the first step in providing personalized emotional support to users. Ratings are user feedback on different situations or content, obtained through user interaction, satisfaction surveys, or direct ratings. By analyzing the ratings, we can understand the user's preference for different situations, thereby providing data support for the subsequent preference matrix construction.

[0061] Then, we performed matrix decomposition with the first associated feature as the row and the second associated feature as the column, and built the basic preference matrix of the target user. The first associated feature reflects the user's attention to different carrier forms, and the second associated feature reflects the user's preference for carrier forms at different stages of emotional change. Through matrix decomposition, the user's overall preference can be decomposed into more specific feature preferences, thereby more accurately describing the user's emotional needs and information acceptance methods.

[0062] Next, the target user's interest information is accessed in turn to build multiple collaborative preference matrices. Each collaborative preference matrix corresponds to the user's preferences in different interest areas. For example, the first collaborative preference matrix may reflect the user's preferences in entertainment, the second collaborative preference matrix may reflect the user's preferences in learning, and so on, until the Mth collaborative preference matrix. The collaborative preference matrix provides specific preference data of users in different fields by segmenting and analyzing the user's interest information.

[0063] Next, collaborative filtering is performed based on the target user's basic preference matrix and multiple collaborative preference matrices to determine multiple virtual adaptation scenarios. Collaborative filtering is a recommendation algorithm that predicts the user's possible preferences for new scenarios by analyzing the user's historical behavior and preference data. By combining the basic preference matrix and multiple collaborative preference matrices, it is possible to more accurately recommend virtual scenarios that best suit the user's current emotional state and interests, thereby obtaining personalized emotional support and experiencing scenarios that are highly consistent with their interests, thereby improving overall satisfaction and emotional experience.

[0064] By obtaining user rating levels, building a basic preference matrix and multiple collaborative preference matrices, and performing collaborative filtering, multiple virtual adaptation scenarios are determined to provide accurate and personalized emotional support and interactive experience.

[0065] Furthermore, the present application also includes:

[0066] Based on the basic preference matrix of the target user, the multiple virtual scenarios are sorted by matching degree to obtain multiple virtual scenarios with priority ranking; based on the multiple virtual scenarios with priority ranking, the first collaborative preference matrix, the second collaborative preference matrix, and up to the Mth collaborative preference matrix corresponding to the target user are combined to perform iterative combination optimization to obtain multiple iterative combinations; the multiple iterative combinations correspond one-to-one to the multiple virtual adaptation scenarios, the virtual adaptation scenarios include combined scenarios formed by N virtual scenarios with optimal introduction timing and optimal duration, and the user engagement incentive strategies corresponding to the N virtual scenarios include passive incentive regulation and active incentive regulation.

[0067] Specifically, multiple virtual scenarios are ranked according to their matching degree based on the target user's basic preference matrix. The basic preference matrix reflects the user's preference for different carrier forms and stages of emotional change. Through the basic preference matrix, the matching degree between each virtual scenario and the user's preference can be calculated. Matching ranking can prioritize the virtual scenarios that best meet the user's needs, ensuring that the user gets the best emotional support experience.

[0068] Then, based on the multiple virtual scenarios ranked by priority, the first collaborative preference matrix, the second collaborative preference matrix, and finally the Mth collaborative preference matrix corresponding to the target user are combined for iterative combination optimization. The collaborative preference matrix reflects the user's preferences in different interest areas. By combining with the basic preference matrix, it can be iteratively optimized, and the combination of virtual scenarios can be continuously adjusted and optimized, so that the optimization process takes into account the user's overall interests and specific emotional needs, ensuring that the final recommended virtual scenario can meet the user's expectations to the greatest extent.

[0069] Then, multiple iterative combinations correspond to multiple virtual adaptation scenarios one by one, and each virtual adaptation scenario includes N virtual scenarios, forming a combination scenario with the optimal introduction timing and optimal duration. The result of iterative combination optimization ensures that each virtual scenario is introduced at the most appropriate time point and lasts for the most appropriate length of time, maximizing its emotional support effect, and thus providing continuous, progressive and refined emotional support, so that users can get the best interactive experience in different emotional states.

[0070] Through matching ranking and iterative combination optimization, multiple virtual adaptation scenarios are finally determined. Each scenario is introduced at the optimal time and lasts for the most suitable time, providing accurate and personalized emotional support.

[0071] Furthermore, the present application also includes:

[0072] Record interaction information corresponding to a first virtual situation, the interaction information including a situation start time point, a situation duration, and user response data, wherein the first virtual situation is any one of the N virtual situations; take entering the first virtual situation as a starting point of an emotion change, and take completion of the first virtual situation as an end point of the emotion change, track the emotion change, and set a first emotion change trajectory; based on the interaction information corresponding to the first virtual situation and the first emotion change trajectory, traverse the N virtual situations and perform random expansion optimization.

[0073] Specifically, recording the interaction information corresponding to the first virtual situation is an important step in tracking and optimizing user emotional feedback. The interaction information includes the situation start time, situation duration, and user response data. The situation start time refers to the specific moment when the user enters the virtual situation, and the situation duration records the user's stay in the situation. User response data includes various behaviors and feedback of users in virtual situations, such as changes in facial expressions, voice intonation, tactile reactions, and physiological signals, which can help understand the user's emotional reactions in different situations, thereby providing basic data for subsequent emotional support and optimization.

[0074] Then, the entry into the first virtual situation is taken as the starting point of the emotion change, and the completion of the first virtual situation is taken as the end point of the emotion change. The emotion change is tracked and the first emotion change trajectory is set. The emotion change trajectory records the user's emotional change process in the entire virtual situation. By analyzing the starting point and end point of the emotion change, the change path of the user's emotion can be depicted. The first emotion change trajectory refers to the specific emotional changes of the user in the first virtual situation, which can reflect the emotional ups and downs and changes of the user in a specific situation, and provide a reference for the refinement of emotional feedback.

[0075] Next, based on the interaction information and the first emotion change trajectory corresponding to the first virtual situation, N virtual situations are visited in sequence for random expansion optimization. Visiting N virtual situations in sequence means analyzing and screening all possible virtual situations, and finding a situation combination that better meets user needs through random expansion optimization. Random expansion optimization is an algorithm that finds the best emotional support solution by randomly generating and evaluating different situation combinations, and then takes into account the emotional changes of users in different situations to provide users with more accurate and personalized emotional support.

[0076] By recording interaction information, setting emotion change trajectories and random expansion optimization, we can fine-tune and optimize users' emotional responses in virtual situations, providing more accurate and personalized emotional support.

[0077] Furthermore, the present application also includes:

[0078] Based on the N virtual situations, N situational engagements are obtained; a user terminal device is connected, and after obtaining read permission, a user engagement incentive framework is built, wherein the user engagement incentive framework enhances the interaction of the target users with gamification elements, reward mechanisms, and social sharing functions; based on the user engagement incentive framework, the optimal introduction timing and optimal duration are configured in comparison with the N virtual situations and the N situational engagements.

[0079] Specifically, based on N virtual situations, N situation engagements are obtained. Situation engagement refers to the frequency and enthusiasm of user interactions in each virtual situation, which can be measured by the user's behavior and feedback in the situation. For example, the user's stay time, number of interactions, and emotional response in a situation can all be used as indicators to evaluate engagement, and then the attractiveness of each situation and the user's interest points can be understood, thereby providing data support for subsequent optimization and configuration.

[0080] Then, the user terminal device is connected, and after obtaining the user's read permission, a user engagement incentive framework is built. The user engagement incentive framework includes gamification elements, reward mechanisms, and social sharing functions, which can enhance the user's interactive experience. Gamification elements can include points, levels, and achievement systems. The reward mechanism can motivate users through virtual prizes or real rewards, and the social sharing function allows users to share their experiences and achievements with friends. The incentive means are designed to increase the user's participation in the virtual situation and make the emotional feedback process more interesting and attractive.

[0081] Next, based on the user engagement incentive framework, the optimal introduction time and duration are configured against N virtual scenarios and N scenario engagements. By analyzing the user engagement data in different scenarios, the optimal introduction time for each scenario can be determined, that is, when to introduce the scenario to maximize user interest and engagement. In addition, the optimal duration of each scenario is determined, that is, the optimal length of time the user stays in the scenario, to ensure that the user's emotional experience is not affected by too long or too short an interaction time. The optimization configuration can be dynamically adjusted based on real-time data to ensure that users get the best experience in every interaction.

[0082] The user participation incentive strategies corresponding to the N virtual situations include passive incentive regulation and active incentive regulation. Furthermore, a composite strategy design of situational awareness and user behavior analysis is adopted, such as dynamically adjusting the incentive mechanism to adapt to user reactions in different situations, using machine learning algorithms to predict user participation trends, etc., to further accelerate the user feedback loop and participation improvement, improve the incentive effect, and ensure the consistency of user participation and the stability of experience quality in different virtual situations. The user participation incentive strategy includes passive incentive regulation and active incentive regulation. Method 1: Passive incentive regulation relies on the internal mechanism design in the incentive architecture to achieve the improvement of user participation. For example, by setting up an achievement system and reward mechanism closely related to the virtual situation, users are attracted and driven in the natural experience process, thereby increasing participation. This method does not rely on additional external stimuli and has the advantage of low implementation cost. Method 2: Active incentive regulation uses external stimuli (such as limited-time activities, social sharing rewards, personalized recommendations, etc.) to achieve rapid improvement of user participation. For example, using limited-time activities to create a sense of urgency and encourage users to actively participate in order to obtain unique rewards. This method has the advantages of fast response speed and significant incentive effect.

[0083] By obtaining situational engagement, building a user engagement incentive framework, and configuring the optimal introduction timing and duration based on engagement data, the user's interactive experience and emotional support in virtual situations can be improved.

[0084] Furthermore, the present application also includes:

[0085] Based on the user engagement incentive architecture, the N virtual situations and N situation engagements are compared, and random expansion optimization is performed with the positive impact corresponding to the positive emotion indicator, and the authorization protection specification is read; based on the authorization protection specification, an authorization control license is issued, and after the authorization control license is authenticated, home appliances are connected, and the home appliances include lighting equipment, audio equipment, and display equipment; the home appliances are integrated and linked control is performed in accordance with the N virtual situations.

[0086] Specifically, based on the user engagement incentive framework, N virtual situations and N situation engagements are compared, and the positive impact corresponding to the positive emotion indicator is randomly expanded and optimized. Positive emotion indicators refer to the positive emotions of users in emotional interactions, such as happiness and excitement. Through random expansion optimization, the gamification elements, reward mechanisms, and social sharing functions in virtual situations are selected and adjusted to maximize the user's positive emotional response, thereby continuously improving the user's interactive experience and emotional feedback effect.

[0087] Then, based on the authorization protection specification, after issuing the authorization control license, authentication is performed, and the home devices are connected after authentication. The authorization protection specification is a standard to ensure user privacy and security. Issuing the authorization control license means that the user's home devices can be accessed and controlled only after the user's explicit permission. Home devices include lighting equipment, audio equipment, and display equipment, which can be linked with virtual scenarios to enhance the user's immersion and interactive experience.

[0088] Next, integrate home devices and perform linkage control in accordance with N virtual situations. Link virtual situations with home devices. For example, when users experience virtual situations, adjust the brightness and color of lighting devices, play sound effects suitable for the situation, and present relevant visual content on display devices. Linkage control can be dynamically adjusted according to the user's situational engagement and positive emotion indicators to ensure that each interaction can provide the best emotional support and immersive experience, thereby extending the impact of virtual situations to real environments and further enhancing the user's emotional experience.

[0089] Through random expansion optimization and authorization protection specifications based on the user engagement incentive architecture, home devices are integrated and linked, controlled according to virtual scenarios, providing immersive and personalized emotional support and interactive experience.

[0090] In summary, the robot adaptive emotion feedback system based on multimodal intelligent perception provided by this application has the following technical effects:

[0091] A multimodal perception data acquisition module is used to acquire the multimodal perception data of the target user through the multimodal perception data acquisition module, and the multimodal perception data acquisition module includes a visual perception device, an auditory perception device, a tactile perception device, and a physiological signal perception device; an emotion category recognition module is used to identify the emotion category based on the multimodal perception data and the initial emotion recognition model corresponding to the pre-training associated with the personality type framework, and the emotion category includes a positive emotion indicator, a neutral emotion indicator, and a reverse emotion indicator; a virtual situation construction module is used to build a first virtual situation based on the multimodal perception data and the positive emotion indicator in the emotion category; to build a second virtual situation based on the multimodal perception data and the neutral emotion indicator in the emotion category; to build a third virtual situation based on the multimodal perception data and the reverse emotion indicator in the emotion category; a first emotion feedback set acquisition module is used to, in the first virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the first emotion feedback A second emotion feedback set obtaining module is used to, in the second virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the second emotion feedback set; a third emotion feedback set obtaining module is used to, in the third virtual situation, combine the personality type defined by the personality type framework of the target user, and perform emotion regulation and emotion change analysis through the change trend of the multimodal perception data to obtain the third emotion feedback set; an adaptive emotion feedback result obtaining module is used to, based on the historical conversation content of the target user, use natural language processing technology to perform deep emotion experience mining on the user's language use, expression mode and emotional vocabulary, obtain deep emotion indicators, and the deep emotion indicators reflect the user's deep and lasting emotional experience, and perform contradictory adaptive assimilation on the first emotion feedback set, the second emotion feedback set, and the third emotion feedback set to obtain adaptive emotion feedback results, thereby achieving the technical goals of accurate emotion recognition and personalized emotion feedback, and achieving the technical effect of improving user interaction satisfaction and emotional experience.

[0092] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0093] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A robot adaptive emotional feedback system based on multimodal intelligent perception, characterized in that: include: A multimodal perception data acquisition module, used to acquire multimodal perception data of a target user through the multimodal perception data acquisition module, wherein the multimodal perception data acquisition module includes a visual perception device, an auditory perception device, a tactile perception device, and a physiological signal perception device; An emotion category recognition module is used to identify emotion categories based on the multimodal perception data using an initial emotion recognition model corresponding to the personality type framework association pre-training, wherein the emotion categories include positive emotion indicators, neutral emotion indicators, and negative emotion indicators; A virtual situation building module, used to build a first virtual situation based on the multimodal perception data and the positive emotion indicators in the emotion category; to build a second virtual situation based on the multimodal perception data and the neutral emotion indicators in the emotion category; and to build a third virtual situation based on the multimodal perception data and the negative emotion indicators in the emotion category. A first emotion feedback set obtaining module is used to obtain a first emotion feedback set by performing emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in the first virtual scenario in combination with the personality type defined by the personality type framework of the target user; A second emotion feedback set obtaining module is used to obtain a second emotion feedback set by performing emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in the second virtual situation in combination with the personality type defined by the personality type framework of the target user; A third emotion feedback set obtaining module is used to obtain a third emotion feedback set by performing emotion regulation and emotion change analysis based on the change trend of the multimodal perception data in the third virtual scenario in combination with the personality type defined by the personality type framework of the target user; An adaptive emotion feedback result acquisition module is used to perform deep emotion experience mining on the user's language usage, expression mode and emotional vocabulary based on the historical conversation content of the target user through natural language processing technology to obtain deep emotion indicators, which reflect the user's deep and lasting emotional experience, and to perform contradictory adaptive assimilation on the first emotion feedback set, the second emotion feedback set and the third emotion feedback set to obtain adaptive emotion feedback results, including identifying emotion change obstacles based on the change trend of the multimodal perception data; Collecting the interest information of the target user, and using a generative adversarial network to create multiple virtual situations that attract the target user with respect to the emotion change obstacle point, wherein the multiple virtual situations include a first virtual situation, a second virtual situation, and a third virtual situation; Based on the multiple virtual scenarios, matching the target user's attention level and the progress of the emotion change process corresponding to the emotion change obstacle point is performed to determine multiple virtual adaptation scenarios, including: based on the physiological signal sensing device, using signal processing technology to analyze the physiological signal and extract physiological signal features; Based on the physiological signal feature, the carrier form is associated with the attention degree of the target user to obtain a first associated feature, wherein the carrier form includes a picture, text, and audio; Based on the physiological signal feature, the carrier form is associated with the progress of the emotion change process corresponding to the emotion change obstacle point to obtain a second associated feature; Based on the carrier form, matching and filtering the multiple virtual scenarios by comparing the first correlation feature and the second correlation feature to determine multiple virtual adaptation scenarios, including obtaining user rating levels, and performing matrix decomposition with the first correlation feature as a row and the second correlation feature as a column to build a basic preference matrix of the target user; Traversing the interest information of the target user, and building a first collaborative preference matrix, a second collaborative preference matrix, and finally an Mth collaborative preference matrix corresponding to the target user; Based on the basic preference matrix of the target user, and the corresponding first collaborative preference matrix, second collaborative preference matrix, and up to the Mth collaborative preference matrix, collaborative filtering is performed to determine multiple virtual adaptation scenarios, including, based on the basic preference matrix of the target user, sorting the multiple virtual scenarios by matching degree to obtain multiple virtual scenarios sorted by priority; Based on the plurality of virtual scenarios ranked by priority, performing iterative combination optimization in combination with a first collaborative preference matrix, a second collaborative preference matrix, and up to an Mth collaborative preference matrix corresponding to the target user, to obtain a plurality of iterative combinations; The multiple iterative combinations correspond one-to-one to the multiple virtual adaptation scenarios, and the virtual adaptation scenarios include a combination scenario formed by N virtual scenarios with optimal introduction timing and optimal duration, and the user engagement incentive strategies corresponding to the N virtual scenarios include passive incentive regulation and active incentive regulation.

2. The robot adaptive emotional feedback system based on multimodal intelligent perception as claimed in claim 1, characterized in that: The virtual adaptation scenario includes a combination scenario formed by N virtual scenarios with optimal introduction timing and optimal duration, and also includes: Recording interaction information corresponding to a first virtual situation, the interaction information including a situation start time point, a situation duration, and user response data, wherein the first virtual situation is any one of the N virtual situations; Taking entering the first virtual situation as the starting point of the emotion change and completing the first virtual situation as the end point of the emotion change, tracking the emotion change, and setting a first emotion change trajectory; Based on the interaction information corresponding to the first virtual situation and the first emotion change trajectory, the N virtual situations are traversed to perform random expansion optimization.

3. The robot adaptive emotional feedback system based on multimodal intelligent perception as claimed in claim 2, characterized in that: Based on the interaction information corresponding to the first virtual situation and the first emotion change trajectory, traversing the N virtual situations includes: Based on the N virtual situations, obtaining N situation participation degrees; Connecting to the user terminal device, after obtaining the read permission, building a user engagement incentive framework, the user engagement incentive framework enhances the interaction of the target user with gamification elements, reward mechanism, and social sharing function; Based on the user engagement incentive framework, the optimal introduction timing and the optimal duration are configured in comparison with the N virtual situations and the N situation engagements.

4. The robot adaptive emotional feedback system based on multimodal intelligent perception as claimed in claim 3, characterized in that: include: Based on the user engagement incentive framework, the N virtual situations and the N situation engagements are compared, the positive impact corresponding to the positive emotion indicator is randomly expanded and optimized, and the authorization protection specification is read; Based on the authorization protection specification, issuing an authorization control license, and after the authorization control license is authenticated, connecting home devices, the home devices include lighting devices, audio devices, and display devices; The home appliances are integrated to perform linkage control in accordance with the N virtual situations.

Citation Information

Cited By

  • Multi-modal sensing humanoid robot action self-adaptive control method and multi-modal sensing humanoid robot action self-adaptive control system

    CN119610112A

  • Multimodal perception humanoid robot motion adaptive control method and system

    CN119610112B

  • Cognitive training emotion interaction method and system based on large model

    CN120148767A

  • Optimization method of robot control model and related equipment

    CN120715901A

  • Interaction system based on emotion recognition

    CN121614838A