Intelligent agent-based personalized story generation and broadcasting method, device and terminal

Abstract

This invention discloses a personalized story generation and broadcasting method, device, and terminal based on intelligent agents, relating to the field of artificial intelligence technology. The method includes: for a timed story generation process, reading a preset user profile and acquiring real-time contextual data; for a story generation process triggered by an active story generation command, parsing the active story generation command, extracting key elements including story theme, style, and / or characters, and acquiring a corresponding preset user profile; deeply fusing the parsed key elements with the corresponding preset user profile data and / or real-time contextual data to obtain story generation prompts; inputting the fused prompts into a preset story generation engine, dynamically generating an original story that integrates user personal characteristics and real-time context using a large-scale language model; and broadcasting the story immersively through voice and / or visual means. This invention has the advantages of high personalization, dynamic generation of original stories, and support for both active and timed triggering.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, smart terminal, and storage medium for personalized story generation and broadcasting based on intelligent agents. Background Technology

[0002] With the development of technology and the continuous improvement of people's living standards, the use of various smart terminals is becoming increasingly widespread. Smart terminals have become indispensable communication and entertainment tools in people's lives. Sometimes, users can use smart terminals such as smartphones to play bedtime stories for children.

[0003] Current technologies for children's bedtime stories or daily story entertainment mainly rely on parents telling and playing fixed audio or video resources, which has obvious limitations: 1) Homogeneous and depleted content: Pre-recorded story resources are limited, and children easily lose interest after repeated listening, while parents also face the problem of running out of story material. 2) Lack of personalization: Existing story content is monotonous and cannot incorporate personalized elements such as the child's name, preferences, and daily experiences, lacking a sense of immersion and closeness. 3) Weak interactivity: Most story machines or apps can only passively play stories and cannot dynamically generate new story content based on the child's immediate interests and instructions, resulting in a single interaction mode. Current technologies lack a smart service that can proactively and timely integrate into children's daily lives (such as fixed bedtimes), requiring manual operation by parents, which is not convenient enough.

[0004] Therefore, there is an urgent need for a solution that can leverage the home-centric large TV screen, combined with AI technology, to automatically generate highly personalized and interactive story content on demand or in real time. Summary of the Invention

[0005] The technical problem this invention aims to solve is to address the shortcomings of the prior art by providing a personalized story generation and broadcasting method, device, smart terminal, and storage medium based on intelligent agents. This invention offers advantages such as high personalization, dynamic generation of original stories, and support for both active and timed triggering, thus resolving the issues of content homogenization, lack of personalized customization, weak interaction, and insufficient convenience in the prior art. This invention can dynamically generate original stories that integrate user characteristics and real-time context based on preset conditions (such as timed events) or user-initiated commands, and deliver them in an immersive broadcasting format through voice and visual means, thereby providing a unique entertainment and companionship experience.

[0006] The technical solution adopted by this invention to solve the problem is as follows: A personalized story generation and broadcasting method based on intelligent agents, comprising: When a story generation instruction is received or the current time reaches the preset time for timed story generation, the story generation process is triggered. When the story generation process is triggered on a timer, it controls the reading of preset user profiles and the acquisition of real-time contextual data. When the story generation process is triggered by an active story generation instruction, the active story generation instruction is parsed, key elements including story theme, style and / or characters are extracted, and corresponding preset user profiles are obtained. The extracted key elements are deeply integrated with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The fused prompts are input into a preset story generation engine, which uses a large language model to dynamically generate original stories that blend user personal characteristics with real-time context. The generated original stories will be narrated in an immersive manner through voice and / or visual means.

[0007] The aforementioned agent-based personalized story generation and broadcasting method, wherein the step of controlling the triggering of the story generation process before receiving an active story generation instruction or the current time reaching the preset time for timed story generation includes: The system collects and obtains personalized data from at least one child user in advance. The personalized data includes: name, nickname, age, gender, favorite animal, color, food, hobbies, fears, and / or family member information.

[0008] The aforementioned agent-based personalized story generation and broadcasting method, wherein the step of controlling the triggering of the story generation process before receiving an active story generation instruction or the current time reaching the preset time for timed story generation includes: Preset the preset time for automatically starting the story generation process.

[0009] The aforementioned agent-based personalized story generation and broadcasting method, wherein the story generation process, which is triggered at a time, includes controlling the reading of a preset user profile and obtaining real-time contextual data, comprises: Real-time access to the family calendar to retrieve specified events, including birthdays, holidays, and / or travel plans; It can connect to smart home devices in real time and obtain environmental information.

[0010] The aforementioned agent-based personalized story generation and broadcasting method includes the following step: deeply fusing the parsed key elements with corresponding preset user profile data and / or real-time context data to obtain story generation prompts. When the story generation process is triggered on a timed basis, the system reads a preset user profile by default to obtain real-time contextual data, including the date and weather conditions, and uses this data to generate prompts for the timed story generation. When the story generation process is triggered by an active story generation command, the system parses the active story generation command, extracts key elements including story theme, style and / or characters, and obtains the corresponding preset user profile; and deeply integrates the parsed key elements with the corresponding preset user profile data to obtain the prompt words for story generation of the active command.

[0011] The aforementioned agent-based personalized story generation and broadcasting method further includes the step of immersive broadcasting of the generated original story via voice and / or visual means: Obtain the generated original story; The generated original stories are reviewed and filtered through a preset safety filter to output original stories that meet children's safety and values ​​standards. The original stories that have been filtered and approved are processed using voice synthesis technology to be read in a tone and rhythm suitable for bedtime stories. And / or analyze the original stories that have been filtered and reviewed, and dynamically switch the corresponding background images or animated scenes based on the key plot points of the analyzed story text to provide an immersive visual presentation.

[0012] The aforementioned agent-based personalized story generation and broadcasting method further includes, after the step of immersively broadcasting the generated original story via voice and / or visual means: When the generated original story broadcast ends, ask questions proactively according to the preset question prompts; Receive user feedback on proactively asked questions; The next story generation will be optimized based on the feedback information.

[0013] A personalized story generation and broadcasting device based on intelligent agents, wherein the device includes: The story generation process trigger module is used to control and trigger the story generation process when a story generation instruction is received or the current time reaches the preset time for timed story generation. The timed trigger data acquisition module is used to control the reading of preset user profiles and the acquisition of real-time contextual data when the story generation process is timed. The active data acquisition module is used to control the parsing of the active story generation instruction when the story generation process is triggered by the active story generation instruction, extract key elements including story theme, style and / or characters, and obtain the corresponding preset user profile; The data fusion and prompt word generation module is used to deeply fuse the parsed key elements with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The original story generation module is used to input the fused prompts into the preset story generation engine and dynamically generate original stories that integrate user personal characteristics and real-time context using a large language model; The original story playback module is used to immerse the generated original stories through voice and / or visual presentations.

[0014] A smart terminal includes a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising the method for performing any one of the methods.

[0015] A computer-readable storage medium, wherein, when instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform any of the methods described herein.

[0016] The present invention provides a personalized story generation and broadcasting method, device, smart terminal, and storage medium based on intelligent agents. By dynamically generating original stories that integrate user personal characteristics and real-time context, the present invention solves the problems in the background technology and has the advantages of high personalization, dynamic generation of original stories, and support for active and timed triggering. The present invention adds new functions to smart terminals: it can dynamically generate original stories that integrate user personal characteristics and real-time context according to preset conditions (such as timed triggering) or user active commands, and broadcast them in an immersive manner through voice and visual means; the generated stories are more suitable for the listener, providing convenience for the user. Furthermore, the present invention also has the following advantages: 1) The generated stories are highly personalized: The story content is deeply customized, incorporating children themselves and elements of their lives into the plot, which greatly enhances the story's sense of immersion, intimacy, and appeal.

[0017] 2) It has great content unlimitedness and innovation: Utilizing AI generation capabilities, it is theoretically possible to generate an unlimited number of non-repeating story contents, completely solving the problem of story resource depletion.

[0018] 3) Intelligent and convenient: It supports automatic generation at set times and can be seamlessly integrated into the daily rhythm of children's lives, providing parents with a worry-free and labor-saving parenting aid.

[0019] 4) Immersive multi-sensory experience: Combining the visual advantages of a large TV screen with high-quality voice broadcasting, it creates an immersive story experience that far surpasses traditional story machines, helping to cultivate children's listening skills and imagination.

[0020] 5) Promote interaction and expression: Through the question and feedback mechanism after the story, children are encouraged to interact with the system and exercise their language expression and thinking skills. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the personalized story generation and broadcasting method based on intelligent agents provided in Embodiment 1 of the present invention.

[0023] Figure 2 This is a flowchart illustrating the personalized story generation and broadcasting method based on intelligent agents provided in Embodiment 2 of the present invention.

[0024] Figure 3 The principle block diagram of the embodiment of the personalized story generation and broadcasting device based on intelligent agents provided by the present invention.

[0025] Figure 4 This is a block diagram illustrating the internal structure of a smart terminal provided in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0027] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0028] In traditional children's story generation systems, pre-recorded story resources are stored on fixed media, making content homogenization and depletion unavoidable, and dynamic content updates impossible. Furthermore, the lack of customization capabilities based on user-specific data prevents the integration of elements such as children's names, preferences, and daily experiences into the story content. In addition, the interaction mechanism is weak; the system can only passively respond to preset commands, and real-time input cannot be used to generate new content. The absence of a timed trigger function prevents story generation from automatically integrating into daily routines, relying on manual intervention. Content depletion directly impacts the system's content freshness index, lack of personalization reduces user engagement, weak interactivity limits system responsiveness, and the lack of a timed trigger function increases the frequency of user operations.

[0029] For example, in a home environment, children ask for stories every night before bed, and the system's pre-stored story library is repeatedly accessed, causing children to become bored with the repetitive content. When a child asks, "Tell me a story about playing with my puppy today," this instruction may not be parsed by the system, and the relevant story cannot be generated. Meanwhile, parents need to manually start story playback every night because a timed trigger mechanism is not configured, making automatic execution at fixed times impossible. Specifically, children's real-time contextual data, such as their experiences that day, cannot be obtained, the integration of user profile data and instruction elements cannot be completed, the story generation process becomes rigid, and the system cannot adapt to children's dynamic needs.

[0030] If the above problems are not addressed, it will be difficult to maintain users' long-term interest, the freshness of content will continue to decline, and the system usage rate will decrease; insufficient interactivity will prevent the system from adapting to children's dynamic needs, and the educational and entertainment value will be reduced; the lack of automatic triggering mechanisms will increase the user's operational burden, hinder the seamless integration of the system into the smart home environment, and limit the depth of AI technology application in the field of children's services.

[0031] To address this, this application proposes a personalized story generation and broadcasting method based on intelligent agents. For example... Figure 1 As shown in Embodiment 1 of the present invention, a personalized story generation and broadcasting method based on an intelligent agent includes the following steps: S100: When a story generation instruction is received or the current time reaches the preset time for timed story generation, the story generation process is triggered. The intelligent agent refers to an autonomous entity capable of sensing the environment, making decisions, and executing actions. In the method of this embodiment, the intelligent agent is responsible for coordinating all aspects of story generation and broadcasting, ensuring the automation and intelligence of the entire process.

[0032] The personalized story generation and broadcasting method in this embodiment refers to a method that can dynamically create and present original stories in an immersive manner based on the personal characteristics and real-time context of a specific user, providing a highly customized content experience.

[0033] The personalized story generation and broadcasting in this embodiment of the invention can be triggered in two ways: one is that the user actively issues a story generation command; the other is that a timer is set, and the story generation process is triggered when the timer expires.

[0034] S200, When the story generation process is triggered on a timer, control the reading of preset user profiles and the acquisition of real-time contextual data; The user persona in this step refers to a set of user characteristics constructed by collecting and analyzing multi-dimensional user data (such as age, gender, interests, preferences, and dislikes). The persona is used to guide the personalized creation of story content, making it more closely aligned with user preferences.

[0035] In this embodiment, real-time contextual data refers to instantaneous information obtained from the external environment or device at the time of story generation. This data may include, but is not limited to, current time, weather conditions, information about the user's physical environment, and schedule, and is used to combine the story content with the user's current state or surrounding environment.

[0036] S300. When the story generation process is triggered by an active story generation instruction, the active story generation instruction is parsed, key elements including story theme, style and / or characters are extracted, and corresponding preset user profiles are obtained. S400: Deeply integrate the extracted key elements with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; S500: Input the fused prompts into the preset story generation engine and use a large language model to dynamically generate an original story that integrates user personal characteristics and real-time context. S600 immersively broadcasts the generated original stories via voice and / or visual means.

[0037] In this embodiment, the story generation engine refers to an integrated software module or system whose core function is to receive input (e.g., prompts) and use internal algorithms or models (e.g., large language models) to create story text. In this embodiment, the story generation engine is the core component for achieving dynamic story generation.

[0038] The large-scale language model in this embodiment is a neural network model with massive parameters trained based on deep learning technology. It can understand, generate and process human language, and has strong language understanding and generation capabilities, as well as text creation, continuation writing, and style transfer capabilities. It is a key technology for realizing the dynamic generation of original stories.

[0039] Immersive broadcasting refers to presenting story content in an attractive way by combining multiple sensory forms such as voice and / or vision, making users feel as if they are in the story scene, thereby enhancing the realism and immersion of the story experience.

[0040] This invention provides a personalized story generation and broadcasting method based on intelligent agents. Its core lies in dynamically generating and presenting customized story content according to user needs or preset conditions. In specific implementation, the story generation process can be triggered in various ways. For example, the user can issue an active story generation command via voice command, touch screen, or application interface, which is then received and recognized by the system. Alternatively, the system can be configured to automatically check whether the current time matches a preset timed story generation time at a specific time, such as before bedtime each day or on specific holidays. When any triggering condition is met, the story generation process is initiated.

[0041] When the story generation process is triggered periodically, the system will automatically perform data acquisition operations. Specifically, pre-stored user profile data will be read, which includes the user's basic information, interests, and preferences. Simultaneously, the system will acquire current real-time contextual data, such as ambient temperature, light intensity, and the current date obtained through sensors or external interfaces. This data acquisition aims to provide basic personalized and contextualized input for subsequent story generation.

[0042] When the story generation process is triggered by an actively generated story instruction, the system will parse that instruction. The parsing process aims to identify key elements from the user's instructions, such as the explicitly specified story theme (e.g., "a story about space exploration"), story style (e.g., "a lighthearted and humorous style"), or desired characters (e.g., "a dog as the protagonist"). While extracting these key elements, the system also obtains pre-defined user profile data corresponding to the current user to ensure the story's personalization.

[0043] Subsequently, the system of this embodiment of the invention deeply integrates the parsed key elements with corresponding preset user profile data and / or real-time contextual data. This integration process aims to combine the user's personalized preferences, real-time environmental information, and user-specified story requirements to form a unified input containing contextual information. For example, the user profile's "likes dinosaurs" can be combined with the real-time context's "rainy day" and the active command's "adventure story" to generate a prompt word about "dinosaur adventure on a rainy day." The prompt word is key to guiding large-scale language models in generating stories.

[0044] Next, the fused prompts are input into a pre-defined story generation engine. This engine integrates a large language model, which, based on its strong language understanding and generation capabilities, dynamically creates original stories based on the prompts. These stories not only incorporate the user's personal characteristics, such as integrating the user's name or preferences into the plot, but also combine real-time context, making the story content highly relevant to the user's current state or environment. For example, if the user profile shows that the user likes blue, and the real-time context is nighttime, the story will include phrases like "a blue moon" or "a little protagonist wearing blue pajamas."

[0045] Finally, the generated original story will be presented to users through an immersive narration. The narration can be purely audio, playing the story content through a speaker, with adjustable speed and tone to suit different scenarios. Alternatively, it can be purely visual, such as displaying the story text on a screen accompanied by static illustrations or simple animations. It can also combine audio and visual elements, for example, simultaneously displaying images or text related to the story content on the screen while the audio story is playing, to enhance user immersion and comprehension.

[0046] The following example will provide a more detailed explanation of the above technical solution: For example, user A is a five-year-old child. Their user profile records their name as "Xiaoming," age as "five," favorite animal as "dinosaur," favorite color as "green," and fear of "thunder." In this embodiment of the invention, the system is configured to automatically read a bedtime story to user A every evening at 8:30 PM.

[0047] When the time reaches 8:30 PM, the system receives a preset time signal for the story generation, thus triggering the story generation process. At this time, the system automatically reads user A's preset user profile and obtains information such as "Xiaoming," "five years old," "dinosaur," "green," and "afraid of thunder." Simultaneously, the system in this embodiment of the invention acquires real-time contextual data, for example, through smart home devices, it obtains information such as the current indoor temperature of 25 degrees Celsius, dim lighting, and an outdoor weather forecast indicating light rain tonight.

[0048] Because it's a timed trigger, the system will by default read user A's user profile and combine it with real-time contextual data. For example, it will deeply integrate elements such as "Xiaoming," "dinosaur," and "green" from the user profile with information such as "night" and "light rain" from the real-time context. The integration process will generate a prompt, such as: "Create a bedtime story for five-year-old Xiaoming about a green dinosaur going on an adventure on a rainy night. The story should not include any thunder." In this embodiment, the fused prompts are then input into a pre-defined story generation engine. The engine utilizes a large language model to dynamically generate an original story based on the prompts. For example, the story could revolve around the plot of "Xiaoming and his green dinosaur friend searching for undiscovered treasure in the rainy night, encountering friendly animals along the way, and finally successfully returning home before dawn." The story would cleverly incorporate the name "Xiaoming," describe the image of the "green dinosaur," and avoid mentioning "thunder" to mitigate user A's fear.

[0049] Once the story is generated, the system of this embodiment can present the original story through immersive broadcasting. For example, the story text is converted into speech with a tone and rhythm suitable for a bedtime story and played through a smart speaker. Simultaneously, background images or simple animated scenes matching the storyline are displayed on the connected smart TV screen. For instance, when the story describes a "rainy night," an animation of raindrops falling is displayed on the screen; when it describes a "green dinosaur," an image of a green dinosaur appears on the screen. This broadcasting method combining voice and visuals allows user A to be fully immersed in the story, obtaining a highly personalized and immersive experience. As can be seen from the above examples, the method of this embodiment can dynamically generate and broadcast unique original stories according to the user's personalized needs and real-time context, effectively solving the problems of content homogenization, lack of personalization, and weak interactivity in existing technologies.

[0050] As can be seen from the above, compared with existing technologies that rely on pre-recorded, fixed content for children's stories, the method of this invention achieves dynamic generation and strong personalization of story content by introducing intelligent agents, user profiles, real-time contextual data, and large-scale language models. For example, in the example of user A, traditional methods can only play preset dinosaur stories, while the method of this invention can generate a unique, original story highly relevant to user A based on user A's name, preferences (green dinosaur), and even avoiding things they fear (thunder), combined with the real-time context (rainy night). This capability effectively solves the problems of homogenized and depleted story content in existing technologies, significantly enhancing the story's appeal and user engagement.

[0051] Furthermore, this method enhances the system's interactivity and ability to integrate into daily life rhythms by supporting both actively generated instructions and timed triggering modes. Traditional story machines or apps typically only play passively, requiring parents to manually select and operate them. This invention, however, not only allows user A to customize stories through instructions but also enables automatic story playback at preset times, such as providing a story service before bedtime each night, significantly improving convenience and making story playback a part of the user's daily routine.

[0052] Furthermore, by deeply integrating user profiles, real-time contextual data, and key elements of user commands, the method of this invention can generate highly contextualized prompts, thereby guiding large-scale language models to create original stories that both conform to user preferences and fit the current environment. This multi-dimensional data fusion strategy makes the generated stories not only personalized but also more relevant to reality, further enhancing the user's immersive experience.

[0053] Ultimately, through immersive storytelling in the form of voice and / or visuals, the method of this invention provides users with a multi-sensory story experience. Compared with existing technologies that only provide single audio or video playback, this method can dynamically adjust the storytelling format according to the story content, such as combining voice suitable for bedtime stories with synchronized visual scenes, making the story presentation expressive and attractive, thereby effectively solving the problems of weak interactivity and lack of immersion in existing technologies.

[0054] In summary, this method, through an intelligent story generation and broadcasting mechanism, overcomes the limitations of existing technologies in terms of content personalization, interactivity, contextual adaptability, and integration into daily life, providing users with a novel and highly customized story experience.

[0055] In other embodiments, this application proposes a personalized story generation and broadcasting method based on intelligent agents, and further proposes to collect and obtain the personalized data storage of at least one child user in advance before controlling the triggering of the story generation process. The personalized data includes: name, nickname, age, gender, favorite animal, color, food, hobbies, fears and / or family member information.

[0056] The system pre-collects and acquires personalized data from at least one child user to ensure that detailed information related to the child user is actively or passively acquired and stored before the story generation process begins. This pre-processing data mechanism allows the system to instantly access rich user background data when a story needs to be generated, laying a solid foundation for subsequent personalized story creation. For example, this data can be collected through guided questionnaires when the user first uses the system, manual input by parents, or through gradual learning and accumulation during the system's interaction with the child. Furthermore, authorized connections can be established with third-party applications (such as children's educational apps and smart wearable devices) to obtain user behavior data and preference information on other platforms.

[0057] The personalized data includes: name, nickname, age, gender, favorite animal, color, food, hobbies, fears, and / or family member information. These specific data items constitute the core elements of a child user's personalized profile. They comprehensively cover a child's basic identity information, physiological characteristics, emotional preferences, and social relationships, and are essential for constructing a three-dimensional and vivid user image. This information can directly or indirectly influence the story's theme, character settings, plot development, and emotional tone, making the generated story more immersive and engaging. This data can be stored in a structured form in a database, such as a relational or non-relational database, for rapid retrieval and fusion. Alternatively, it can be stored in the form of a knowledge graph, using different personalized data items as nodes connected by relationships to better capture the complex correlations between data, supporting deeper story generation.

[0058] This application embodiment ensures that the system has access to rich and in-depth user background information during the subsequent story generation process by pre-collecting and storing detailed personalized data of child users before the story generation process is triggered. When the story generation process is triggered, whether by a timed trigger or an active command, the system will obtain this pre-collected and stored detailed data, including name, nickname, age, gender, favorite animals, colors, foods, hobbies, fears, and / or family member information, when reading the preset user profile. When this detailed personalized data is deeply integrated with the parsed key elements (such as story theme, style, and / or characters) and real-time contextual data, more accurate and personalized story generation prompts can be generated. After receiving these highly customized prompts, the large language model can dynamically generate original stories that truly integrate user personal characteristics and real-time context. For example, the story can naturally incorporate the child's favorite animals as characters, or their preferred colors as the main color scheme of the scene, while avoiding the appearance of things they fear, thus making the generated story content more targeted and attractive.

[0059] The following is a specific example. For instance, when a smart story-telling device is first activated, the system of this embodiment guides parents through a child information entry process. During this process, parents can enter the child's name (e.g., "Xiaoming"), nickname (e.g., "Mingming"), age (e.g., "5 years old"), gender (e.g., "male"), favorite animal (e.g., "puppy," "dinosaur"), color (e.g., "blue," "green"), food (e.g., "pizza," "strawberry"), hobbies (e.g., "drawing," "building blocks"), fears (e.g., "thunder"), and family member information (e.g., "father," "mother," "grandparents"). This detailed personalized data is collected and stored in a structured data format in the device's local storage or a cloud server, forming Xiaoming's unique user profile. When the user subsequently issues a story-generating command (e.g., "tell a story about dinosaurs") or when the preset scheduled playback time arrives, the system can directly access this preset personalized data. When generating a story, the system will set Xiaoming as the protagonist and incorporate his favorite dinosaur elements and blue scenes, while avoiding thunder scenes, thus generating a highly personalized original story that fits Xiaoming's interests and preferences.

[0060] By pre-collecting and storing detailed personalized data of child users before the story generation process is triggered, this application ensures that the system possesses sufficiently rich and in-depth user background information during story generation. This enables the subsequent story generation engine to generate original stories highly tailored to the individual characteristics of child users based on these detailed preferences and features during deep integration. For example, the story can include characters, scenes, or plots that children like, while avoiding elements that frighten them, thereby significantly enhancing the story's appeal, immersion, and educational value, effectively solving the problem of story content generalization caused by a lack of in-depth personalized information.

[0061] In other embodiments, this application proposes a personalized story generation and broadcasting method based on intelligent agents. This method can trigger the story generation process based on a received active story generation command or the current time reaching a preset time for story generation. However, in practical applications, relying solely on manual user triggering or passive system triggering at specific times may result in insufficient flexibility in story generation, failing to meet users' personalized needs for story generation timing in specific scenarios. For example, users might want stories to be automatically generated at a fixed bedtime each day without requiring manual operation each time.

[0062] In this regard, this application further proposes that when a story generation instruction is received or the current time reaches the preset time for timed story generation, the step of controlling the triggering of the story generation process includes: pre-setting the preset time for automatically starting the story generation process.

[0063] This embodiment of the steps pre-setting the preset time for automatically starting the story generation process means that before the actual triggering of the story generation process, the system allows users or administrators to pre-configure one or more specific time points. When the system detects that the current time matches these preset times, it will automatically start the story generation process. Its purpose is to provide an automated, periodic story generation mechanism, thereby reducing repetitive user operations and improving the system's intelligence and user experience. One implementation is that the system provides a user interface that allows users to set specific dates and times, such as 9 PM every night, through a calendar selector or time input box. Another implementation is that the system can intelligently recommend and allow users to confirm a preset time based on user habits or through machine learning algorithms. For example, after detecting that a user typically engages in bedtime activities during a certain time period, it can recommend automatically generating stories during that time period.

[0064] Building upon the aforementioned method, this application's solution pre-sets a preset time for automatically initiating the story generation process before it is triggered. This allows the system to automatically trigger the story generation process at the designated time based on user or administrator configuration. Specifically, the system continuously monitors the current time during operation. Once the current time matches any of the pre-set times, the system immediately triggers the story generation process. This mechanism, along with the triggering method that receives proactive story generation commands, constitutes the two main triggering pathways for the story generation process. In this way, the system can not only respond to users' immediate needs but also provide personalized story services according to a predetermined schedule without user intervention. The introduction of this preset time transforms story generation from a passive response into a proactive service, significantly enhancing the convenience and intelligence of the user experience.

[0065] The above technical solution introduces a preset time for automatically starting the story generation process in the agent-based personalized story generation and broadcasting method. This makes the story generation process no longer entirely dependent on user commands, but can automatically start according to the preset time. This automation mechanism greatly improves the convenience of the user experience, especially suitable for scenarios requiring regular story broadcasting, such as children's bedtime stories. Users only need to set it once, and the system can automatically provide personalized stories at the specified time, thereby reducing the user's operational burden, ensuring the timeliness and continuity of the story service, and making the personalized story broadcasting service more intelligent and user-friendly.

[0066] In some of the embodiments described above in this application, when the story generation process is triggered on a timer, it is proposed to control the reading of a preset user profile and the acquisition of real-time contextual data. However, in its implementation, if the method of acquiring real-time contextual data is too generalized or lacks specific sources, the acquired contextual information is not rich enough or closely related to the user's actual life, thereby affecting the personalization and attractiveness of the generated story, and making the story content not strongly related to the user's current environment or upcoming events.

[0067] In response, this application further proposes a story generation process that is triggered on a timer, which controls the reading of preset user profiles and the acquisition of real-time contextual data, including: real-time access to the family calendar to obtain specified events, including birthdays, holidays and / or travel plans; and real-time access to smart home devices to obtain environmental information.

[0068] This feature involves real-time access to the family calendar to retrieve specific events, including birthdays, holidays, and / or travel plans. The aim is to proactively obtain information on specific events closely related to the user's life by connecting in real-time with the calendar system used daily by family members. These specified events can include, but are not limited to, the user's birthday, important holidays, family anniversaries, and pre-planned travel plans. Implementation methods include integration with mainstream digital calendar services (such as Google Calendar, Outlook Calendar, and Apple Calendar) via an Application Programming Interface (API) to periodically synchronize calendar data; or, the system can provide an interface allowing users to manually input or import family calendar data and set event types and reminders. Another approach is to directly access the linked family calendar information and extract relevant events through smart terminal devices such as smart speakers or smart displays.

[0069] The real-time access to smart home devices and acquisition of environmental information aims to obtain various physical parameters and status information of the current environment through real-time interaction with the user's smart home environment. This environmental information may include, but is not limited to, indoor temperature, humidity, light intensity, air quality (such as PM2.5 and CO2 concentration), the open / closed status of windows or doors, and even the user's activity status detected by sensors. In terms of implementation, this can be achieved by connecting various smart home devices through a unified interface of the smart home platform (e.g., Matter, HomeKit, Zigbee, Z-Wave protocols) to receive data reported by the devices in real time; or, the system can directly communicate with specific smart devices (such as smart thermostats, smart light bulbs, environmental sensors, etc.) to obtain their current status data. Furthermore, a smart home hub or gateway device can centrally collect and process sensor data from different smart devices and provide it as environmental information to the story generation process.

[0070] This application's solution concretizes the "acquiring real-time contextual data" step in the timed story generation process, enabling it to obtain information from more personalized and real-time data sources. Specifically, when the system detects the need to generate a story at a specific time, it no longer simply acquires general contextual data, but actively accesses the family calendar in real time to obtain specific events closely related to the user's life trajectory, such as birthdays, holidays, or travel plans. Simultaneously, the system also accesses smart home devices in real time to obtain real-time physical information about the user's environment, such as current temperature, lighting conditions, or weather. This specific and dynamic real-time contextual data obtained from the family calendar and smart home devices, along with preset user profile data, is deeply integrated to generate story prompts. This mechanism ensures that the generated prompts not only include the user's personal preferences but also incorporate the user's current real-world environment and upcoming events, allowing a large language model to dynamically generate original stories highly aligned with the user's personal characteristics, real-time context, and life events. In this way, this application effectively solves the problem of insufficient or generalized real-time contextual data acquisition leading to low story personalization in traditional methods, significantly improving the relevance and appeal of the stories.

[0071] The following is a specific example. Imagine a child user whose family calendar indicates next week is their birthday, and it's raining outside today while the indoor smart lighting system detects dim lighting. When the system reaches the preset story generation time, this embodiment of the invention first accesses the child's digital calendar in real-time to identify the specified event, "next week's birthday." Simultaneously, the system connects to the family's smart home devices in real-time, obtaining weather information ("rainy") from a smart weather station and environmental information ("dim indoor lighting") from the smart lighting system. This specific real-time contextual data (next week's birthday, rain, dim indoor lighting) is deeply integrated with the child's preset user profile (e.g., liking dinosaurs, the color blue, eating apples, fearing thunder, etc.). For example, the system generates a prompt containing the phrase, "A story about a little dinosaur preparing a surprise for his upcoming birthday on a rainy day. The story takes place indoors, in dim lighting, but full of warmth." This prompt is then input into the story generation engine, generating a personalized original story that aligns with the child's personal preferences and is closely related to the current environment and the upcoming event.

[0072] Through the aforementioned technical solutions, this application significantly enhances the personalization and contextual relevance of timed story generation. By real-time access to the family calendar, the system can capture important events in the user's life, such as birthdays, holidays, or travel plans, allowing the generated stories to revolve around these events, giving them deeper meaning and emotional connection. Simultaneously, real-time access to smart home devices to obtain environmental information enables the stories to integrate into the user's real physical environment, such as setting story scenes based on weather, lighting, or temperature, greatly enhancing the immersion and engagement of the stories. This refined and multi-dimensional acquisition of real-time contextual data effectively solves the problem of story content being disconnected from the user's actual life due to relying solely on generalized contextual data. This allows large-scale language models to generate more user-friendly and engaging original stories, thus providing children with a more unique and meaningful personalized story experience.

[0073] In some of the embodiments described above in this application, although it is proposed to deeply integrate the parsed key elements with user profile data and / or real-time contextual data to obtain prompt words for story generation, the data fusion strategies and prompt word generation methods under different triggering mechanisms (timed triggering or active command triggering) are not carefully distinguished, which may result in the prompt word generation being inaccurate or inefficient.

[0074] To address this, this application further proposes a step of deeply integrating the parsed key elements with corresponding preset user profile data and / or real-time contextual data to obtain prompts for story generation. This step includes: when the story generation process is timed, controlling the default reading of preset user profiles to obtain real-time contextual data including the date and weather conditions, forming prompts for timed story generation; when the story generation process is triggered by an active story generation command, controlling the parsing of the active story generation command to extract key elements including story theme, style, and / or characters, and obtaining corresponding preset user profiles; and deeply integrating the parsed key elements with the corresponding preset user profile data to obtain prompts for story generation from the active command.

[0075] The "default reading of preset user profiles" refers to the system automatically retrieving and loading personalized data associated with the target user from storage when the story generation process is triggered on a scheduled basis, without requiring additional user instructions or confirmation. This ensures that story generation is always based on user preferences and characteristics. For example, the system can pre-configure a default user ID or user group, and automatically load the user profile data corresponding to that ID or user group when the scheduled task starts; alternatively, the system can intelligently determine the currently active user based on the current time or device usage and automatically read the preset profile of that active user.

[0076] The acquisition of real-time contextual data, including the date and weather conditions, refers to the system's ability to dynamically collect and integrate external information related to the current environment and time during story generation, making the story content more timely and immersive. For example, in this embodiment of the invention, the date information can be obtained through an internal clock module, and the real-time weather conditions for a specified geographical location can be obtained by calling a third-party weather service API; alternatively, the system can integrate smart home devices, acquire data such as ambient temperature and humidity through sensors, and obtain date information in conjunction with network time protocols.

[0077] The aforementioned timed story generation prompt refers to the structured combination of default-read user profiles and acquired real-time contextual data to generate a text instruction that guides a large language model in generating stories. This aims to ensure that the timed generated stories align with user preferences and fit the current context. For example, key information from the user profile (such as name and preferences) can be concatenated with real-time contextual data (such as date and weather) using a preset template; alternatively, the system can employ more sophisticated natural language generation techniques to transform the fused data into more natural and guiding prompts.

[0078] The process of parsing the actively generated story instructions and extracting key elements, including story theme, style, and / or characters, refers to the system performing semantic analysis on the story generation request input by the user via voice or text. This analysis identifies and extracts the core components of the story explicitly specified by the user, ensuring that the user's intent is accurately understood and reflected. For example, the system can utilize natural language processing techniques, such as named entity recognition and intent recognition, to identify the story's theme words, desired style descriptions, and specified story characters from the user's instructions. Alternatively, the system can employ rule-based pattern matching or machine learning models to classify and extract information from the user's instructions.

[0079] The acquisition of the corresponding preset user profile refers to the system in this embodiment of the invention retrieving and loading corresponding personalized data from storage based on the current user identity or user information implicit in the instruction when receiving an active story generation instruction, so that the actively generated story can also fully consider the user's personalized needs. For example, the system in this embodiment of the invention can automatically associate and read the user's preset profile based on the user's login information or the user ID bound to the device; or, if the user instruction contains a mention of a specific user, the system can identify the information and load the profile of the corresponding user.

[0080] The deep integration of the parsed key elements with corresponding preset user profile data refers to the organic combination of user-specified key elements with the system's preset user profile data to form a comprehensive input that reflects both the user's immediate needs and long-term preferences, aiming to generate a highly personalized story that meets the user's expectations. For example, the system can use the parsed key elements as the main component and the user profile data as supplementary and constraining conditions to jointly construct prompt words; or, the system can use a weighted fusion approach, assigning different weights to the key elements and the user profile based on their clarity and importance, and then integrating them into a unified semantic representation.

[0081] The prompts generated from the user's active instructions refer to transforming deeply fused user-specified key elements and user profile data into clear and specific text instructions. These instructions guide a large language model to generate a personalized story that aligns with the user's active intent. For example, the fused information can be filled into a preset template; alternatively, the system can utilize more advanced natural language generation techniques to transform the fused data into more expressive and guiding prompts.

[0082] This application's solution differentiates the triggering methods of the story generation process, thereby employing different data acquisition and fusion strategies to generate more targeted story prompts. Specifically, the story generation process is triggered when the system receives an active story generation instruction or when the current time reaches the preset time for timed story generation. If it is timed, the system will control the default reading of a preset user profile and acquire real-time contextual data, including the date and weather conditions. This data is integrated to form timed story generation prompts, ensuring that the story content can naturally integrate into the user's daily life context, such as telling a relevant story based on the day's weather or generating a celebratory story for a specific date. This method provides continuous personalized content without requiring active user input.

[0083] On the other hand, when the story generation process is triggered by an active story generation command, the system of this invention first parses the user command, extracting key elements such as story theme, style, and / or characters. Simultaneously, it acquires preset user profile data corresponding to the current user. Subsequently, these parsed key elements are deeply integrated with the user profile data. This integration strategy focuses on combining the user's immediate needs (expressed through commands) with the user's long-term preferences (reflected through profiles), thereby generating a highly customized story generation prompt for the active command. For example, if a user actively requests an "adventure" story, the system of this embodiment will adjust the story details based on the user's "favorite animals" or "fears" in the user profile to better suit the user's personality. Through these two different data acquisition and integration paths, the solution of this application can flexibly adjust the prompt generation strategy according to the source of the story trigger. Timed trigger story prompts focus more on contextualization and automation, while active command story prompts emphasize the combination of user intent and personalized preferences. This differentiated processing enables large language models to obtain more accurate and richer guiding information when generating stories, thus effectively solving the problems of insufficiently accurate prompts or insufficient personalization of stories under a single fusion strategy, and significantly improving the quality of story generation and user experience.

[0084] The following example illustrates this. Suppose there is a child user named "Xiao Le," whose user profile records the name "Xiao Le," age "5 years old," favorite animal "puppy," favorite color "blue," and fear of "thunder." When the system detects that the current time has reached the preset bedtime story time (e.g., 9 PM), and the weather forecast shows "sunny," the system will initiate a timed story generation process. At this time, the system will by default read Xiao Le's user profile and obtain the current date and "sunny" weather. Subsequently, this information will be integrated to form a timed story generation prompt, such as: "Please generate a bedtime story for 5-year-old Xiao Le about a blue puppy's adventure on a sunny night, without any thunder in the story." In another scenario, if Xiaole's parents proactively issue a voice command: "Please generate a story about a brave firefighter, with a positive and uplifting style, and the protagonist named Xiaole," then the system in this embodiment of the invention will parse this proactive story generation command and extract key elements: the theme "firefighter," the style "positive and uplifting," and the character "Xiaole." Simultaneously, the system will obtain Xiaole's preset user profile. Then, these key elements will be deeply integrated with Xiaole's user profile data. For example, the system will combine the "firefighter" theme with Xiaole's preference for "loving puppies" to generate a story about a firefighter rescuing a puppy, ensuring that the story's tone aligns with "positive and uplifting" and Xiaole's age characteristics. Finally, a prompt for proactive story generation is generated, such as: "Please generate a positive and uplifting story for 5-year-old Xiaole about a brave firefighter rescuing a blue puppy, avoiding scenes of thunder in the story." Through the aforementioned technical solution, this application can generate story prompts using differentiated data acquisition and fusion strategies based on the type of story trigger (timed trigger or active command trigger). This allows timed stories to better integrate into real-time contexts, such as automatically adjusting story content based on date and weather, enhancing the story's immersion and timeliness. Simultaneously, for stories triggered by active commands, the system can more accurately combine key elements of user commands with the user's personalized profile, ensuring that the generated story not only meets the user's immediate needs but also deeply aligns with their long-term preferences, thereby significantly improving the story's personalization and user satisfaction. This refined prompt generation mechanism effectively solves the problem of story content not adequately matching user needs or contexts under a single fusion strategy, enabling large-scale language models to generate more attractive and relevant original stories.

[0085] In some of the embodiments described above in this application, a personalized story generation and broadcasting method based on intelligent agents is proposed. This method can dynamically generate original stories based on user profiles and real-time context. However, in its implementation, directly broadcasting the original stories generated by a large language model has problems such as unsuitable content for children, monotonous broadcasting format, and lack of immersion, thereby affecting the educational value of the story and the user experience.

[0086] In response, this application further proposes that the steps for immersive broadcasting of the generated original story through voice and / or visual means include: acquiring the generated original story; reviewing and filtering the content of the generated original story through a preset safety filter, and outputting an original story that meets the standards of children's safety and values; processing the filtered and reviewed original story into a tone and rhythm suitable for bedtime stories through voice synthesis technology; and / or parsing the filtered and reviewed original story, and dynamically switching the corresponding background images or animation scenes according to the key plots of the parsed story text to conduct immersive broadcasting in visual form.

[0087] Specifically, acquiring the generated original story refers to the personalized story content in text form, received from the large language model output by the story generation engine, which serves as the raw input for subsequent processing and broadcasting. For example, the complete text string can be received from the story generation engine through an API interface or internal data transfer mechanism; alternatively, after the story generation engine completes the generation, it stores the story text in a temporary storage area, from which the broadcasting module reads it.

[0088] By using pre-set safety filters to review and filter the content of generated original stories, the output of original stories that meet children's safety and values ​​standards is produced. This refers to using a content review mechanism to identify, remove, or modify content inappropriate for children, ensuring that the broadcasted stories are healthy, positive, and in line with children's cognitive level and moral standards, thus avoiding negative impacts on children from harmful information. For example, safety filters can use technologies such as keyword matching, sensitive word database comparison, and semantic analysis to identify potentially illegal or inappropriate content in the story text and replace, delete, or mark it; alternatively, safety filters can integrate machine learning models to learn from a large corpus of children's stories, determine whether the overall emotional tone and theme of the story are positive and uplifting, and rewrite or reject stories that do not meet the standards.

[0089] Original stories that have passed filtering and review are processed using speech synthesis technology to be read in a tone and rhythm suitable for bedtime stories. This involves using text-to-speech (TTS) technology to convert the reviewed story text into audio, and optimizing the tone, speed, and timbre of the audio to make it more suitable as a bedtime story. This provides a gentle, soothing, and emotionally resonant auditory experience, helping children relax and fall asleep. For example, speech synthesis technology can use deep learning models to generate speech with specific emotions (such as calm and gentle) and speeds (such as slow) by adjusting parameters; or, multiple preset timbre templates and tone patterns for bedtime stories can be used, selecting the most suitable template during synthesis and making fine adjustments based on the story content, such as appropriately slowing down the speed or increasing the volume at key points.

[0090] This involves analyzing filtered and approved original stories, dynamically switching corresponding background images or animated scenes based on the key plot points of the analyzed story text, and providing an immersive visual presentation. This refers to semantic analysis of the story text, extracting key information such as core events, character actions, and scene descriptions, and matching and displaying corresponding visual elements. Through the combination of visual and auditory elements, the immersiveness and appeal of the story are enhanced, helping children better understand and imagine the content. For example, the story analysis module can utilize Natural Language Processing (NLP) technology to identify entities (characters, locations, objects), actions, and emotions in the story, and then map this information to corresponding background images or animated clips using a pre-set visual material library. Alternatively, rule-based or machine learning methods can be used to segment the story text and generate a visual tag for each segment or key sentence. The system then retrieves and plays corresponding visual content from the material library based on these tags, achieving a smooth transition between visual scenes.

[0091] The solution in this application, after receiving a story generation instruction and generating an original story that integrates user personal characteristics and real-time context, first acquires the generated original story to ensure the suitability of the broadcast content and the immersive experience of the broadcast format. Subsequently, the story text is sent to a preset safety filter for rigorous content review. The safety filter reviews the story content based on preset child safety and values ​​standards, identifying and processing any inappropriate elements, thereby outputting a purified story version that meets children's reading standards. This filtering process is a crucial step in ensuring the healthy and positive nature of the story content, avoiding uncontrollable or inappropriate content generated by large language models. Next, the original story, after passing the filter review, enters the broadcast stage. To provide an immersive broadcast experience, the solution of this embodiment provides two or a combination of broadcast formats. In terms of voice broadcasting, the reviewed story text is converted into audio using speech synthesis technology. During this process, the speech synthesis parameters are specially adjusted to generate a tone and rhythm suitable for bedtime stories, such as using a soft timbre, slow speed, and soothing intonation, aiming to create a calm and warm auditory environment to help children relax. In terms of visual broadcasting, the system of this invention will perform in-depth analysis of the reviewed story text to identify key plot points, scene changes and character actions in the story.

[0092] Based on this extracted key information, the system dynamically selects and switches corresponding background images or animated scenes from the visual material library. For example, when the story describes a forest scene, a forest background will be displayed on the screen; when a character performs a certain action, a corresponding animation will play. This organic combination of auditory and visual elements makes the story presentation more vivid and engaging, greatly enhancing children's sense of immersion and comprehension. Through this mechanism, this solution not only solves the problem of inappropriate content that may result from directly broadcasting AI-generated stories, but also transforms the original text story into a multi-sensory, immersive experience through refined voice processing and dynamic visual presentation. This makes the broadcasting of personalized stories more in line with children's cognitive characteristics and emotional needs, effectively enhancing the story's appeal and educational value, thus providing a more complete and humanized user experience on top of the basic story generation function.

[0093] The following is a specific example to illustrate this. When the system generates an original story about "Little Rabbit Lele searching for colorful mushrooms in the forest," the story text is first acquired and sent to a security filter. This filter contains a sensitive word library, such as "dangerous" and "bad guy," and uses natural language processing technology to analyze the overall emotional tone of the story. If the story contains a scene like "Lele accidentally falls into a trap," the filter in this embodiment will modify it to "Lele accidentally took a small step but quickly regained his footing," to avoid unnecessary fright. The reviewed story text, such as "Little Rabbit Lele hops and skips in the sunny forest, searching for colorful mushrooms," will be sent to the speech synthesis module. This module selects a preset "gentle female voice" tone, adjusts the speech rate to approximately 120 words per minute, and slightly raises the tone of words like "hops and skips" to make it sound more lively. A calm tone is used at the end of the story to create a bedtime story atmosphere. Simultaneously, to provide visual immersion, the story text is parsed. When the story is interpreted as "in a sunny forest," the invention can dynamically switch to displaying a background image of sunlight filtering through leaves and dappling the forest floor. When the story is interpreted as "hopping and skipping in search of colorful mushrooms," an animated clip of a rabbit jumping across the grass and finding mushrooms can be played. When the storyline changes to "Lele meets a friendly little squirrel," the screen switches to an animated scene of the squirrel interacting with the rabbit. In this way, children can not only listen to the story but also visually experience the story content, thus gaining a richer and more immersive story experience.

[0094] Through the aforementioned technical solution, this application effectively addresses the potential uncontrollability of large-scale language models when generating stories, avoiding the output of content unsuitable for children and ensuring the health and safety of the story content. Simultaneously, by refining the processing of voice intonation and rhythm, and dynamically switching visual scenes according to the storyline, the immersiveness and appeal of the storytelling are greatly enhanced. This not only provides children with a more comfortable and enjoyable experience while listening to stories, helping them relax and promote better sleep, but also deepens their understanding and memory of the story content through the combination of audiovisual elements, stimulating their imagination, thereby significantly improving the overall quality and user satisfaction of personalized storytelling.

[0095] In some of the embodiments described above in this application, a personalized story generation and broadcasting method based on intelligent agents is proposed, which can dynamically generate and broadcast original stories according to user profiles and real-time context. However, after the story broadcast is completed, this method lacks a further interaction mechanism with the user and cannot obtain the user's immediate feedback on the current story. This limits the ability of the story generation system to continuously learn and optimize user preferences, which may lead to a gradual decrease in the relevance of subsequently generated stories to user interests.

[0096] In response, this application further proposes to proactively ask questions based on preset questions after the generated original story broadcast ends; to receive user feedback on the proactive questions; and to optimize the next story generation based on the feedback.

[0097] In this embodiment, after the generated original story broadcast ends, the system actively asks pre-set questions. This step aims to proactively guide user interaction after the story broadcast to collect immediate user feedback. This can be achieved in various ways. For example, the system can automatically play a pre-set voice question, such as "Did you like this story?" or "Which character in the story do you like the most?"; or, on devices with a visual interface, the system can display pre-set text questions on the screen and provide options or input boxes for the user to answer.

[0098] The step of receiving user feedback on proactively asked questions is used to obtain user evaluation or preference data regarding the story content. Specifically, the system of the present invention can use speech recognition technology to receive the user's verbal answers and convert them into processable text information; or, it can receive preset options selected by the user through interactive interfaces such as touch screens and buttons, or receive text information input by the user.

[0099] The core of the step of optimizing the next story generation based on the feedback information lies in using the collected user feedback data to adjust and improve the story generation strategy, thereby enhancing the personalization and appeal of subsequent stories. For example, the system in this embodiment can update the user's profile data with the user's preferences for story themes, styles, characters, or plots reflected in the feedback information, or adjust them as weight parameters of the story generation engine; alternatively, the system can analyze the feedback information, identify story elements that users dislike, and avoid or reduce the occurrence of these elements in subsequent story generation processes.

[0100] In the aforementioned agent-based personalized story generation and broadcasting method, after dynamically generating an original story that integrates user's personal characteristics and real-time context using a large language model, and then broadcasting it immersively through voice and / or visual means, this application further introduces a user feedback and optimization mechanism. Specifically, the system in this embodiment of the invention no longer simply ends the broadcasting process, but actively raises questions to the user based on preset questions. This aims to break the one-way broadcasting mode and establish two-way interaction with the user. Subsequently, the system receives feedback information provided by the user in response to these proactive questions. This feedback information directly reflects the user's likes, preferences, or dissatisfactions with the current story. This feedback information is not isolated but is further utilized by the system as an important basis for optimizing the next story generation. Through this closed-loop feedback mechanism, the system can continuously learn users' deep preferences, such as their inclination towards specific story themes, styles, characters, or plot developments. This allows the system to adjust the story generation prompts or directly influence the generation strategy of the large language model during subsequent story generation, enabling newly generated stories to better match users' personalized needs and real-time contexts. This effectively avoids the problem of repetitive story content or disconnection from user interests, significantly improving user experience and story appeal.

[0101] The following is a specific example. For instance, after an original story about "The Little Rabbit and Its Forest Adventure" is narrated via a smart speaker, the system of this embodiment can automatically play a voice question: "What do you think was the most interesting thing the little rabbit encountered in the forest?" or "Do you like the little rabbit or the little squirrel in the forest?" The user can answer by voice, for example, saying, "I like the little rabbit, it's very brave!" or "I think the flowers in the forest are beautiful." The system uses speech recognition technology to convert the user's answer into text and analyzes the keywords and sentiment. For example, if the user repeatedly expresses a liking for the theme of "bravery" or a preference for "forest animals," this information will be recorded and updated in the user's personalized data. Thus, in the next story generation, the system of this embodiment will prioritize story elements containing the theme of "bravery" or featuring "forest animals" as protagonists, thereby generating a new story that better suits the user's preferences, such as "The Brave Little Squirrel Searches for Treasure in the Forest."

[0102] As can be seen, through the above technical solution, this application can proactively interact with users and collect their real-time feedback after the story broadcast ends. This feedback is effectively utilized to form a continuously optimizing closed loop, enabling subsequent stories to more accurately match users' personalized preferences and real-time contexts. This not only significantly enhances the story's appeal and user immersion but also avoids a decline in experience due to a disconnect between story content and user interests, thus achieving intelligent and personalized iteration of story generation and broadcasting.

[0103] The present invention will be further described in detail below through another specific application embodiment: This specific application embodiment provides a personalized story generation and broadcasting method based on intelligent agents, such as... Figure 2 As shown, it includes the following steps: Step S11, Story generation process triggered; triggering methods include: 1) Timed trigger: This means that in this embodiment, the system automatically starts the story generation process at a preset time (such as 8 pm every night).

[0104] 2) Voice command trigger: The user issues a story generation command to the TV smart agent via voice (such as "Little TV, tell a story about the magic forest"). Then proceed to step S12.

[0105] Step S12: Extraction and fusion of generated elements.

[0106] In this specific embodiment, if it is a timed trigger, the system will read the user profile by default and combine it with contextual information such as the date and weather to form a story generation prompt, such as "Generate a heartwarming bedtime story containing [user name] and [user's favorite animal - puppy] adventuring in the forest of [current season - autumn]".

[0107] In this specific embodiment, if it is triggered by an instruction, the system parses the instruction and forcibly integrates the core information in the user profile (such as the user's name) into the instruction theme to form a prompt word, such as "Generate a story about [user's name]'s adventure in the universe, in which a planet [user's favorite color - blue] should appear", and then proceeds to step S13.

[0108] Step S13: AI story generation and review.

[0109] In this embodiment, the fused prompts are input into the story generation engine, which uses a large language model to generate story text. The system incorporates a built-in safety filter to review the generated content, ensuring it meets child safety and values ​​standards, before proceeding to step S14.

[0110] Step S14: Multimodal immersive broadcasting.

[0111] In this embodiment, a story can be generated using high-quality speech synthesis technology, with a tone and rhythm suitable for a bedtime story.

[0112] Simultaneously, corresponding background images or simple animated scenes can be dynamically switched on the TV screen based on key plot points of the story text, creating an immersive story atmosphere. Then proceed to step S15.

[0113] Step S15: Interaction and Continuation.

[0114] In this embodiment, after the story is read aloud, the system can proactively ask questions such as "Did you like this story?" or "Guess where [user name] will go on an adventure tomorrow?", and optimize the next story generation based on the user's voice feedback. In this embodiment, the system records the generated story text and audio, which can be played back by the user.

[0115] The present invention will be described in detail below through specific application scenarios and embodiments.

[0116] Example of a first application scenario: timed triggering; including the following steps: 1) When the smart terminal system clock reaches the set time, such as 8 pm (corresponding to step S11 above: timed trigger).

[0117] 2) The intelligent terminal system in this embodiment reads the image of the user "Lele", such as 5 years old, likes dinosaurs and strawberries, and it is raining outside today; and combined with the context, it generates the prompt words: "Generate a heartwarming adventure bedtime story containing Lele and a gentle Stegosaurus, who discover a mysterious cave indoors on a rainy evening." (corresponding to step S12 above).

[0118] 3) The AI ​​story generation engine generates an original story of about 500 words based on this prompt (corresponding to step S13 above).

[0119] 4) The TV screen dims, displays a dynamic background of the night sky, and begins to narrate the story in a gentle male voice: "Lele, it's time to sleep! Tonight's story is 'Lele and the Stegosaurus on a Rainy Day'..." (corresponding to step S14 above).

[0120] The second application scenario: command-triggered scenario, including the following steps: 11) The user says to the TV: "Generate an adventure story about a child in the universe." The smart terminal receives the user's voice command (corresponding to step S11 above: command triggering).

[0121] 12) The intelligent terminal system parses the instruction, the theme of which is "Cosmic Adventure", and forcibly incorporates the user profile "Beibei" (who likes blue). The prompt message is: "Generate a story about Beibei's cosmic adventure, in which blue planets and friendly alien friends appear." (Corresponding to step S12 above).

[0122] 13) AI generates stories (corresponding to step S3 above).

[0123] 14) The TV responds: "Okay, Beibei! This story is 'Beibei's Adventures on the Blue Planet'..." At the same time, the screen displays an animation of the universe and the blue planet (corresponding to step S4 above).

[0124] The user profile data described in this embodiment is configured and maintained by parents through the accompanying mobile app upon initial use. All generated story content can be viewed and replayed in the app's story history library.

[0125] In summary, compared with the prior art, the present invention has the following significant advantages: 1) The generated stories are highly personalized: The story content is deeply customized, incorporating children themselves and elements of their lives into the plot, which greatly enhances the story's sense of immersion, intimacy, and appeal.

[0126] 2) It has great content unlimitedness and innovation: Utilizing AI generation capabilities, it is theoretically possible to generate an unlimited number of non-repeating story contents, completely solving the problem of story resource depletion.

[0127] 3) Intelligent and convenient: It supports automatic generation at set times and can be seamlessly integrated into the daily rhythm of children's lives, providing parents with a worry-free and labor-saving parenting aid.

[0128] 4) Immersive multi-sensory experience: Combining the visual advantages of a large TV screen with high-quality voice broadcasting, it creates an immersive story experience that far surpasses traditional story machines, helping to cultivate children's listening skills and imagination.

[0129] 5) Promote interaction and expression: Through the question and feedback mechanism after the story, children are encouraged to interact with the system and exercise their language expression and thinking skills.

[0130] Exemplary device like Figure 3 As shown, this embodiment of the invention provides a personalized story generation and broadcasting device based on an intelligent agent, comprising: The story generation process trigger module 310 is used to control and trigger the story generation process when it receives an active story generation instruction or when the current time reaches the preset time for timed story generation. The timed trigger data acquisition module 320 is used to control the reading of preset user profiles and the acquisition of real-time context data when the story generation process is timed. The active trigger data acquisition module 330 is used to control the parsing of the active story generation instruction when the story generation process is triggered by the active story generation instruction, extract key elements including story theme, style and / or characters, and obtain the corresponding preset user profile; The data fusion and prompt word generation module 340 is used to deeply fuse the parsed key elements with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The original story generation module 350 is used to input the fused prompt words into the preset story generation engine and dynamically generate original stories that integrate user personal characteristics and real-time context using a large language model; The original story playback module 360 ​​is used to immerse the generated original story through voice and / or visual presentation, as described above.

[0131] Based on the above embodiments, the present invention also provides a smart terminal, the principle block diagram of which can be as follows: Figure 4 As shown, the smart terminal can be a smart large-screen TV or a smart computer. The smart terminal includes a processor, memory, network interface, display screen, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the smart terminal is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a personalized story generation and broadcasting method based on intelligent agents. The database of the smart terminal stores the personalized story generation and broadcasting program based on intelligent agents.

[0132] Those skilled in the art will understand that Figure 4 The block diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the smart terminal to which the present invention is applied. A specific smart terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0133] In one embodiment, a smart terminal is provided, including a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors. The one or more programs include instructions for performing the following operations: When a story generation instruction is received or the current time reaches the preset time for timed story generation, the story generation process is triggered. When the story generation process is triggered on a timer, it controls the reading of preset user profiles and the acquisition of real-time contextual data. When the story generation process is triggered by an active story generation instruction, the active story generation instruction is parsed, key elements including story theme, style and / or characters are extracted, and corresponding preset user profiles are obtained. The extracted key elements are deeply integrated with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The fused prompts are input into a preset story generation engine, which uses a large language model to dynamically generate original stories that blend user personal characteristics with real-time context. The generated original stories will be narrated immersively through voice and / or visual means, as described above.

[0134] The step of controlling the triggering of the story generation process before receiving an active story generation instruction or when the current time reaches the preset time for timed story generation includes: The system collects and obtains personalized data from at least one child user in advance. The personalized data includes: name, nickname, age, gender, favorite animal, color, food, hobbies, fears, and / or family member information.

[0135] The step of controlling the triggering of the story generation process before receiving an active story generation instruction or when the current time reaches the preset time for timed story generation includes: Preset the preset time for automatically starting the story generation process.

[0136] The aforementioned agent-based personalized story generation and broadcasting method, wherein the story generation process, which is triggered at a time, includes controlling the reading of a preset user profile and obtaining real-time contextual data, comprises: Real-time access to the family calendar to retrieve specified events, including birthdays, holidays, and / or travel plans; It can connect to smart home devices in real time and obtain environmental information.

[0137] The step of deeply integrating the parsed key elements with corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation includes: When the story generation process is triggered on a timed basis, the system reads a preset user profile by default to obtain real-time contextual data, including the date and weather conditions, and uses this data to generate prompts for the timed story generation. When the story generation process is triggered by an active story generation command, the system parses the active story generation command, extracts key elements including story theme, style and / or characters, and obtains the corresponding preset user profile; and deeply integrates the parsed key elements with the corresponding preset user profile data to obtain the prompt words for story generation of the active command.

[0138] The step of immersively narrating the generated original story through voice and / or visual means further includes: Obtain the generated original story; The generated original stories are reviewed and filtered through a preset safety filter to output original stories that meet children's safety and values ​​standards. The original stories that have been filtered and approved are processed using voice synthesis technology to be read in a tone and rhythm suitable for bedtime stories. And / or analyze the original stories that have been filtered and reviewed, and dynamically switch the corresponding background images or animated scenes based on the key plot points of the analyzed story text to provide an immersive visual presentation.

[0139] The step of immersively narrating the generated original story through voice and / or visual means further includes: When the generated original story broadcast ends, ask questions proactively according to the preset question prompts; Receive user feedback on proactively asked questions; The next story generation will be optimized based on the feedback information, as described above.

[0140] In other embodiments, this application proposes a computer-readable storage medium that, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to perform the above-described method.

[0141] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0142] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A personalized story generation and broadcasting method based on intelligent agents, characterized in that, include: When a story generation instruction is received or the current time reaches the preset time for timed story generation, the story generation process is triggered. When the story generation process is triggered on a timer, it controls the reading of preset user profiles and the acquisition of real-time contextual data. When the story generation process is triggered by an active story generation instruction, the active story generation instruction is parsed, key elements including story theme, style and / or characters are extracted, and corresponding preset user profiles are obtained. The extracted key elements are deeply integrated with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The fused prompts are input into a preset story generation engine, which uses a large language model to dynamically generate original stories that blend user personal characteristics with real-time context. The generated original stories will be narrated in an immersive manner through voice and / or visual means.

2. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, Before the step of controlling the triggering of the story generation process when a story generation instruction is received or the current time reaches the preset time for timed story generation, the following steps are included: The system collects and obtains personalized data from at least one child user in advance. The personalized data includes: name, nickname, age, gender, favorite animal, color, food, hobbies, fears, and / or family member information.

3. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, Before the step of controlling the triggering of the story generation process when a story generation instruction is received or the current time reaches the preset time for timed story generation, the following steps are included: Preset the preset time for automatically starting the story generation process.

4. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, The story generation process, which is triggered at a set time, includes controlling the reading of preset user profiles and obtaining real-time contextual data, including: Real-time access to the family calendar to retrieve specified events, including birthdays, holidays, and / or travel plans; It can connect to smart home devices in real time and obtain environmental information.

5. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, The step of deeply integrating the parsed key elements with corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation includes: When the story generation process is triggered on a timed basis, the system reads a preset user profile by default to obtain real-time contextual data, including the date and weather conditions, and uses this data to generate prompts for the timed story generation. When the story generation process is triggered by an active story generation command, the system parses the active story generation command, extracts key elements including story theme, style and / or characters, and obtains the corresponding preset user profile; and deeply integrates the parsed key elements with the corresponding preset user profile data to obtain the prompt words for story generation of the active command.

6. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, The step of immersively narrating the generated original story through voice and / or visual means also includes: Obtain the generated original story; The generated original stories are reviewed and filtered through a preset safety filter to output original stories that meet children's safety and values ​​standards. The original stories that have been filtered and approved are processed using voice synthesis technology to be read in a tone and rhythm suitable for bedtime stories. And / or analyze the original stories that have been filtered and reviewed, and dynamically switch the corresponding background images or animated scenes based on the key plot points of the analyzed story text to provide an immersive visual presentation.

7. The personalized story generation and broadcasting method based on intelligent agents according to claim 1, characterized in that, The step of immersively narrating the generated original story via voice and / or visual means also includes: When the generated original story broadcast ends, ask questions proactively according to the preset question prompts; Receive user feedback on proactively asked questions; The next story generation will be optimized based on the feedback information.

8. A personalized story generation and broadcasting device based on intelligent agents, characterized in that, The device includes: The story generation process trigger module is used to control and trigger the story generation process when a story generation instruction is received or the current time reaches the preset time for timed story generation. The timed trigger data acquisition module is used to control the reading of preset user profiles and the acquisition of real-time contextual data when the story generation process is timed. The active data acquisition module is used to control the parsing of the active story generation instruction when the story generation process is triggered by the active story generation instruction, extract key elements including story theme, style and / or characters, and obtain the corresponding preset user profile; The data fusion and prompt word generation module is used to deeply fuse the parsed key elements with the corresponding preset user profile data and / or real-time context data to obtain prompt words for story generation; The original story generation module is used to input the fused prompts into the preset story generation engine and dynamically generate original stories that integrate user personal characteristics and real-time context using a large language model; The original story playback module is used to immerse the generated original stories through voice and / or visual presentations.

9. A smart terminal, characterized in that, It includes a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, wherein the one or more programs include methods for performing any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is able to perform the method as described in any one of claims 1-7.

Images