Scenario simulation method and system based on on-board large language model

Abstract

The present disclosure provides a scenario simulation method and system based on an on-board large language model. The method comprises: on the basis of a simulation scenario type and a role selected by a user, performing simulation scenario construction and role setting, and generating a virtual dialogue scene; analyzing voice features and an emotional state of the user, and dynamically adjusting the simulation scenario construction and the role setting; on the basis of the voice features and the emotional state, selecting a voice synthesis role from an on-board voice database to represent the voice of a counterpart role; on the basis of the language and the expression content of the user, using a large language model to generate utterances and logic of the counterpart role; and on the basis of the voice synthesis role and the utterances and logic of the counterpart role and in combination with the voice features and the emotional state of the user obtained by real-time analysis, providing an instant scenario simulation feedback of the counterpart role.

Description

A Context Simulation Method and System Based on In-Vehicle Large Language Model

[0001] This application claims priority to Chinese Patent Application No. 202510358589.0, filed on March 25, 2025, entitled "A Context Simulation Method and System Based on an In-Vehicle Large Language Model", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This disclosure belongs to the field of in-vehicle artificial intelligence technology, and specifically relates to a scenario simulation method and system based on an in-vehicle large language model. Background Technology

[0003] With social development and intensifying workplace competition, people face increasing pressure in their work and life. Especially in complex situations such as arguments, interviews, and debates, users often find themselves at a disadvantage due to insufficient expression or poor emotional control. While in-vehicle systems possess some voice interaction capabilities, these systems primarily focus on basic functions such as entertainment, navigation, and voice assistants. Summary of the Invention

[0004] The technical solution adopted in this disclosure is as follows:

[0005] Firstly, this disclosure provides a scenario simulation method based on an in-vehicle large language model, including:

[0006] Obtain the user's command to start the simulation scenario and display the simulation scenario type generated based on the scenario model preset by the in-vehicle system;

[0007] Based on the user-selected simulation scenario type and role, the simulation scenario is constructed and the role is set, and a virtual dialogue scene is generated;

[0008] In virtual dialogue scenarios, based on the user's language and expression content, the user's voice characteristics and emotional state are analyzed to dynamically adjust the construction of the simulation scenario and the setting of the role; according to the voice characteristics and emotional state, a voice synthesis character is selected from the vehicle voice database to represent the voice of the other party.

[0009] Generate the other party's speech and logic based on the user's language and expression using a large language model;

[0010] Based on the speech and logic of the synthesized voice character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis, the system provides real-time situational simulation feedback to the other party character.

[0011] In some embodiments, obtaining the user's command to initiate the simulation scenario and displaying the simulation scenario type generated based on the scenario model preset by the in-vehicle system includes:

[0012] Users can select the simulation scenario type or describe a specific scenario using voice or touch.

[0013] The system uses NLP to analyze the user's selected simulation scenario type or the specific scenario described, and generates simulation scenarios based on a pre-set scenario model.

[0014] In some embodiments, the process of constructing a simulation scenario and setting roles based on the user-selected simulation scenario type and role, and generating a virtual dialogue scene, includes:

[0015] Obtain the user's selected simulation scenario type and role, construct the simulation scenario and set the role information. The simulation scenario includes the scenario type and difficulty; the role information includes the other party's personality traits, behavior patterns and speech style.

[0016] In some embodiments, the step of analyzing a user's voice characteristics and emotional state based on their language and expressions in a virtual dialogue scenario, and dynamically adjusting the simulation scenario construction and role setting, includes:

[0017] Real-time acquisition of users' language and expressions; NLP analysis of users' speech and expressions to identify user performance characteristics and thus obtain users' emotional state;

[0018] Based on emotional state analysis, the simulation scenario construction and role setting are dynamically adjusted;

[0019] The step of selecting a synthesized voice character from the vehicle's voice database to represent the other party's voice based on voice features and emotional state includes:

[0020] The system selects appropriate tone, emotional state, and corresponding voice synthesis characters from the vehicle's voice database to represent the other party's voice; it then adjusts the voice synthesis parameters based on the other party's words and emotional state to match the voice with the dialogue content.

[0021] In some embodiments, the generation of the other party's speech and logic based on the user's language and expressions using a large language model includes:

[0022] The system uses a pre-trained large language model to generate the speech and logic of the other party, and generates coherent and reasonable dialogue based on contextual information. The generated dialogue is then fine-tuned based on the user-defined personality traits and speech style of the character.

[0023] In some embodiments, providing real-time situational simulation feedback to the other party based on the speech and logic of the synthesized speech character and the other party character, combined with the user's speech characteristics and emotional state obtained through real-time analysis, includes:

[0024] Real-time analysis of users' expressive content and emotional state, including logical clarity, tone intensity, and emotional fluctuations;

[0025] Using natural language processing technology, based on the speech and logic of the synthesized voice character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis, virtual character feedback is generated using speech synthesis technology.

[0026] Through the interactive process, real-time situational simulation feedback of the other party's role is provided, enabling real-time interaction between the user and the virtual character;

[0027] Based on a feedback generation algorithm, it generates feedback suggestions according to the user's expression and emotional state, and provides real-time feedback and suggestions, including logical suggestions and tone adjustments; it also records the user's practice process, including dialogue content, emotional state, and feedback suggestions.

[0028] Secondly, this disclosure provides a context simulation system based on an in-vehicle large language model, comprising:

[0029] The acquisition module is used to acquire the user's command to start the simulation scenario and to display the simulation scenario type generated based on the scenario model preset by the vehicle system.

[0030] The scene generation module is used to construct simulated scenarios and set roles based on the simulated scenario type and role selected by the user, and generate virtual dialogue scenarios;

[0031] The voice synthesis module is used in virtual dialogue scenarios to analyze the user's voice features and emotional state based on the user's language and expression content, and dynamically adjust the construction of the simulation scenario and the setting of the role; according to the voice features and emotional state, it selects a voice synthesis character from the vehicle voice database to represent the voice of the other party.

[0032] The speech and logic generation module is used to generate the speech and logic of the other party based on the user's language and expression content using a large language model;

[0033] The scenario simulation module is used to provide real-time scenario simulation feedback to the other party based on the speech and logic of the speech synthesis character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis.

[0034] Thirdly, this disclosure provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the scenario simulation method based on an in-vehicle large language model.

[0035] Fourthly, this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the scenario simulation method based on an in-vehicle large language model.

[0036] Fifthly, this disclosure provides a computer program product, the computer program product including computer instructions, the computer instructions instructing a computer to execute the scenario simulation method based on an in-vehicle large language model. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this disclosure or related technologies, the accompanying drawings of the solutions in the embodiments of this disclosure or related technologies are described below. It should be understood that the accompanying drawings described below are merely for the convenience of clearly illustrating some embodiments of the technical solutions in this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0038] Figure 1 is a flowchart of a scenario simulation method based on an in-vehicle large language model according to an embodiment of this application;

[0039] Figure 2 is another flowchart illustrating the scenario simulation method based on an in-vehicle large language model according to an embodiment of this application;

[0040] Figure 3 illustrates a scenario simulation system based on an in-vehicle large language model according to an embodiment of this application;

[0041] Figure 4 is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0042] The embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this disclosure, and should not be construed as limiting this disclosure. The numbering of the steps in the following embodiments is only for the convenience of illustrating the embodiments. This disclosure does not limit the order between steps, and the execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0043] In the description of this disclosure, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly. Those skilled in the art can reasonably determine the specific meaning of these terms in this disclosure based on the specific content of the technical solution.

[0044] While in-vehicle systems possess some voice interaction capabilities, these systems primarily focus on basic functions such as entertainment, navigation, and voice assistants, lacking specialized training tools to help users handle complex interpersonal interactions. Technically, sentiment analysis is mainly applied in areas like social media analysis and customer service systems, typically analyzing emotional states in text or speech. However, these applications largely rely on non-real-time batch data processing, making it difficult to meet the demands of real-time, highly interactive scenarios. Although Natural Language Processing (NLP) has made significant progress in text generation and machine translation, it still faces numerous challenges in practical applications, particularly in simulating interpersonal interactions using sentiment analysis. In in-vehicle environments, NLP models are mainly used for navigation commands, entertainment searches, and basic conversational assistant functions, lacking the ability to deeply understand and handle complex situations. Furthermore, while speech recognition and synthesis technologies are relatively mature, in in-vehicle systems, they are primarily used for recognizing voice commands and simple conversational interactions.

[0045] In modern society, people often need to communicate effectively in complex social situations, such as arguments, interviews, and debates. However, a lack of effective training and preparation can lead to communication failures and even psychological stress. Existing in-vehicle systems mainly focus on entertainment, navigation, and voice assistant functions, lacking systems capable of simulating complex situations in the private environment of a car, resulting in low utilization of users' time in the vehicle.

[0046] Based on this, this application aims to develop a new technical solution that can organically combine natural language processing, sentiment analysis, speech recognition and synthesis technologies to form an integrated situation simulation system, especially for simulating complex situations such as disputes, interviews, and debates, to help users improve their expression and emotion management abilities in the car and enhance the vehicle's functionality.

[0047] This disclosure creates a comprehensive, multi-scenario situational simulation system by combining sentiment analysis and NLP technology. This system is integrated into a vehicle to help users conduct effective expression training and emotion management training in the private environment of the car.

[0048] This disclosure aims to help users improve their expressive abilities and emotional control in specific situations such as disputes, interviews, and debates, avoiding unfavorable situations caused by insufficient expressive ability or improper emotional control. This disclosure provides a situational simulation method and system based on an in-vehicle large language model. This method helps users improve their expressive abilities and emotional management in specific scenarios, allowing users to select simulated scenarios of interest from a preset scenario library for simulation training.

[0049] This disclosed system is used to simulate and analyze emotional processing and expression training in complex interpersonal scenarios such as disputes, interviews, and debates. The system aims to combine technologies such as Natural Language Processing (NLP), sentiment analysis, and speech recognition to help users improve their expressive and emotional management abilities in real-world interpersonal conflicts, interviews, and debates through in-car scenario simulations.

[0050] This disclosure provides a scenario simulation method based on an in-vehicle large language model, as shown in Figure 1, including:

[0051] S101: Obtain the user's command to start the simulation scenario and display the simulation scenario type generated based on the scenario model preset by the vehicle system; that is, when the command to start the simulation scenario is detected, control the display device to display the simulation scenario type generated by the preset scenario model.

[0052] S102, based on the simulation scenario type and role selected by the user, construct the simulation scenario and set the role, and generate a virtual dialogue scene; that is, when an instruction indicating that a simulation scenario type or role is selected is detected, construct the corresponding simulation scenario and set the corresponding role, and generate a virtual dialogue scene.

[0053] S103, in virtual dialogue scenarios, analyzes the user's voice characteristics and emotional state based on the user's language and expression content, and dynamically adjusts the construction of the simulation scenario and the setting of the role; according to the voice characteristics and emotional state, selects an appropriate voice synthesis role from the vehicle voice database to represent the voice of the other party.

[0054] S104, Based on the user's language and expression content, use a large language model to generate the other party's speech and logic;

[0055] S105, based on the speech and logic of the synthesized character and the generated counterpart, combined with real-time analysis of the user's voice characteristics and emotional state, provides immediate situational simulation feedback for the counterpart. In other words, based on the speech and logic of the synthesized character and the counterpart, combined with real-time analysis of the user's voice characteristics and emotional state, it provides immediate situational simulation feedback for the counterpart.

[0056] This disclosed scenario simulation method based on an in-vehicle large language model involves a series of steps to simulate real or fictional scenarios, helping users improve their expressive abilities and emotional management skills in specific situations. Furthermore, it allows users to select simulated scenarios of interest from a pre-set scenario library, effectively enriching vehicle functionality and increasing vehicle utilization. This method can be implemented by the vehicle's in-vehicle terminal or by a scenario simulation system integrated into the vehicle; the following description uses a scenario simulation system to illustrate the method.

[0057] In some embodiments, the system further constructs specific virtual dialogue scenarios and sets corresponding roles based on user selection. During the virtual dialogue, the system analyzes the user's voice characteristics and emotional state to identify their emotional responses in real time. Based on the identification results, the system selects a synthesized voice role from the voice database that matches the current situation and emotion to represent the other party's voice. Using a large language model, the system generates logical speech and dialogue content based on the user's identified emotional responses and the set other party's role (i.e., the system generates logical speech and dialogue content based on the identified user's emotional responses and the set other party's role information using a large language model). This helps to make the simulated situation more realistic and challenges the user's coping abilities. During the dialogue, the system continuously analyzes the user's expressions and emotions and adjusts the simulated situation and the other party's role's feedback as needed. This real-time feedback mechanism helps users better understand and adapt to communication needs in different situations.

[0058] By simulating different dialogue scenarios and role settings, users can hone their communication skills in various situations, including verbal expression, body language, and emotional management. Through continuous simulation of various situations, users can gradually enhance their adaptability and improve their decision-making abilities in emergency or complex situations. The method disclosed herein allows users to choose simulation scenarios and role settings according to their interests and needs, thus providing a personalized training experience. A real-time feedback mechanism helps users understand their performance and problems in a timely manner, and make adjustments and improvements as needed.

[0059] This disclosed scenario simulation system provides a private and safe training environment within a vehicle, helping users improve their communication skills and emotional control in complex situations such as arguments, interviews, and debates. Through repeated practice and system feedback, users can effectively improve their performance and reduce the psychological pressure associated with these scenarios. Furthermore, the system features intelligent debriefing and personalized optimization suggestions, enabling users to continuously improve their coping strategies and psychological resilience, ultimately achieving better performance in real-world situations.

[0060] The present disclosure will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0061] This disclosure provides a scenario simulation method and system based on an in-vehicle large language model, which can be mainly used for user expression training in complex social situations, such as disputes, interviews, and debates; including the following methods:

[0062] Natural Language Processing (NLP): The system utilizes advanced NLP techniques to generate dialogue content that matches the context selected by the user. The NLP model can not only understand the user's speech input but also automatically generate realistic dialogues based on the context, simulating complex real-world situations such as arguments, interviews, and debates.

[0063] Speech Recognition and Synthesis: The system employs high-precision speech recognition technology to capture user voice input in real time and generates natural and fluent dialogue feedback through speech synthesis technology. This process ensures that the interaction between the user and the system is highly real-time and natural.

[0064] Sentiment Analysis: The system incorporates sentiment analysis technology to assess the user's emotional state during the simulation in real time and provides contextual feedback based on the analysis results. This feature helps users better understand their emotional responses during practice and learn how to control their emotions in complex situations.

[0065] Scenario debriefing: After the simulation, the system generates a detailed debriefing report, including analysis of the user's expression skills, emotional control, and dialogue strategies. The system also provides personalized optimization suggestions based on the user's performance to help them perform better in similar future situations.

[0066] Multiple scenario selection: Users can choose from various simulated scenarios within the system, such as arguments, interviews, and debates. The system will generate corresponding dialogues and scenario settings based on the selected scenario, providing users with targeted training.

[0067] Personalized settings: Users can customize the difficulty and complexity of the simulated scenarios according to their individual needs. The system will automatically adjust the intensity of dialogue generation and emotional feedback to meet the training needs of different users.

[0068] This disclosed scenario simulation system provides a private and safe training environment within a vehicle, helping users improve their communication skills and emotional control in complex situations such as arguments, interviews, and debates. Through repeated practice and system feedback, users can effectively improve their performance and reduce the psychological pressure associated with these scenarios. Furthermore, the system features intelligent debriefing and personalized optimization suggestions, enabling users to continuously improve their coping strategies and psychological resilience, ultimately achieving better performance in real-world situations.

[0069] In a specific embodiment of this disclosure, as shown in Figure 2, situation simulation and emotion processing are achieved through the following steps:

[0070] Step 1: The user starts the system and selects a scenario.

[0071] The system acquires the user's command to initiate a simulated scenario and displays the simulated scenario type generated based on a preset scenario model of the in-vehicle system. In some embodiments, this process includes: the user initiating the system via voice or touch interface, selecting a simulated scenario type, or describing a specific scenario. The system analyzes the user's input using NLP and generates an appropriate simulated scenario based on the preset scenario model.

[0072] Startup method: Users can start the system via the in-vehicle system's voice assistant (e.g., "XX vehicle, start the scenario simulation system") or by clicking the icon on the touchscreen. The screen here refers to the in-vehicle terminal's screen.

[0073] Context Selection: The system provides a user-friendly interface listing preset context types (such as driving conflict, emergency rescue, daily communication, or arguments, interviews, debates, etc.) or allowing users to describe specific scenarios using natural language. The system's built-in NLP module parses the user input, matching the closest preset context or generating a custom context. In other words, the system can have a context selection interface listing multiple context types for direct user selection, as well as custom options. Users can trigger these custom options to describe the context via voice or text input; the system can then use its built-in NLP to parse the user's voice or text input, matching the closest preset context based on the parsing results, or automatically and intelligently defining a context based on the parsing results.

[0074] Context generation: Once a context is selected or intelligently defined, the system will generate an initial simulated scenario based on the preset context model (including scene description, role setting, dialogue template, etc.).

[0075] Step 2: Scenario Construction and Role Setting

[0076] The system constructs a simulated scenario and sets up roles based on the specific scenario type and role selected by the user, generating a virtual dialogue scene. In some embodiments, this process includes: the user recounting the actual scenario, and the system generating a virtual dialogue scene based on the user's input, including the other party's words and reactions. Users can customize the difficulty of the scenario and the personality traits of the other party to make it more realistic.

[0077] User repetition: Users can further refine the scene description through voice or text input, including specific locations, times, weather conditions, traffic conditions, etc.

[0078] Role setting: The system generates virtual dialogue scenarios based on user input and allows users to customize the other party's personality traits (such as friendly, hostile, neutral), behavior patterns (such as proactive, passive, provocative), and speech style (such as formal, casual, humorous).

[0079] Difficulty Adjustment: Users can also adjust the difficulty of the scene, such as increasing the complexity of the dialogue, increasing the degree of emotional conflict, or introducing additional interference factors (such as noise or distracting tasks).

[0080] Step 3: The sentiment analysis module analyzes the user's voice features and emotional state.

[0081] In virtual dialogue scenarios, the system analyzes users' voice characteristics and emotional states based on their language and expressions, dynamically adjusting the simulated scenario construction and role settings. In some embodiments, the system identifies user performance characteristics, such as logical clarity, tone intensity, and emotional fluctuations, by analyzing users' voices and expressions.

[0082] The system analyzes the user's voice characteristics (such as speech rate, tone, and emotional intensity) to determine the user's emotional state, such as anger, tension, and anxiety.

[0083] Based on emotional state analysis, the system can dynamically adjust the reactions of the simulated other party to make them more realistic, or help users improve their emotional expression and control abilities through practice.

[0084] Voice feature analysis: The system captures the user's voice signal through the microphone and uses voice processing technology to extract features (such as speech rate, pitch, volume, and emotional intensity).

[0085] Emotion recognition: Based on the extracted features, the system uses sentiment analysis algorithms (such as machine learning models or deep learning networks) to determine the user's emotional state (such as anger, tension, anxiety, calmness, etc.).

[0086] Dynamic adjustment: Based on the user's emotional state, the system can dynamically adjust the simulated opponent's reaction, such as increasing or decreasing provocative language, adjusting tone and attitude, to better adapt to the user's practice needs.

[0087] In some embodiments, the process of acquiring and analyzing user language and expressions in real time is implemented through speech recognition technology or NLP technology.

[0088] Speech recognition technology is used to convert speech signals into text. It is a technology in which machines recognize speech signals and convert them into text, and then understand the text and convert it into instructions.

[0089] NLP technology performs semantic analysis on text obtained through speech recognition or text directly input by the user to understand the user's intent and expressed emotions. Key information is extracted from the text for subsequent user sentiment analysis.

[0090] In some embodiments, methods for identifying user performance characteristics and emotional states include: emotion recognition algorithms and multimodal emotion analysis methods. The emotion recognition algorithm utilizes machine learning or deep learning algorithms, such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM), to perform sentiment analysis on the user's speech and text content, identifying the user's emotional state, such as happiness, sadness, or anger. Multimodal emotion analysis combines nonverbal information such as facial expressions, postures, and movements with verbal information such as speech and text to perform multimodal emotion analysis, improving the accuracy and robustness of emotion recognition.

[0091] During scenario construction, the simulated scenario is dynamically adjusted based on the user's emotional state. For example, when the user is happy, positive and interesting scenario elements can be added; when the user is sad, a warm and comforting scenario can be constructed. In some embodiments, the roles in the simulated scenario can be dynamically adjusted based on the user's emotional state and behavioral characteristics. For example, when the user shows a strong desire to explore, an adventurous role can be created; when the user shows negative emotions, a role that can provide comfort and support can be created.

[0092] During the simulated scenario, the user's emotional state and behavioral characteristics are continuously monitored, and real-time feedback and adjustments are made as needed. For example, when a user shows dissatisfaction or confusion, the scenario or role settings can be adjusted in a timely manner to better meet the user's needs and expectations.

[0093] Step 4: The large speech model generates the other party's speech and logic.

[0094] Based on the user's language and expression content, the system generates the speech and logic of the other party using a large language model. Specifically, the system generates the speech and logic of the other party using a large language model (such as a Generative Pre-trained Transformer, GPT).

[0095] Large Language Models: The system uses pre-trained large language models (such as the GPT series) to generate the speech and logic of the other party. These models can generate coherent and reasonable dialogue content based on contextual information.

[0096] Personalized customization: The system can also fine-tune the generated dialogue based on the user's character traits and speech style to better match the simulation effect expected by the user.

[0097] Step 5: The system's speech database synthesizes a speaker representing the other party's voice.

[0098] Based on speech characteristics and emotional state, an appropriate synthesized voice character is selected from an in-vehicle speech database to represent the voice of the other party. In some embodiments, this process includes: the system selecting appropriate tone, emotional state, and corresponding synthesized voice character from the speech database to represent the voice of the other party.

[0099] Voice database: The system has a rich voice database containing voice samples of different genders, ages, accents and speaking speeds.

[0100] Speech synthesis: Based on the user-defined characteristics and emotional state of the other party, the system selects the most suitable speech sample from the speech database and uses speech synthesis technology to generate a speaker representing the other party's voice.

[0101] Emotion matching: The system will also adjust the parameters of speech synthesis (such as tone, speech rate, volume, etc.) according to the other party's words and emotional state to ensure that the voice matches the content of the conversation.

[0102] Step 6: Scenario Simulation and Real-time Feedback

[0103] Based on the speech and logic of the synthesized and generated opposing characters, and combined with real-time analysis of the user's voice features and emotional state, the system provides immediate situational simulation feedback to the opposing character. In other words, based on the speech and logic of the synthesized and opposing characters, and combined with the user's voice features and emotional state obtained through real-time analysis, the system provides immediate situational simulation feedback to the opposing character. In some embodiments, this process includes: the system begins simulating a dispute scenario, and the user interacts with the virtual character through voice or text input. The system analyzes the user's expressed content and emotional state in real time, and combined with the user's voice features and emotional state obtained through real-time analysis, provides immediate situational simulation feedback to the opposing character, such as logical suggestions, tone adjustments, etc. (i.e., providing immediate situational simulation feedback to the opposing character based on the user's voice features and emotional state obtained through real-time analysis).

[0104] Scene simulation: The system begins to simulate arguments or other types of dialogue scenarios, and users interact with virtual characters through voice or text input.

[0105] Real-time analysis: The system analyzes the user's expression and emotional state in real time, including logical clarity, tone intensity, and emotional fluctuations.

[0106] Real-time feedback: Based on the analysis results, the system provides real-time feedback and suggestions, such as pointing out logical flaws and suggesting adjustments to tone or emotional expression. This feedback can be presented to the user through voice prompts, screen displays, or vibration.

[0107] Practice records: The system also records the user's practice process, including dialogue content, emotional state, feedback and suggestions, so that the user can review and evaluate their progress later.

[0108] The above plan, when implemented, includes the following:

[0109] Conflict Scenario Construction: The system constructs specific conflict scenarios based on user-selected contexts or custom descriptions. Scenarios should include key elements such as location, time, relationships between individuals, and the cause of the conflict. Conflict scenarios should possess a certain degree of complexity and diversity to simulate various situations in real life.

[0110] Virtual Character Design: Based on the conflict scenario, the system sets the personality traits, speech style, and behavior patterns of the virtual characters. Virtual characters should have distinct personalities and characteristics so that users can feel a sense of realism when interacting with them.

[0111] Interaction methods: Users can interact with the virtual character through voice or text input. The system should support multiple input methods to meet the needs of different users. The system should respond to user input in real time and generate corresponding dialogue content and feedback.

[0112] Content Analysis: The system uses Natural Language Processing (NLP) technology to analyze the user's expression, identifying key information and logical structure. The system should be able to determine whether the user's expression is clear and accurate, and point out logical flaws or illogical points.

[0113] Emotional state analysis: The system determines the user's emotional state by analyzing the user's voice characteristics (such as speech rate, tone, and volume) and text content. Emotional state analysis should be accurate and timely so that the system can adjust the virtual character's feedback and speech style accordingly.

[0114] Logical suggestions: When a user's expression contains logical flaws or illogical points, the system should provide specific logical suggestions to help the user improve their expression. These suggestions should be concise, clear, and easy for the user to understand and accept.

[0115] Tone Adjustment: The system should adjust the tone and style of its feedback based on the user's emotional state and the virtual character's settings. When the user exhibits negative emotions such as anger or tension, the system should provide appropriate reassurance and guidance to help the user alleviate these emotions.

[0116] Role-playing guidance: The system can provide users with role-playing guidance to help them better understand and imitate the personality traits and speech style of virtual characters. Role-playing guidance can include suggestions on language skills, emotional expression, and body language.

[0117] Real-time feedback presentation: The system should present real-time feedback in multiple ways, such as voice prompts, screen displays, and vibration. The presentation of feedback should be intuitive and clear so that users can obtain and understand the feedback content in a timely manner.

[0118] Utilizing advanced natural language processing technologies, such as semantic understanding and sentiment analysis, it achieves accurate analysis of user expressions and emotional states. High-quality speech synthesis technology is used to generate virtual character responses with natural and fluent speech.

[0119] By combining various interaction methods such as speech recognition and text input, real-time interaction between users and virtual characters can be achieved. An intelligent feedback generation algorithm is developed to generate targeted and practical feedback suggestions based on the user's expression and emotional state.

[0120] Step 7: Review and Optimization

[0121] After the simulation ends, the system generates a debriefing report, detailing the user's performance and providing optimization suggestions. Users can re-enter the simulation based on the report, trying different approaches until a satisfactory result is achieved. In other words, the debriefing report or related pages offer an option to re-enter the simulation; users can trigger this option to re-enter the simulation.

[0122] Step 8: Data Privacy and Management

[0123] The system ensures that all user data is processed locally. Users can choose to delete or modify historical scenario data, and the system can also dynamically adjust the settings of future simulation scenarios based on user choices.

[0124] Based on the above detailed description, this disclosure can realize scenario simulation in the vehicle environment, helping users to effectively manage and improve their expression ability and emotional control level in interpersonal disputes, thereby improving the functionality of the vehicle.

[0125] This disclosed scenario simulation method based on an in-vehicle large language model can simulate complex social situations and help users improve their expressive abilities and emotional control through real-time feedback and debriefing reports. It utilizes advanced NLP technology to generate dialogue content that matches the user's selected scenario, understands the user's voice input, and automatically generates realistic dialogue. High-precision speech recognition technology captures the user's voice input, and speech synthesis technology generates natural and fluent dialogue feedback. Combined with sentiment analysis technology, it assesses the user's emotional state in real time during the simulation and provides scenario feedback.

[0126] After the simulation, the system generates a detailed debriefing report, including analysis of the user's expressive abilities, emotional control, and dialogue strategies, and provides personalized optimization suggestions. Users can select from various simulation scenarios within the system, which will generate corresponding dialogues and scenario settings based on the selected scenario. Users can customize the difficulty and complexity of the simulation scenarios according to their individual needs, and the system will automatically adjust the intensity of dialogue generation and emotional feedback.

[0127] The second aspect of this disclosure provides a scenario simulation system based on an in-vehicle large language model. Figure 3 is a structural diagram of the system, which includes:

[0128] The acquisition module 100 is used to acquire the user's command to start the simulation scenario and to display the simulation scenario type generated based on the scenario model preset by the vehicle system.

[0129] The scene generation module 200 is used to construct simulation scenarios and set roles based on the simulation scenario type and role selected by the user, and generate virtual dialogue scenarios;

[0130] The voice synthesis module 300 is used in virtual dialogue scenarios to analyze the user's voice features and emotional state based on the user's language and expression content, and dynamically adjust the construction of the simulation scenario and the setting of the role; according to the voice features and emotional state, it selects an appropriate voice synthesis role from the vehicle voice database to represent the voice of the other party.

[0131] The speech and logic generation module 400 is used to generate the speech and logic of the other party based on the user's language and expression content using a large language model;

[0132] The context simulation module 500 is used to provide real-time context simulation feedback to the other party character based on the speech and logic of the synthesized character and the generated other party character, combined with real-time analysis of the user's voice features and emotional state. In other words, it provides real-time context simulation feedback to the other party character based on the speech and logic of the synthesized character and the other party character, combined with real-time analysis of the user's voice features and emotional state.

[0133] In some embodiments, the context simulation system based on the vehicle-mounted large language model includes a natural language processing module, a speech recognition and synthesis module, a sentiment analysis module, a scene review module, a multi-context selection module, and a personalization setting module.

[0134] The Natural Language Processing (NLP) module understands user voice input and automatically generates realistic dialogue. The Speech Recognition and Synthesis (SRS) module captures user voice input in real time and generates natural and fluent dialogue feedback through speech synthesis technology. The Sentiment Analysis (SSA) module assesses the user's emotional state during the simulation in real time and provides contextual feedback. The Scenario Review module generates a detailed review report after the simulation ends and provides personalized optimization suggestions. The Multi-Scenario Selection module allows users to choose from a variety of different simulation scenarios. The Personalization Settings module allows users to customize the difficulty and complexity of the simulation scenarios.

[0135] The system also includes a data privacy and management module to ensure that all user data is processed locally and to allow users to manage historical data and adjust future simulation settings.

[0136] This system utilizes key technologies such as natural language processing, speech recognition and synthesis, sentiment analysis, scenario deconstruction, multi-scenario selection, and personalized settings to simulate complex social situations, helping users improve their expressive abilities and emotional control. The system provides a private and secure training environment within the vehicle, enabling users to effectively improve their performance in complex situations, reduce psychological stress, and enhance coping strategies and mental resilience through repeated practice and system feedback.

[0137] This system is based on the aforementioned scenario simulation method using an in-vehicle large language model. Further details of this system can be found in the aforementioned method implementation examples.

[0138] As shown in Figure 4, a third aspect of this disclosure provides an electronic device including a memory 701, a processor 702, and a computer program stored in the memory 701 and executable on the processor. When the processor executes the computer program, it implements the scenario simulation method based on an in-vehicle large language model. The electronic device also includes a communication interface 703 and a bus 704.

[0139] A fourth aspect of this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the scenario simulation method based on an in-vehicle large language model.

[0140] A fifth aspect of this disclosure provides a computer program product including computer instructions that instruct a computer to execute the scenario simulation method based on an in-vehicle large language model.

[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0143] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, readable storage media, optical storage, etc.) containing computer-usable program code.

[0144] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more flowchart illustrations and / or one or more block diagrams.

[0145] Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort should fall within the scope of protection of this disclosure.

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this disclosure and not to limit them. Although this disclosure has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of this disclosure. Any modifications or equivalent substitutions that do not depart from the spirit and scope of this disclosure should be covered within the protection scope of the claims of this disclosure.

Claims

1. A scenario simulation method based on an in-vehicle large language model, comprising: Obtain the user's command to start the simulation scenario and display the simulation scenario type generated based on the scenario model preset by the in-vehicle system; Based on the user-selected simulation scenario type and role, the simulation scenario is constructed and the role is set, and a virtual dialogue scene is generated; In virtual dialogue scenarios, based on the user's language and expression content, the user's voice characteristics and emotional state are analyzed to dynamically adjust the construction of the simulation scenario and the setting of the role; according to the voice characteristics and emotional state, a voice synthesis character is selected from the vehicle voice database to represent the voice of the other party. Generate the other party's speech and logic based on the user's language and expression using a large language model; Based on the speech and logic of the synthesized voice character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis, the system provides real-time situational simulation feedback to the other party character.

2. The method according to claim 1, wherein, The process of obtaining the user's startup simulation scenario command and displaying the simulation scenario types generated based on the scenario model preset by the in-vehicle system includes: Users can select the simulation scenario type or describe a specific scenario using voice or touch. The system uses NLP to analyze the user's selected simulation scenario type or the specific scenario described, and generates simulation scenarios based on a pre-set scenario model.

3. The method according to claim 1, wherein, The process of constructing and setting simulation scenarios and roles based on user-selected simulation scenario types and roles, and generating virtual dialogue scenarios, includes: Obtain the user's selected simulation scenario type and role, construct the simulation scenario and set the role information. The simulation scenario includes the scenario type and difficulty; the role information includes the other party's personality traits, behavior patterns and speech style.

4. The method according to claim 1, wherein, In the virtual dialogue scenario, based on the user's language and expression content, the system analyzes the user's voice characteristics and emotional state, and dynamically adjusts the simulation scenario construction and role setting, including: Real-time acquisition of users' language and expressions; NLP analysis of users' speech and expressions to identify user performance characteristics and thus obtain users' emotional state; Based on emotional state analysis, the simulation scenario construction and role setting are dynamically adjusted; The step of selecting a synthesized voice character from the vehicle's voice database to represent the other party's voice based on voice features and emotional state includes: The system selects appropriate tone, emotional state, and corresponding voice synthesis characters from the vehicle's voice database to represent the other party's voice; it then adjusts the voice synthesis parameters based on the other party's words and emotional state to match the voice with the dialogue content.

5. The method according to claim 1, wherein, The generation of the other party's speech and logic based on the user's language and expressions using a large language model includes: The system uses a pre-trained large language model to generate the speech and logic of the other party, and generates coherent and reasonable dialogue based on contextual information. The generated dialogue is then fine-tuned based on the user-defined personality traits and speech style of the character.

6. The method according to claim 1, wherein, The method provides real-time situational simulation feedback to the other party based on the speech and logic of the synthesized voice character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis, including: Real-time analysis of users' expressive content and emotional state, including logical clarity, tone intensity, and emotional fluctuations; Using natural language processing technology, based on the speech and logic of the synthesized voice character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis, virtual character feedback is generated using speech synthesis technology. Through the interactive process, real-time situational simulation feedback of the other party's role is provided, enabling real-time interaction between the user and the virtual character; Based on a feedback generation algorithm, it generates feedback suggestions according to the user's expression and emotional state, and provides real-time feedback and suggestions, including logical suggestions and tone adjustments; it also records the user's practice process, including dialogue content, emotional state, and feedback suggestions.

7. A scenario simulation system based on an in-vehicle large language model, comprising: The acquisition module is used to acquire the user's command to start the simulation scenario and to display the simulation scenario type generated based on the scenario model preset by the vehicle system. The scene generation module is used to construct simulated scenarios and set roles based on the simulated scenario type and role selected by the user, and generate virtual dialogue scenarios; The voice synthesis module is used in virtual dialogue scenarios to analyze the user's voice features and emotional state based on the user's language and expression content, and dynamically adjust the construction of the simulation scenario and the setting of the role; according to the voice features and emotional state, it selects a voice synthesis character from the vehicle voice database to represent the voice of the other party. The speech and logic generation module is used to generate the speech and logic of the other party based on the user's language and expression content using a large language model; The scenario simulation module is used to provide real-time scenario simulation feedback to the other party based on the speech and logic of the speech synthesis character and the other party character, combined with the user's voice characteristics and emotional state obtained through real-time analysis.

8. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the scenario simulation method based on an in-vehicle large language model as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the scenario simulation method based on an in-vehicle large language model as described in any one of claims 1-6.

10. A computer program product comprising computer instructions, wherein, The computer instructions instruct the computer to execute the scenario simulation method based on the vehicle-mounted large language model as described in any one of claims 1-6.

Images