Method and device for generating virtual role of in-vehicle infotainment system, and electronic device

Abstract

The embodiment of the invention provides a method and device for generating a virtual role of a vehicle-mounted infotainment system and an electronic device, and the method comprises the steps: obtaining the role attribute information of a to-be-generated virtual role; generating a first virtual role according to the role attribute information by adopting a first role generation model; in response to the situation that the use data of the first virtual role meets the updating condition, updated role attribute information is obtained, and the use data of the first virtual role is used for reflecting the use frequency and satisfaction of the riding object on the first virtual role; performing model training on the first role generation model based on the updated role attribute information to obtain a second role generation model; and adopting a second role generation model to generate a second virtual role according to the updated role attribute information so as to perform voice interaction with the riding object by using the second virtual role. According to the invention, the technical problem that the vehicle voice assistant is fixed in image and single in type in the prior art is solved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of vehicle networking technology, and in particular to a method, device, and electronic device for generating a virtual character for a vehicle computer. Background Art

[0002] With the rapid development of automotive electronics, intelligent voice assistants are now widely installed in vehicles. However, the current in-vehicle voice assistants are factory-installed and have a relatively simple image, typically a cartoon character or a robot. Because these fixed voice assistants are monotonous and cannot be modified, user usage and user experience are generally low.

[0003] There is currently no effective solution to the above technical problems. Summary of the Invention

[0004] The embodiments of the present application provide a method, device, and electronic device for generating a virtual character for an in-vehicle computer, aiming to improve the technical problem in the related art that the image of an in-vehicle computer voice assistant is fixed and of a single type.

[0005] According to one aspect of an embodiment of the present invention, a method for generating a virtual character for a vehicle is provided, comprising: obtaining character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for performing voice interaction with a passenger in a vehicle, and the character attribute information is used for describing the appearance characteristics and expression characteristics of the to-be-generated virtual character; using a first character generation model to generate a first virtual character according to the character attribute information; in response to usage data of the first virtual character satisfying an update condition, obtaining updated character attribute information, wherein the usage data of the first virtual character is used for reflecting the frequency of use and satisfaction of the passenger with the first virtual character; performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model; and using the second character generation model to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passenger.

[0006] Based on the above technical solution as a whole, personalized generation and continuous optimization of the virtual character of the car computer are achieved through customized character attribute information and dynamic model updates. This not only enhances the interactive experience and emotional connection of the virtual character of the car computer, but also effectively improves user satisfaction and usage frequency.

[0007] Furthermore, obtaining the character attribute information of the virtual character to be generated includes: obtaining the character attribute information customized by the passenger; or, obtaining the data information of the passenger, wherein the data information is used to represent the appearance characteristics and preference characteristics of the passenger; determining the character attribute information based on the data information; or, obtaining the initial character attribute information customized by the passenger and the data information of the passenger; determining the character attribute information based on the initial character attribute information and the data information.

[0008] Based on the above technical solution, the user customizes the generation of personalized car-machine virtual characters or the system generates personalized car-machine virtual characters based on user attribute information, which effectively enhances the user experience, deepens the emotional connection, and increases the frequency of use and satisfaction.

[0009] Furthermore, using the first character generation model to generate the first virtual character according to the character attribute information includes: using the first character generation model to generate a third virtual character according to the character attribute information; adjusting the image of the third virtual character based on the image adjustment requirement to obtain the first virtual character, wherein the image adjustment requirement is the personalized adjustment requirement of the passenger for the appearance of the third virtual character.

[0010] Based on the above technical solution, by supporting the image of the car voice assistant to be personalized based on user needs, adapt to the environment, and dynamically adjust the image, the interaction quality and emotional connection strength between the car virtual character and the user are significantly improved, thereby improving the user experience.

[0011] Furthermore, the image of the first virtual character is adjusted based on the current environmental state of the vehicle and the adjustment time interval.

[0012] Based on the above technical solution, by adjusting the image of the virtual character to match the environmental state of the vehicle, the first virtual character can present an appearance that is more in line with the current situation, and by adjusting the image of the first virtual character based on a certain time interval, the display image of the first virtual character can be enriched.

[0013] Furthermore, in response to the usage data of the first virtual character satisfying the update conditions, obtaining updated character attribute information includes: obtaining the passenger's usage time, the number of times the first virtual character is used, the usage time of the first virtual character and the passenger's satisfaction; determining the usage frequency based on the passenger's usage time, the number of times the first virtual character is used and the usage time of the first virtual character, wherein the usage frequency is the frequency with which the passenger uses the first virtual character; determining a model training parameter adjustment value based on the usage frequency and the passenger's satisfaction; in response to the model training parameter adjustment value being greater than a preset threshold, obtaining the passenger's current attention information; updating the character attribute information based on the current attention information to obtain updated character attribute information.

[0014] Based on the above technical solution, through dynamic monitoring and intelligent adjustment, continuous personalization and intelligent optimization of the virtual role of the vehicle computer are achieved, which not only significantly improves the user's emotional experience and satisfaction, but also promotes the efficient use of vehicle computer system resources and the precise matching of service content.

[0015] Furthermore, the current attention information includes information related to the passenger's current attention area and information related to the passenger's customized area.

[0016] Based on the above technical solution, by integrating the user's current attention information and customized attention areas, it not only enhances the personalized service capabilities of the virtual character of the car computer, but also further deepens the emotional interaction between the user and the virtual character of the car computer, thereby improving the overall level of user experience.

[0017] Furthermore, the first character generation model is trained based on the updated character attribute information to obtain the second character generation model, including: adjusting the training parameters of the first character generation model based on the updated character attribute information to obtain the adjusted training parameters; and the first character generation model is trained based on the adjusted training parameters and the updated character attribute information to obtain the second character generation model.

[0018] Based on the above technical solution, by adjusting the model training parameters based on the updated role attribute information, and then generating a more optimized second role generation model based on the adjusted training parameters and the updated role attribute information, the adaptability and intelligence level of the role generation model are significantly enhanced, the service quality is improved, and efficient use of resources is achieved, while providing users with a continuous and personalized experience.

[0019] According to another aspect of an embodiment of the present invention, a device for generating a virtual character of a vehicle is also provided, including: a first acquisition module, used to obtain character attribute information of a virtual character to be generated, wherein the virtual character to be generated is used for voice interaction with passengers in the vehicle, and the character attribute information is used to describe the appearance characteristics and expression characteristics of the virtual character to be generated; a first generation module, used to use a first character generation model to generate a first virtual character according to the character attribute information; a second acquisition module, used to obtain updated character attribute information in response to usage data of the first virtual character meeting an update condition, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passengers with the first virtual character; a model training module, used to perform model training on the first character generation model based on the updated character attribute information to obtain a second character generation model; a second generation module, used to use the second character generation model to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passengers.

[0020] Furthermore, the first acquisition module is also used to obtain the customized character attribute information of the passenger; or, obtain the data information of the passenger, wherein the data information is used to represent the appearance characteristics and preference characteristics of the passenger; determine the character attribute information based on the data information; or, obtain the initial character attribute information customized by the passenger and the data information of the passenger; determine the character attribute information based on the initial character attribute information and the data information.

[0021] Furthermore, the first generation module is also used to use the first character generation model to generate a third virtual character according to the character attribute information; adjust the image of the third virtual character based on the image adjustment requirements to obtain the first virtual character, wherein the image adjustment requirements are the personalized adjustment requirements of the passenger for the appearance of the third virtual character.

[0022] Furthermore, the method for generating a vehicle-mounted virtual character further includes adjusting the image of the third virtual character based on the current environmental state of the vehicle and the adjustment time interval.

[0023] Furthermore, the second acquisition module is also used to obtain the passenger's use time, the number of times the first virtual character is used, the use time of the first virtual character and the passenger's satisfaction; determine the usage frequency based on the passenger's use time, the number of times the first virtual character is used and the use time of the first virtual character, wherein the usage frequency is the frequency with which the passenger uses the first virtual character; determine the model training parameter adjustment value based on the usage frequency and the passenger's satisfaction; in response to the model training parameter adjustment value being greater than a preset threshold, obtain the passenger's current attention information; update the character attribute information based on the current attention information to obtain the updated character attribute information.

[0024] Furthermore, the current attention information includes information related to the passenger's current attention area and information related to the passenger's customized area.

[0025] Furthermore, the model training module is also used to adjust the training parameters of the first character generation model based on the updated character attribute information to obtain adjusted training parameters; and perform model training on the first character generation model based on the adjusted training parameters and the updated character attribute information to obtain a second character generation model.

[0026] According to another aspect of an embodiment of the present invention, a vehicle is provided for executing any of the above methods for generating a vehicle-mounted virtual character.

[0027] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute any of the above-mentioned methods for generating a virtual character for a vehicle when running on a computer or a processor.

[0028] According to another aspect of an embodiment of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute any of the above methods for generating a vehicle-mounted virtual character.

[0029] In an embodiment of the present invention, a custom attribute and a dynamic update model are adopted to obtain character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for voice interaction with passengers in the vehicle, and the character attribute information is used to describe the appearance characteristics and expression characteristics of the to-be-generated virtual character; a first character generation model is adopted to generate a first virtual character according to the character attribute information; in response to usage data of the first virtual character satisfying an update condition, updated character attribute information is obtained, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passengers with the first virtual character; the first character generation model is trained based on the updated character attribute information to obtain a second character generation model; the second character generation model is adopted to generate a second virtual character according to the updated character attribute information, and the overall technical solution of using the second virtual character for voice interaction with passengers achieves the purpose of enhancing the emotional connection between the user and the vehicle-mounted voice assistant, increasing the usage rate of the voice assistant, and improving user satisfaction, thereby realizing the technical effect of image diversity and personalization of the vehicle-mounted voice assistant, and thus solving the technical problem of the fixed image and single type of the vehicle-mounted voice assistant in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of a method for generating a virtual character in a vehicle computer provided in one embodiment of the present application;

[0031] Figure 2 This is a flowchart of a system automatically generating a virtual character image provided by an embodiment of the present application;

[0032] Figure 3 This is a flowchart of a user manually generating a virtual character image provided by an embodiment of the present application;

[0033] Figure 4 This is a flow chart of a model training process for automatically generating a virtual character image by a system according to an embodiment of the present application;

[0034] Figure 5 This is a structural diagram of a device for generating a virtual character on a vehicle provided in an embodiment of the present application;

[0035] Figure 6 It is a structural diagram of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to make the technical problems, technical solutions and beneficial effects solved by this application more clearly understood, this application is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0037] In order to help those skilled in the art better understand the present technical solution, the following terminology is explained:

[0038] The image of the car voice assistant refers to the visual image in the voice assistant software interface installed inside the car. It can be an anthropomorphic character used to enhance the intuitiveness and emotional connection of human-computer interaction.

[0039] Character attribute information refers to a set of characteristic data related to the image of the car voice assistant, such as name, age, gender, language, clothing style, personality traits, etc., which is used to define and distinguish different virtual images.

[0040] The character generation model is an algorithm-based model used to generate virtual character images based on character attribute information. The character generation model is the basis for virtual character design and customization.

[0041] A virtual image refers to a character created through digital technology with a specific appearance and personality traits. It can be the image embodiment of a car voice assistant, interacting with users and providing services.

[0042] In the current technological landscape, automakers often pre-set one or several standardized voice assistant avatars when designing in-vehicle systems. These avatars are designed to appeal to a wide range of user groups, striving to be neutral, friendly, and recognizable. However, they also suffer from a high degree of homogeneity. While these fixed, lacking personalization can provide basic voice interaction services, they are limited in terms of deep emotional resonance and personalized service.

[0043] Users' use of voice assistants isn't limited to obtaining information or controlling vehicle functions; they also want to experience an interaction with them that's similar to communicating with a real person. However, current voice assistant images, due to their fixed and single nature, fail to meet users' diverse needs for voice assistant images in different contexts, such as age, gender, appearance, language style, and clothing. For example, a 40-year-old female user might want her voice assistant to resemble a close friend, while a 20-year-old male user might prefer his to resemble a like-minded friend.

[0044] Therefore, how to break the fixed pattern of traditional voice assistant image and provide a more personalized and emotional voice assistant image has become a key issue in improving user experience and increasing the usage rate of voice assistants.

[0045] An embodiment of the present application provides a method for generating a virtual character for a vehicle computer, comprising: obtaining character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for performing voice interaction with passengers in the vehicle, and the character attribute information is used to describe the appearance characteristics and expression characteristics of the to-be-generated virtual character; using a first character generation model to generate a first virtual character based on the character attribute information; in response to usage data of the first virtual character satisfying an update condition, obtaining updated character attribute information, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passengers with the first virtual character; performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model; and using the second character generation model to generate a second virtual character based on the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passengers.

[0046] Based on the above technical solution as a whole, personalized generation and continuous optimization of the virtual character of the car computer are achieved through customized character attribute information and dynamic model updates. This not only enhances the interactive experience and emotional connection of the virtual character of the car computer, but also effectively improves user satisfaction and usage frequency.

[0047] This application embodiment provides a method for generating a virtual character in a vehicle. Figure 1 , including the following steps:

[0048] Step S10: Acquiring character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for voice interaction with a passenger of the vehicle, and the character attribute information is used to describe the appearance and expression characteristics of the to-be-generated virtual character;

[0049] In the embodiment of the present invention, the virtual character to be generated can be understood as a digital image to be created in the vehicle system for voice interaction with passengers.

[0050] Character attribute information can be understood as a series of information used to describe the characteristics of a virtual character. For example, character attribute information may include, but is not limited to, name, age, gender, language style, clothing preferences, hobbies, personality traits, etc., and is used to guide the virtual character generation model to produce an image that matches the user's expectations. This information is not limited here.

[0051] The passenger can be understood as the main user who interacts with the virtual character through voice. For example, the passenger can include the driver and other passengers, which are not limited here.

[0052] The appearance characteristics of a virtual character can be understood as the visual appearance characteristics of the virtual character, including but not limited to gender, age, face shape, hairstyle, eye color, clothing style, body posture, etc.

[0053] The expression characteristics of a virtual character can be understood as other characteristics of the virtual character, including but not limited to the language, hobbies, personality, etc. of the virtual character.

[0054] Acquiring the character attribute information of the virtual character to be generated can be understood as acquiring the appearance characteristics and expression characteristics of the virtual character, aiming to achieve a high degree of personalization of the virtual character, making it an ideal reflection of the user's emotional and functional needs.

[0055] In an embodiment of the present invention, a virtual character is generated by collecting user preferences. The virtual character can interact with passengers through voice. The attribute information covers appearance and expression characteristics, such as gender, age, language style, etc., making the character personalized and capable of emotional communication, enhancing user experience and system interactivity, and improving user satisfaction and the humanized service level of the vehicle system.

[0056] Step S12: using the first character generation model to generate a first virtual character according to the character attribute information;

[0057] In this embodiment of the present invention, the first character generation model can be understood as a pre-trained machine learning model used to generate a preliminary image of a virtual character based on input character attribute information. The first character generation model is capable of understanding and transforming various character attribute information, converting this character attribute information into the visual image and behavioral characteristics of the virtual character. The first character generation model utilizes an algorithm to generate an initial virtual character image that matches the user's attributes, laying the foundation for subsequent personalized adjustments.

[0058] The first avatar can be understood as the initial appearance of the car voice assistant, customized according to the user's specific attributes and requirements. The first avatar reflects the user's basic preferences and serves as the foundation for subsequent personalization and optimization.

[0059] Using the first character generation model to generate the first virtual character according to the character attribute information can be understood as generating the first virtual character corresponding to the character attribute information based on the first character generation model and the character attribute information.

[0060] In an embodiment of the present invention, by adopting a first character generation model and generating a first virtual character based on character attribute information, it is intended to provide users with a more friendly, personalized and intelligent voice assistant image, thereby improving the satisfaction and experience of human-computer interaction.

[0061] Step S14: in response to the usage data of the first virtual character meeting the update condition, obtaining updated character attribute information, wherein the usage data of the first virtual character is used to reflect the passenger's usage frequency and satisfaction with the first virtual character;

[0062] In the embodiments of the present invention, the usage data of the first virtual character can be understood as the collected detailed records and feedback reflecting the passenger's interaction with the first virtual character. For example, the usage data of the first virtual character can include usage frequency, interaction duration, function usage, and user satisfaction, etc., which are not limited here.

[0063] In response to the usage data of the first virtual character meeting the update conditions, obtaining the updated character attribute information can be understood as the system judging whether the virtual character needs to be adjusted or upgraded based on the detailed data collected on the user's interaction with the first virtual character. If the judgment result indicates that the virtual character needs to be adjusted or upgraded, it is necessary to obtain the updated character attribute information to better adapt to user needs and improve user satisfaction.

[0064] In an embodiment of the present invention, the system determines whether the update conditions are met based on data such as the usage frequency and user satisfaction of the first virtual character. Once the conditions are met, new character attribute information is collected to optimize the virtual character, thereby ensuring that the virtual character can dynamically adapt to changes in user needs, improve the attractiveness and satisfaction of long-term interactions, and effectively enhance the stickiness between the user and the vehicle system and the personalized service experience.

[0065] Step S16: performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model;

[0066] In the embodiment of the present invention, the second character generation model can be understood as a model that is optimized and upgraded based on the first character generation model by using data feedback and learning.

[0067] The first character generation model is trained based on the updated character attribute information to produce the second character generation model. This means that the updated character attribute information is used to train or fine-tune the first character generation model. The first character generation model then generates a new virtual character image that better meets user expectations based on the updated attribute information. After training and optimization, the first character generation model is upgraded to the second character generation model, which can better understand and generate virtual characters that meet user preferences, thereby improving the level of personalization and user satisfaction of the virtual characters.

[0068] In an embodiment of the present invention, by training the first character generation model based on the updated character attribute information, the second character generation model obtained is an optimization model that performs self-optimization and upgrading based on user usage and feedback information. The purpose is to generate a more personalized virtual character that meets user needs in a timely manner, and improve the interaction quality and emotional connection between the user and the vehicle system.

[0069] Step S18: using the second character generation model to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passenger.

[0070] In the embodiment of the present invention, the second virtual character can be understood as a virtual character that is regenerated based on the optimized second character generation model and according to updated character attribute information.

[0071] Adopting the second character generation model to generate a second virtual character according to the updated character attribute information, and using the second virtual character to perform voice interaction with the passenger can be understood as generating a second virtual character that meets the user's latest preferences based on the second character generation model and the updated character attribute information, and using the second virtual character to perform voice interaction with the passenger, aiming to provide users with a constantly improving virtual partner, thereby enhancing the emotional connection and practicality of human-computer interaction.

[0072] In an embodiment of the present invention, an optimized second character generation model is used to generate a second virtual character that is more personalized and more adaptable to user needs based on updated character attribute information, so that the car voice assistant can dynamically adjust according to user usage habits and preferences, enhance the user's emotional connection and experience satisfaction, and effectively improve the quality of human-computer voice interaction.

[0073] In an embodiment of the present invention, a custom attribute and a dynamic update model are adopted to obtain character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for voice interaction with passengers in the vehicle, and the character attribute information is used to describe the appearance characteristics and expression characteristics of the to-be-generated virtual character; a first character generation model is adopted to generate a first virtual character according to the character attribute information; in response to usage data of the first virtual character satisfying an update condition, updated character attribute information is obtained, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passengers with the first virtual character; the first character generation model is trained based on the updated character attribute information to obtain a second character generation model; the second character generation model is adopted to generate a second virtual character according to the updated character attribute information, and the overall technical solution of using the second virtual character for voice interaction with passengers achieves the purpose of enhancing the emotional connection between the user and the vehicle-mounted voice assistant, increasing the usage rate of the voice assistant, and improving user satisfaction, thereby realizing the technical effect of image diversity and personalization of the vehicle-mounted voice assistant, and thus solving the technical problem of the fixed image and single type of the vehicle-mounted voice assistant in the related art.

[0074] Optionally, in step S10, obtaining the character attribute information of the virtual character to be generated includes the following execution steps:

[0075] Step S101, obtaining the customized character attribute information of the passenger; or

[0076] Step S102, obtaining data information of the passenger, wherein the data information is used to represent the appearance characteristics and preference characteristics of the passenger;

[0077] Step S103: determining the role attribute information based on the data information; or,

[0078] Step S104: Obtain the passenger's customized initial role attribute information and the passenger's data information; determine the role attribute information based on the initial role attribute information and the data information.

[0079] In this embodiment of the present invention, the passenger's appearance characteristics can be understood as specific features of the passenger's appearance, including but not limited to the passenger's age, gender, facial features, body shape, hairstyle, etc. The passenger's appearance characteristics are used to create a virtual character that resonates more emotionally with the passenger. For example, if the passenger is a middle-aged male, the system may generate a mature virtual character with corresponding appearance characteristics.

[0080] A passenger's preference profile can be understood as information about their behavior, preferences, and interests. This includes information about their favorite music genres, favorite radio stations, preferred communication styles (e.g., formal vs. humorous), frequently used apps, vehicle usage habits (e.g., long-distance driving vs. city commuting), and even their emotional state and behavioral patterns. These profiles can help the system understand the passenger's lifestyle and personality, thereby generating a virtual character whose functionality and personality more closely align with the user's needs.

[0081] Obtaining customized character attribute information of passengers can be understood as allowing passengers to actively input or select the characteristics they want the virtual character to have, including but not limited to appearance, voice style, personality traits, skill preferences, etc., thereby giving users a high degree of personalized choice and making the virtual character closer to the user's ideal image or functional needs.

[0082] For example, take a 40-year-old female user A as an example. A regards cars as her best friend. Based on personal preferences, the character attribute information of A's customized virtual character can be: name is Lao Lin, gender is female, virtual age is 41 years old, hair attribute is short black hair, skin is yellow, communication language is Chinese Chongqing dialect, clothing features are capable and concise, mainly brown (which can also be associated with the color of the car), likes to wear skirts, hobbies are photography and chatting, personality is calm and a bit verbose, and areas of expertise are fast speed and good sense of direction.

[0083] For example, let's take a 20-year-old male user B as an example. B regards cars as his friends. Based on personal preferences, the character attribute information of B's ​​customized virtual character can be: name is Lin, gender is male, virtual age is 21 years old, hair attribute is long brown hair, communication language is English, clothing features include liking to wear short sleeves and shorts, hobby is discussing topics related to basketball, personality is impatient but helpful, area of ​​expertise is basketball, and occupation is basketball player.

[0084] There are no restrictions on the character attribute information of the above-mentioned virtual characters, which can be set according to the preferences of the passengers. In addition, the character attribute information of the virtual characters can also be filled with the attributes of the birthplace. Based on the birthplace, the language attributes of the virtual characters can be switched to the dialect of the birthplace, which is not restricted here.

[0085] Obtaining data about passengers can be understood as the system automatically collecting and analyzing user behavior data, preferences, vehicle usage habits, and other information to achieve a more comprehensive understanding of the user's characteristics and needs. For example, with the user's consent, the system automatically collects personal information about User C, including: User C's gender is male, date of birth is March 2, 1985, vehicle usage habits are 7:00 AM to 8:00 AM for commuting and 6:00 PM to 7:00 PM for commuting, preferred listening content is hot news and Jacky Cheung's songs, and preferred listening language is Mandarin. These are not restricted here.

[0086] Determining character attributes based on data can be understood as the system utilizing collected user data, through algorithmic analysis and matching, to automatically determine the attributes that a virtual character should possess, in order to achieve the best fit for the user's needs. For example, based on the aforementioned personal information of User C, the generated voice avatar might be a typical 40-year-old, wearing glasses and a suit, who is born in the 1980s and enjoys discussing hot news. This is not a limitation.

[0087] Determining the role attribute information based on the initial role attribute information and the data information can be understood as also being able to simultaneously determine the role attribute information based on the initial role attribute information and the user's data information. By dynamically adjusting the role attribute information by integrating the initial role attributes with the user's data, the interactivity and emotional connection with the user can be enhanced, bringing the user a more satisfying and enjoyable riding experience.

[0088] Figure 2 This is a flowchart of a system automatically generating a virtual character image according to an embodiment of the present application. Figure 2As shown, with the user's consent, the system automatically collects user attributes such as gender, voice, age, preferences, and habits. Based on these attributes, it builds a model for the virtual character and recommends a virtual character image that is relevant to the user's usage data. For example, the information collection function can autonomously collect a large amount of past and current vehicle usage information from the user and create a virtual character image based on this data. The created virtual character image will also undergo large-scale model training, accumulating knowledge and developing expertise in areas of interest to the user, thereby reducing response time during subsequent user interactions. Furthermore, this can significantly reduce the likelihood of the virtual character's responses being inconsistent with facts, illogical, or out of context when interacting with the user. After the system automatically generates the virtual character image based on user data, the user can choose to help the virtual character make changes to its clothing and makeup, and can also set the virtual character to change its outfit on a daily or weekly basis, thereby enhancing the user's interactive experience with the virtual character.

[0089] In the embodiment of the present invention, through the above steps, not only can the passenger be allowed to directly customize the character attributes to ensure that the image is close to personal preferences, but a virtual image that conforms to the user's characteristics can also be generated based on the passenger's appearance characteristics and preference characteristic data, thereby providing the fun of human-computer interaction. Big data analysis capabilities can also be used to achieve more accurate image generation that is closer to the user's actual needs, effectively enhancing the emotional connection of human-computer interaction and significantly improving user experience and satisfaction.

[0090] Optionally, in step S12, using the first character generation model to generate the first virtual character according to the character attribute information includes the following execution steps:

[0091] Step S121, using the first character generation model to generate a third virtual character according to the character attribute information;

[0092] Step S122 : adjusting the image of the third virtual character based on the image adjustment requirement to obtain the first virtual character, wherein the image adjustment requirement is the passenger's personalized adjustment requirement for the appearance of the third virtual character.

[0093] In the embodiment of the present invention, the third virtual character can be understood as an initial, non-optimized virtual character image generated based on the character attribute information using the first character generation model.

[0094] Image adjustment requests can be understood as personalized modifications to the appearance of the generated third virtual character, allowing passengers to fine-tune the character's visual presentation to better suit their personal aesthetic or situational needs. For example, these image adjustment requests may include specific modifications to the character's hairstyle, clothing, accessories, expression, and posture, such as changing the hairstyle to a short one, choosing casual clothing instead of formal attire, adding specific accessories, or adjusting the character's expression and posture, though these are not limited here.

[0095] Using the first character generation model to generate a third virtual character based on character attribute information can be understood as using the first character generation model to generate an initial third virtual character that conforms to the user-defined character attribute information (such as age, gender, personality traits, voice type, preferences, etc.).

[0096] Adjusting the image of the third virtual character based on the image adjustment requirements to obtain the first virtual character can be understood as allowing the passenger to further put forward personalized requirements on the basis of the third virtual character, such as changing the hairstyle, adjusting the clothing, modifying the appearance details, etc., and fine-tuning the image of the third virtual character according to the image adjustment requirements, so as to obtain a first virtual character that better meets the user's aesthetics and needs.

[0097] Figure 3 This is a flowchart of a user manually generating a virtual character image provided by an embodiment of the present application, such as Figure 3 As shown, based on the attributes set by the user for the virtual character, the first character generation model can be used to generate an initial image (the third virtual character). The initial image has its own dressing room and wardrobe. The user can choose suitable clothes, shoes, necklaces, and other accessories for the voice assistant character in the predefined wardrobe according to their preferences, and choose the appropriate makeup in the dressing room to obtain the target image (the first virtual character).

[0098] In an embodiment of the present invention, a third virtual character can be generated through a first character generation model based on the character attribute information customized by the passenger, and the passenger can then make personalized image adjustments to the third virtual character to obtain the first virtual character. This allows the user to create a unique virtual assistant image based on personal preferences and emotional needs, greatly enhancing the user's sense of identification and satisfaction with the virtual character.

[0099] Optionally, the method for generating a vehicle-mounted virtual character further includes: adjusting the image of the first virtual character based on the current environmental state of the vehicle and an adjustment time interval.

[0100] In an embodiment of the present invention, adjusting the image of the third virtual character based on the current environmental state of the vehicle and the adjustment time interval can be understood as the virtual character automatically adjusting its image according to the actual environment of the vehicle (such as weather, geographical location, interior atmosphere, etc.) and the preset time interval to adapt to different external conditions and user situations.

[0101] For example, the virtual character image can also choose its own clothing and makeup according to the current season, vehicle location and the climate of the day. For example, if the vehicle is in the Northeast in winter, the virtual character will choose to appear in a thicker down jacket. When the car reaches Sanya in summer, the virtual character will wear beach pants and appear with the current weather data. There is no restriction here.

[0102] In an embodiment of the present invention, the third virtual character can also be automatically updated based on the current environmental status of the vehicle (such as weather, geographical location) and a preset adjustment time interval to obtain the first virtual character, thereby creating a highly satisfying and immersive first virtual character that can reflect the user's personality and flexibly respond to different environments, significantly improving the human-computer interaction experience.

[0103] In addition, virtual characters can grow. If the image chosen when the car first accompanies the user is a 20-year-old young person, 10 years later, the virtual character will be a middle-aged image, accompanying the user as he grows up, and will have a better understanding of the user, thereby strengthening the emotional connection between the car and the user.

[0104] Optionally, in step S14, in response to the usage data of the first virtual character meeting the update condition, obtaining updated character attribute information includes:

[0105] Step S141, obtaining the passenger's vehicle usage time, the number of times the first virtual character is used, the first virtual character's usage time, and the passenger's satisfaction;

[0106] Step S142, determining a usage frequency based on the passenger's vehicle usage time, the number of times the first virtual character is used, and the usage time of the first virtual character, wherein the usage frequency is the frequency with which the passenger uses the first virtual character;

[0107] Step S143, determining the model training parameter adjustment value based on the usage frequency and the passenger's satisfaction;

[0108] Step S144, in response to the model training parameter adjustment value being greater than a preset threshold, obtaining the current attention information of the passenger;

[0109] Step S145: Update the role attribute information based on the current attention information to obtain updated role attribute information.

[0110] In the embodiment of the present invention, the vehicle usage time of the passenger can be understood as the accumulated time of the passenger using the vehicle. For example, the vehicle usage time of the passenger can be expressed as T, which is not limited here.

[0111] The number of times the first virtual character is used can be understood as the total number of times the passenger interacts with the first virtual character. For example, the number of times the first virtual character is used can be expressed as N, which is not limited here.

[0112] The usage time of the first virtual character can be understood as the total time spent by the passenger during the interaction with the first virtual character. For example, the usage time of the first virtual character can be expressed as T1, which is not limited here.

[0113] The passenger's satisfaction can be understood as the result of the passenger's evaluation of their overall experience using the first virtual character. For example, the passenger's satisfaction can be represented by M, with a value range of 0-1. M can be obtained from the user satisfaction survey form of the vehicle computer and is not limited here.

[0114] The frequency of use can be understood as the intensiveness of the use of the virtual character calculated by combining the passenger's use time, the number of times the first virtual character is used, and the use time.

[0115] The preset threshold value can be understood as a pre-set standard value used to determine whether the model training parameters need to be adjusted. For example, the preset threshold value can be set to 1, which is not limited here.

[0116] Determining the usage frequency based on the vehicle usage time of the passenger, the number of times the first virtual character is used, and the usage time of the first virtual character can be understood as combining the above-collected vehicle usage time, the number of times the first virtual character is used, and the usage time to form a quantitative indicator that reflects the user's dependence on the virtual character and the density of use. For example, the user's usage frequency can be calculated according to the formula: usage frequency =

[0117] N*T1 / T is obtained, which is not limited here.

[0118] Determining the model training parameter adjustment value based on usage frequency and passenger satisfaction can be understood as calculating the adjustment range of the model training parameter using the two indicators of usage frequency and satisfaction. For example, the model training parameter adjustment value = usage frequency * M, which is not limited here.

[0119] In response to the model training parameter adjustment value being greater than a preset threshold, obtaining the current attention information of the passenger can be understood as, when the model training parameter adjustment value is less than the preset threshold, the system will start the update mechanism to mine the current focus or interest areas from the user's behavioral data. Exemplarily, within a cycle, if the model training parameter adjustment value is less than 1, the system starts the model training parameter adjustment. For example, when the news that the user cares about changes from entertainment to sports, the current model is no longer applicable, and user satisfaction, user usage frequency, and model training parameter adjustment value will all decrease. When the model training parameter adjustment value is less than 1, the system will start a new model training parameter adjustment, such as adding sports-related training content. In addition, users can also choose to trigger new model training in other fields (such as finance), which is not restricted here.

[0120] The character attribute information is updated based on the current attention information. The updated character attribute information can be understood as the system updating the character attribute information of the virtual character, such as personality traits, knowledge areas, image and dress, based on the current attention information of the passenger, to better cater to the preferences and needs of the passenger.

[0121] In the embodiment of the present invention, by timely collecting and analyzing user data, it is determined whether the current virtual character generation model is suitable for the user's current status, so that the user's preferences and needs can be accurately understood, thereby providing virtual character services that are more in line with the user's personality, significantly improving user experience and satisfaction.

[0122] Optionally, the above-mentioned current attention information includes relevant information of the passenger's current attention field and relevant information of the passenger's customized field.

[0123] In this embodiment of the present invention, the information related to the current focus area can be understood as the area of ​​focus that the passenger is currently or recently paying attention to. This information is typically derived from the passenger's behavioral data, search history, social media activity, or interaction records with other smart devices. For example, the information related to the current focus area can be the user's news interest changing from entertainment to sports, without limitation.

[0124] The relevant information of the custom field can be understood as the personalized preferences or specific field needs clearly specified by the user. For example, the relevant information of the custom field can be the field of finance selected by the user independently, which is not limited here.

[0125] In the embodiment of the present invention, by comprehensively applying the user's current attention information, the human-computer interaction experience between the user and the virtual character is significantly improved, thereby not only providing instant and personalized services, but also deepening emotional connections and promoting knowledge exchange.

[0126] Optionally, in step S16, performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model includes:

[0127] Step S161, adjusting the training parameters of the first character generation model based on the updated character attribute information to obtain adjusted training parameters;

[0128] Step S162: Perform model training on the first character generation model based on the adjusted training parameters and the updated character attribute information to obtain a second character generation model.

[0129] In an embodiment of the present invention, the training parameters of the first character generation model are adjusted based on the updated character attribute information. Obtaining the adjusted training parameters can be understood as, based on the collected updated character attribute information, the system will adjust the first character generation model to obtain the adjusted training parameters.

[0130] Training the first character generation model based on the adjusted training parameters and the updated character attribute information to obtain the second character generation model can be understood as updating and optimizing the first character generation model based on the adjusted training parameters and the updated character attribute information, thereby obtaining a second character generation model that is more in line with user needs.

[0131] In the embodiment of the present invention, through continuous model adjustment and training, the virtual character can dynamically adapt to user needs and provide a higher level of personalized services, thereby not only improving user satisfaction but also promoting the emotional connection between the virtual character and the user.

[0132] Figure 4 This is a flow chart of a model training for automatically generating a virtual character image by a system according to an embodiment of the present application. Figure 4As shown, the system first collects information data that the car owner is concerned about (for example, including news, sports, entertainment, etc.), and uploads the above-mentioned information data that the car owner is concerned about to the cloud. The cloud will collect specific relevant information (including newspapers, headlines, etc.) from the information data that the car owner is concerned about, and perform data processing and format conversion on the information data that the car owner is concerned about. Based on the processed information data that the car owner is concerned about, the character generation model is trained as a super-anthropomorphic model. Finally, based on the model output file (i.e., the passenger's car usage time, the number of times the first virtual character is used, the usage time of the first virtual character, and the passenger's satisfaction), the model is evaluated to determine whether the current character generation model needs to be adjusted in parameters. If the current character generation model needs to be adjusted in parameters, the current character generation model is continuously adjusted based on the updated character attribute information and the updated parameters until the preset conditions are met. The model file at this time (i.e., the passenger's car usage time, the number of times the first virtual character is used, the usage time of the first virtual character, and the passenger's satisfaction) is saved in the vehicle system for the next adjustment of the character generation model.

[0133] The present application also provides a device 500 for generating a virtual character on a vehicle. Figure 5 , including: a first acquisition module 501, used to obtain role attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used to perform voice interaction with passengers of the vehicle, and the role attribute information is used to describe the appearance characteristics and expression characteristics of the to-be-generated virtual character; a first generation module 502, used to generate a first virtual character according to the role attribute information using a first role generation model; a second acquisition module 503, used to obtain updated role attribute information in response to usage data of the first virtual character satisfying an update condition, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passengers with the first virtual character; a model training module 504, used to perform model training on the first role generation model based on the updated role attribute information to obtain a second role generation model; a second generation module 505, used to generate a second virtual character according to the updated role attribute information using the second role generation model, so as to use the second virtual character to perform voice interaction with passengers.

[0134] Furthermore, the first acquisition module 501 is also used to obtain the customized role attribute information of the passenger; or, obtain the data information of the passenger, wherein the data information is used to represent the appearance characteristics and preference characteristics of the passenger; determine the role attribute information based on the data information; or, obtain the initial role attribute information customized by the passenger and the data information of the passenger; determine the role attribute information based on the initial role attribute information and the data information.

[0135] Furthermore, the first generation module 502 is also used to use the first character generation model to generate a third virtual character according to the character attribute information; adjust the image of the third virtual character based on the image adjustment requirement to obtain the first virtual character, wherein the image adjustment requirement is the personalized adjustment requirement of the passenger for the appearance of the third virtual character.

[0136] Furthermore, the method for generating a vehicle-mounted virtual character further includes adjusting the image of the third virtual character based on the current environmental state of the vehicle and the adjustment time interval.

[0137] Furthermore, the second acquisition module 503 is also used to obtain the passenger's usage time, the number of times the first virtual character is used, the usage time of the first virtual character and the passenger's satisfaction; determine the usage frequency based on the passenger's usage time, the number of times the first virtual character is used and the usage time of the first virtual character, wherein the usage frequency is the frequency with which the passenger uses the first virtual character; determine the model training parameter adjustment value based on the usage frequency and the passenger's satisfaction; in response to the model training parameter adjustment value being greater than a preset threshold, obtain the passenger's current attention information; update the character attribute information based on the current attention information to obtain the updated character attribute information.

[0138] Furthermore, the current attention information includes information related to the passenger's current attention area and information related to the passenger's customized area.

[0139] Furthermore, the model training module 504 is also used to adjust the training parameters of the first character generation model based on the updated character attribute information to obtain adjusted training parameters; and perform model training on the first character generation model based on the adjusted training parameters and the updated character attribute information to obtain a second character generation model.

[0140] According to another aspect of an embodiment of the present invention, a vehicle is provided for executing any of the above methods for generating a vehicle-mounted virtual character.

[0141] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute any of the above-mentioned methods for generating a virtual character for a vehicle when running on a computer or a processor.

[0142] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0143] Step S10: Acquiring character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for voice interaction with a passenger of the vehicle, and the character attribute information is used to describe the appearance and expression characteristics of the to-be-generated virtual character;

[0144] Step S12: using the first character generation model to generate a first virtual character according to the character attribute information;

[0145] Step S14: in response to the usage data of the first virtual character meeting the update condition, obtaining updated character attribute information, wherein the usage data of the first virtual character is used to reflect the passenger's usage frequency and satisfaction with the first virtual character;

[0146] Step S16: performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model;

[0147] Step S18: using the second character generation model to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passenger.

[0148] Optionally, in this embodiment, the above-mentioned computer-readable storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0149] The present application also provides an electronic device 600. Figure 6 , including a memory 601 and a processor 602, wherein the memory 601 is used to store computer programs; the processor 602 is used to execute the programs stored in the memory 601 to implement the method for generating a vehicle-mounted virtual character introduced in any embodiment of the present application.

[0150] Optionally, in this embodiment, the processor in the electronic device may be configured to run a computer program to perform the following steps:

[0151] Step S10: Acquiring character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used for voice interaction with a passenger of the vehicle, and the character attribute information is used to describe the appearance and expression characteristics of the to-be-generated virtual character;

[0152] Step S12: using the first character generation model to generate a first virtual character according to the character attribute information;

[0153] Step S14: in response to the usage data of the first virtual character meeting the update condition, obtaining updated character attribute information, wherein the usage data of the first virtual character is used to reflect the passenger's usage frequency and satisfaction with the first virtual character;

[0154] Step S16: performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model;

[0155] Step S18: using the second character generation model to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passenger.

[0156] In this application, a plurality refers to two or more.

[0157] In this application, unless otherwise expressly defined, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they can refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediary; and internal communication between two components. A person of ordinary skill in the art will understand the specific meanings of these terms in this application.

[0158] In this application, the terms "first," "second," "third," "fourth," etc. (if any) are used to distinguish similar objects and are not necessarily used to describe a particular sequential order.

[0159] The term "and / or" in this application simply describes an association between related objects, indicating that three possible relationships exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this application generally indicates that the related objects are in an "or" relationship.

[0160] Unless otherwise specified, all steps of the present application may be performed sequentially or randomly. For example, a statement that the method includes steps A and B indicates that the method may include steps A and B performed sequentially, or steps B and A performed sequentially. For example, a statement that the method may also include step C indicates that step C may be added to the method in any order, for example, the method may include steps A, B, and C, or steps A, C, and B, or steps C, A, and B, etc.

[0161] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for generating a virtual character of a vehicle computer, characterized in that: include: Obtaining character attribute information of a virtual character to be generated, wherein the virtual character to be generated is used to perform voice interaction with a passenger of a vehicle, and the character attribute information is used to describe the appearance characteristics and expression characteristics of the virtual character to be generated; generating a first virtual character according to the character attribute information using a first character generation model; In response to the usage data of the first virtual character meeting an update condition, acquiring updated character attribute information, wherein the usage data of the first virtual character is used to reflect the passenger's usage frequency and satisfaction with the first virtual character; Performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model; The second character generation model is used to generate a second virtual character according to the updated character attribute information, so as to use the second virtual character to perform voice interaction with the passenger.

2. The method according to claim 1, characterized in that The step of obtaining the character attribute information of the virtual character to be generated includes: Obtaining the role attribute information customized by the passenger; or, Acquiring data information of the passenger, wherein the data information is used to represent the appearance characteristics and preference characteristics of the passenger; determining the character attribute information based on the data information; or Acquire the customized initial role attribute information of the passenger object and the data information of the passenger object; and determine the role attribute information based on the initial role attribute information and the data information.

3. The method according to claim 1 or 2, characterized in that The generating of the first virtual character according to the character attribute information using the first character generation model includes: generating a third virtual character according to the character attribute information using the first character generation model; The image of the third virtual character is adjusted based on the image adjustment requirement to obtain the first virtual character, wherein the image adjustment requirement is the personalized adjustment requirement of the passenger on the appearance of the third virtual character.

4. The method according to claim 3, characterized in that The method further comprises: The image of the first virtual character is adjusted based on the current environmental state of the vehicle and an adjustment time interval.

5. The method according to claim 1 or 2, characterized in that In response to the usage data of the first virtual character satisfying an update condition, obtaining the updated character attribute information includes: Obtaining the vehicle usage time of the passenger, the number of times the first virtual character is used, the usage time of the first virtual character, and the passenger's satisfaction; determining a usage frequency based on the vehicle usage time of the passenger, the number of times the first virtual character is used, and the usage time of the first virtual character, wherein the usage frequency is the frequency with which the passenger uses the first virtual character; determining a model training parameter adjustment value based on the usage frequency and the passenger's satisfaction; In response to the model training parameter adjustment value being greater than a preset threshold, obtaining current attention information of the passenger; The role attribute information is updated based on the current attention information to obtain the updated role attribute information.

6. The method according to claim 5, characterized in that The current attention information includes information related to the passenger's current attention area and information related to the passenger's customized area.

7. The method according to claim 1 or 2, characterized in that The performing model training on the first character generation model based on the updated character attribute information to obtain a second character generation model includes: Adjusting the training parameters of the first character generation model based on the updated character attribute information to obtain the adjusted training parameters; Model training is performed on the first character generation model based on the adjusted training parameters and the updated character attribute information to obtain the second character generation model.

8. A device for generating a virtual character of a vehicle computer, characterized in that: include: A first acquisition module is configured to acquire character attribute information of a to-be-generated virtual character, wherein the to-be-generated virtual character is used to perform voice interaction with a passenger in a vehicle, and the character attribute information is used to describe the appearance and expression characteristics of the to-be-generated virtual character; A first generating module, configured to generate a first virtual character according to the character attribute information using a first character generating model; a second acquisition module, configured to acquire updated character attribute information in response to the usage data of the first virtual character satisfying an update condition, wherein the usage data of the first virtual character is used to reflect the frequency of use and satisfaction of the passenger with the first virtual character; A model training module, configured to perform model training on the first character generation model based on the updated character attribute information to obtain a second character generation model; The second generation module is used to generate a second virtual character according to the updated character attribute information using the second character generation model, so as to use the second virtual character to perform voice interaction with the passenger.

9. A vehicle, characterized in that: The vehicle is used to execute the method for generating a vehicle-mounted virtual character as described in any one of claims 1 to 7 above.

10. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the method for generating a vehicle-mounted virtual character as described in any one of claims 1 to 7.

Images