CN-121996208-A - System and method for constructing customized scene of vehicle based on SOA LLMs

Abstract

The invention relates to a system and a method for constructing a customized scene of a vehicle based on SOA LLMs. The system comprises a voice recognition device, an SOA atomic function library and a control device, wherein the voice recognition device is configured to acquire user voice information, convert the acquired user voice information into user demand information in a text form, classify the converted user demand information, determine whether the category of the user demand information is a customized scene, and generate the user demand information of the customized scene when the category of the user demand information is determined to be the customized scene, the SOA atomic function library is configured to provide state information of a sensor and an actuator of a vehicle, and the control device is configured to analyze the user demand information and generate a scheme of the customized scene suitable for the user demand based on LLMs by utilizing the user demand information of the customized scene and the SOA atomic function library.

Inventors

• NIU DONG
• YANG FENG
• HUANG RUZHANG

Assignees

• 现代自动车株式会社
• 起亚株式会社

Dates

Publication Date

20260508

Application Date

20241101

Claims (18)

1. 1. A system for constructing a customized scene of a vehicle based on a SOA LLMs, comprising: The voice recognition device is configured to acquire user voice information, convert the acquired user voice information into user demand information in a text form, classify the converted user demand information, determine whether the category of the user demand information is a customized scene, and generate user demand information of the customized scene when determining that the category of the user demand information is the customized scene; An SOA atomic function library configured to provide status information of sensors and actuators of a vehicle, and And a control device communicatively connected to the voice recognition device and the SOA atomic function library, the control device configured to analyze the user demand information based on LLMs and generate a solution for the customized scenario that is appropriate for the user demand using the user demand information for the customized scenario and the SOA atomic function library.
2. 2. The SOA-based LLMs system for constructing a custom scene for a vehicle according to claim 1, further comprising: And a display device communicatively connected with the control device and configured to display the generated scheme of the customized scene to the user.
3. 3. The SOA-based LLMs system for constructing a custom scene for a vehicle according to claim 2, further comprising: a face recognition device communicably connected with the control device and configured to recognize face information of the user, determine whether the user is a registered user, and send the determined user identity information to the control device, and A seat occupancy detection device communicatively connected to the control device and configured to determine an occupancy position of the user and to transmit the determined occupancy position information of the user to the control device; The control device is further configured to analyze the user demand information and generate a solution of the customized scene suitable for the user demand based on LLMs using the user demand information of the customized scene and the SOA atomic function library when it is determined that the user is a registered user or the user is seated in the driver seat.
4. 4. A system for constructing a customized scene of a vehicle based on SOA LLMs as claimed in claim 2 or 3, wherein the display means further comprises: A confirmation key for the user to indicate satisfaction of the generated scheme of the customized scene by selecting the confirmation key, and The reconstruction key is selected by the user to indicate that the scheme of the generated customized scene is unsatisfactory for the scheme of the generated customized scene such that the scheme of the customized scene is regenerated.
5. 5. The SOA-based LLMs build a custom scene for a vehicle according to claim 4, wherein the control device is further configured to: When the user is not satisfied with the generated scheme of the customized scene, regenerating the scheme of the customized scene according to the feedback information of the user until the user is satisfied; When the user is satisfied with the generated scheme of the customized scene, storing the generated scheme of the customized scene or storing the generated scheme of the customized scene and corresponding user identity information or riding position information.
6. 6. The SOA-based LLMs system for constructing a customised scene for a vehicle according to claim 5, wherein the control device is further configured to: when the condition of the stored scheme of the customized scene is satisfied, the stored scheme of the customized scene is executed.
7. 7. The SOA based LLMs system for constructing a custom scene for a vehicle according to claim 1, wherein, Training LLMs and prompting optimization processing are carried out by utilizing preset customized scene information; the LLMs is disposed in a control device or cloud in the vehicle.
8. 8. The SOA-based LLMs system for constructing a customised scene for a vehicle according to claim 1, wherein the speech recognition means comprises: A microphone configured to capture a voice signal of a user; an ASR module configured to convert a speech signal of a user into text information; an NLP module configured to process and parse text information to understand sentence structure and meaning, and And the NLU module is configured to understand the text information and make corresponding decisions according to the requirements.
9. 9. The SOA-based LLMs system for constructing a custom scene of a vehicle according to claim 3, wherein the face recognition means includes: A camera configured to acquire a face image of a user, and A processor configured to convert the acquired face image into face information, extract features of the face information, and compare the extracted features with the face features stored in the database to confirm whether the user is a registered user.
10. 10. A method of constructing a customized scene of a vehicle based on a SOA LLMs, comprising: By the speech recognition device: Acquiring voice information of a user; Converting the acquired user voice information into user demand information in a text form; Classifying the converted user demand information; Determining whether the category of the user demand information is a customized scene; When the category of the user demand information is determined to be a customized scene, generating the user demand information of the customized scene; providing status information of sensors and actuators of a vehicle from an SOA atomic function library, and The control device comprises: Analyzing the user demand information based on LLMs by using the user demand information of the customized scene and the SOA atomic function library, and A solution of a customized scene is generated that is tailored to the needs of the user.
11. 11. The method of constructing a customized scene of a vehicle based on SOA LLMs of claim 10, further comprising: The generated scheme of the customized scene is displayed to the user by the display device.
12. 12. The method of constructing a customized scene of a vehicle based on SOA LLMs of claim 11, further comprising: By the face recognition device: identifying facial information of a user; Determining whether the user is a registered user, and Transmitting the determined user identity information to a control device; seat occupancy detection device: Determining a seating position of a user, and Transmitting the determined user riding position information to a control device; the control device comprises: When the user is determined to be a registered user or the user sits in a driver seat, analyzing the user demand information based on LLMs by utilizing the user demand information of the customized scene and the SOA atomic function library; A solution of a customized scene is generated that is tailored to the needs of the user.
13. 13. The method of constructing a customized scene of a vehicle based on SOA LLMs according to claim 11 or 12, further comprising: Receiving, from a user, a representation generated by selection of a confirmation key on a display device that is satisfactory to a scenario of the customized scenario; representing dissatisfaction with the generated customized scene by the user selecting a reconstruction key on the display device; A scheme of customizing the scene is regenerated.
14. 14. The method of constructing a customized scene of a vehicle based on SOA LLMs of claim 13, further comprising: When the user is not satisfied with the generated scheme of the customized scene, the control device regenerates the scheme of the customized scene according to the feedback information of the user until the user is satisfied; When the user is satisfied with the generated scheme of the customized scene, storing the generated scheme of the customized scene or storing the generated scheme of the customized scene and corresponding user identity information or riding position information.
15. 15. The method of constructing a customized scene of a vehicle based on SOA LLMs as claimed in claim 14, wherein, When the condition of the stored scheme of the customized scene is satisfied, the stored scheme of the customized scene is executed.
16. 16. The method of constructing a customized scene of a vehicle based on SOA LLMs as claimed in claim 10, wherein, Training LLMs and prompting optimization processing are carried out by utilizing preset customized scene information; the LLMs is disposed in a control device or cloud in the vehicle.
17. 17. The method of constructing a customized scene of a vehicle based on SOA LLMs as claimed in claim 10, wherein, Capturing, by a microphone, a voice signal of a user; Converting, by the ASR module, the speech signal of the user into text information; processing and analyzing the text information by the NLP module to understand the structure and meaning of the sentence; The NLU module understands the text information and makes corresponding decisions according to the requirements.
18. 18. The method of constructing a customized scene of a vehicle based on SOA LLMs as claimed in claim 12, wherein, Acquiring a facial image of a user by a camera; The acquired face image is converted into face information by the processor, features of the face information are extracted, and the extracted features are compared with the face features stored in the database to confirm whether the user is a registered user.

Description

System and method for constructing customized scene of vehicle based on SOA LLMs Technical Field The present invention relates to the technical field of intelligent vehicles and software architecture, and more particularly, to a system and method for constructing a customized scene of a vehicle based on a large language model (Large Language Models, LLMs) of a Service-oriented architecture (Service-Oriented Architecture, SOA). Background Currently, vehicle developers are developing smart scenario applications that run under Service-oriented architecture (SOA), allowing users to customize various functions through the sensors and actuators of the vehicle. SOA is a design model in which applications consist of various independent services that can be updated and extended individually without affecting the overall system. In this case, various functions of the vehicle (e.g., driving assistance, entertainment system, navigation, etc.) are designed as independent "atomic" services, which can be combined and customized according to the needs of the user. However, this customization process can encounter challenges. Users may need to have some technical knowledge to understand and operate these complex systems, and manually editing and customizing settings may be difficult and time consuming. Accordingly, there is a need for systems and methods that can accurately understand the needs of a user and eliminate the difficulty of the user to set custom scenes themselves. The above description of the background is only for the purpose of facilitating a thorough understanding of the present invention's aspects (in terms of the means of technology used, the technical problems solved, and the technical effects produced, etc.) and should not be taken as an acknowledgement or any form of suggestion that this message constitutes prior art that is already known to a person skilled in the art. Disclosure of Invention The invention aims to understand the requirements of a user by utilizing a large-scale language model (Large Language Models, LLMs) and generate corresponding output according to the requirements of the user. The user can change the driving mode of the vehicle, adjust the temperature of the seat, change the music playlist, or the like by a simple language command. This may not only improve the user experience, but may also improve the safety and efficiency of the vehicle. According to an embodiment of the present invention, there is provided a system for constructing a customized scene of a vehicle based on an SOA LLMs, which includes a voice recognition device configured to acquire user voice information, convert the acquired user voice information into user demand information in a text form, classify the converted user demand information, determine whether a category of the user demand information is a customized scene, and generate user demand information of the customized scene when it is determined that the category of the user demand information is the customized scene, an SOA atomic function library configured to provide status information of sensors and actuators of the vehicle, and a control device communicably connected with the voice recognition device and the SOA atomic function library, the control device configured to analyze the user demand information based on LLMs and generate a scheme of the customized scene suitable for the user demand using the user demand information of the customized scene and the SOA atomic function library. Preferably, the system for constructing a customized scene of a vehicle based on the SOA LLMs further comprises a display device communicatively connected to the control device and configured to display the generated solution of the customized scene to a user. Preferably, the system for constructing a customized scene of a vehicle based on LLMs of the SOA further comprises a face recognition device communicatively connected to the control device and configured to recognize face information of the user, determine whether the user is a registered user, and send the determined user identity information to the control device, and a seat occupancy detection device communicatively connected to the control device and configured to determine the occupancy position of the user, and send the determined occupancy position information to the control device, the control device being further configured to analyze the user demand information and to generate a plan of the customized scene suitable for the user demand based on LLMs using the user demand information of the customized scene and the SOA atomic function library when it is determined that the user is a registered user or the user is sitting in the driver seat. Preferably, the display device further includes a confirm button, the user indicating satisfaction with the generated scheme of the customized scene by selecting the confirm button, and a reconstruct button, the user indicating dissatisfaction with the generated scheme