A scene-driven simulation training method for an intelligent cockpit agent

By using multi-source sensor data processing and scene-driven simulation training methods, the problem of adaptive interaction in intelligent cockpits under complex scenarios was solved, achieving high-precision dynamic scene mapping and adaptive interaction strategies, thereby improving the accuracy and safety of drivers' operations in emergency situations.

CN122392375APending Publication Date: 2026-07-14SHENYANG QIYUAN TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHENYANG QIYUAN TECHNOLOGY CO LTD
Filing Date
2026-05-06
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing smart cockpits struggle to achieve accurate adaptive interaction and real-time response when faced with highly dynamic and unpredictable complex scenarios, resulting in poor interactive experience or safety hazards.

Method used

By acquiring and preprocessing multi-source sensor data, clustering to construct diverse scene libraries, combining temporal convolutional networks to extract environmental adaptation features, conducting risk assessment and interaction configuration, generating an audiovisual deviation correction model, and realizing dynamic scene mapping and adaptive interaction strategies.

Benefits of technology

It improves the intelligent cockpit's sensitivity to highly dynamic driving scenarios and the rationality of resource allocation, ensuring high accuracy and safety of interactive feedback under extreme and sudden conditions, and enhancing the wake-up rate of emergency alarms and the physical fidelity of immersive simulation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122392375A_ABST
    Figure CN122392375A_ABST
Patent Text Reader

Abstract

The application relates to the field of intelligent transportation, and discloses a scene-driven simulation training method for an intelligent cabin intelligent agent, which comprises the following steps: acquiring multi-source sensing data, preprocessing the multi-source sensing data to obtain a fusion data set, and performing typical driving scene clustering to obtain a diversified scene library; performing environment feature self-adaptive compensation according to light intensity mutation values and driving action time sequence coordinates to obtain environment adaptive features, and performing response priority evaluation to obtain a priority response sequence; performing scene risk mapping evaluation according to environment emergency condition data and a road surface friction coefficient to obtain a risk evaluation value, and generating a prompt sound effect scheme; performing audiovisual synchronous reconstruction verification according to the prompt sound effect scheme and picture rendering instructions to obtain a target simulation model; and performing interactive self-adaptive configuration generation according to scene diversity coding, a target interaction model and historical operation data to obtain a final interactive configuration. The application realizes precise adaptive interaction under high dynamic scenes.
Need to check novelty before this filing date? Find Prior Art