In recent years, there has been remarkable progress in autonomous driving technology. To improve the safety of autonomous driving comprehensively, accurate predictions for all traffic agents are crucial. Typically, the graph neural network is widely employed for the trajectory prediction. To enhance the prediction accuracy rate, this paper utilizes a finetuned vision-to-language large model to extract driving intentions. With the well-designed prompt and the supervision of the specific dataset, the LLM (large language model) can analyze the current traffic condition and give the corresponding driving intention. This paper also combines the result of the LLM and the output of the traditional prediction model, and the future trajectory is modified with the driving intention, which can improve the final prediction accuracy. Finally, in the decision-making part, both the driving intention from the LLM and the trajectory from the traditional prediction model are considered in the boundary-based drivable area, and a safe planning path is then generated. According to the validation in the public motion forecasting dataset, this method has greatly improved the accuracy of the prediction and the safety of route planning.


    Access

    Download


    Export, share and cite



    Title :

    Advancing Autonomous Driving Safety Through LLM Enhanced Trajectory Prediction


    Additional title:

    Lect.Notes Mechanical Engineering


    Contributors:

    Conference:

    Advanced Vehicle Control Symposium ; 2024 ; Milan, Italy September 01, 2024 - September 05, 2024



    Publication date :

    2024-10-04


    Size :

    7 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English





    Autonomous driving map enhanced trajectory prediction method integrating semantic information

    ZHAO RUI / QIAO SENYAO / WU JIANGHANG et al. | European Patent Office | 2024

    Free access


    Intention-Driven Trajectory Prediction for Autonomous Driving

    Fan, Shiwei / Li, Xiangxu / Li, Fei | IEEE | 2021


    INTENTION-DRIVEN TRAJECTORY PREDICTION FOR AUTONOMOUS DRIVING

    Fan, Shiwei / Li, Xiangxu / Li, Fei | British Library Conference Proceedings | 2021