LLaViDA: A Large Language Vision Driving Assistant for Explicit Reasoning and Enhanced Trajectory Planning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Liu, Yudong, Hallyburton, Spencer, Kim, Jiwoo, Lin, Yueqian, Li, Yiming, Wang, Qinsi, Ye, Hui, Sun, Jingwei, Pajic, Miroslav, Chen, Yiran, Li, Hai
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914212193763328
author Liu, Yudong
Hallyburton, Spencer
Kim, Jiwoo
Lin, Yueqian
Li, Yiming
Wang, Qinsi
Ye, Hui
Sun, Jingwei
Pajic, Miroslav
Chen, Yiran
Li, Hai
author_facet Liu, Yudong
Hallyburton, Spencer
Kim, Jiwoo
Lin, Yueqian
Li, Yiming
Wang, Qinsi
Ye, Hui
Sun, Jingwei
Pajic, Miroslav
Chen, Yiran
Li, Hai
contents Trajectory planning is a fundamental yet challenging component of autonomous driving. End-to-end planners frequently falter under adverse weather, unpredictable human behavior, or complex road layouts, primarily because they lack strong generalization or few-shot capabilities beyond their training data. We propose LLaViDA, a Large Language Vision Driving Assistant that leverages a Vision-Language Model (VLM) for object motion prediction, semantic grounding, and chain-of-thought reasoning for trajectory planning in autonomous driving. A two-stage training pipeline--supervised fine-tuning followed by Trajectory Preference Optimization (TPO)--enhances scene understanding and trajectory planning by injecting regression-based supervision, produces a powerful "VLM Trajectory Planner for Autonomous Driving." On the NuScenes benchmark, LLaViDA surpasses state-of-the-art end-to-end and other recent VLM/LLM-based baselines in open-loop trajectory planning task, achieving an average L2 trajectory error of 0.31 m and a collision rate of 0.10% on the NuScenes test set. The code for this paper is available at GitHub.
format Preprint
id arxiv_https___arxiv_org_abs_2512_18211
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LLaViDA: A Large Language Vision Driving Assistant for Explicit Reasoning and Enhanced Trajectory Planning
Liu, Yudong
Hallyburton, Spencer
Kim, Jiwoo
Lin, Yueqian
Li, Yiming
Wang, Qinsi
Ye, Hui
Sun, Jingwei
Pajic, Miroslav
Chen, Yiran
Li, Hai
Robotics
Artificial Intelligence
Trajectory planning is a fundamental yet challenging component of autonomous driving. End-to-end planners frequently falter under adverse weather, unpredictable human behavior, or complex road layouts, primarily because they lack strong generalization or few-shot capabilities beyond their training data. We propose LLaViDA, a Large Language Vision Driving Assistant that leverages a Vision-Language Model (VLM) for object motion prediction, semantic grounding, and chain-of-thought reasoning for trajectory planning in autonomous driving. A two-stage training pipeline--supervised fine-tuning followed by Trajectory Preference Optimization (TPO)--enhances scene understanding and trajectory planning by injecting regression-based supervision, produces a powerful "VLM Trajectory Planner for Autonomous Driving." On the NuScenes benchmark, LLaViDA surpasses state-of-the-art end-to-end and other recent VLM/LLM-based baselines in open-loop trajectory planning task, achieving an average L2 trajectory error of 0.31 m and a collision rate of 0.10% on the NuScenes test set. The code for this paper is available at GitHub.
title LLaViDA: A Large Language Vision Driving Assistant for Explicit Reasoning and Enhanced Trajectory Planning
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2512.18211