VERDI: VLM-Embedded Reasoning for Autonomous Driving

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Feng, Bowen, Mei, Zhiting, Ost, Julian, Ghilotti, Filippo, Li, Baiang, Girgis, Roger, Majumdar, Anirudha, Heide, Felix
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910101562982400
author Feng, Bowen
Mei, Zhiting
Ost, Julian
Ghilotti, Filippo
Li, Baiang
Girgis, Roger
Majumdar, Anirudha
Heide, Felix
author_facet Feng, Bowen
Mei, Zhiting
Ost, Julian
Ghilotti, Filippo
Li, Baiang
Girgis, Roger
Majumdar, Anirudha
Heide, Felix
contents While autonomous driving (AD) stacks struggle with decision making under partial observability and real-world complexity, human drivers are capable of applying commonsense reasoning to make near-optimal decisions with limited information. Recent work has attempted to leverage finetuned Vision-Language Models (VLMs) for trajectory planning at inference time to emulate human behavior. Despite their success in benchmark evaluations, these methods are often impractical to deploy (a 70B parameter VLM inference at merely 8 tokens per second requires more than 160G of memory), and their monolithic network structure prohibits safety decomposition. To bridge this gap, we propose VLM-Embedded Reasoning for autonomous DrIving (VERDI), a training-time framework that distills the reasoning process and commonsense knowledge of VLMs into the AD stack. VERDI augments modular differentiable end-to-end (e2e) AD models by aligning intermediate module outputs at the perception, prediction, and planning stages with text features explaining the driving reasoning process produced by VLMs. By encouraging alignment in latent space, VERDI enables the modular AD stack to internalize structured reasoning, without incurring the inference-time costs of large VLMs. We evaluate VERDI in both open-loop and closed-loop settings. Our method outperforms existing end-to-end approaches without embedded reasoning by up to 11% in $\ell_{2}$ distance, and achieves the best overall driving performance in the closed-loop HugSim simulator, including a 10% improvement in Non-Collision Rate, while maintaining fast inference speed.
format Preprint
id arxiv_https___arxiv_org_abs_2505_15925
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VERDI: VLM-Embedded Reasoning for Autonomous Driving
Feng, Bowen
Mei, Zhiting
Ost, Julian
Ghilotti, Filippo
Li, Baiang
Girgis, Roger
Majumdar, Anirudha
Heide, Felix
Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
While autonomous driving (AD) stacks struggle with decision making under partial observability and real-world complexity, human drivers are capable of applying commonsense reasoning to make near-optimal decisions with limited information. Recent work has attempted to leverage finetuned Vision-Language Models (VLMs) for trajectory planning at inference time to emulate human behavior. Despite their success in benchmark evaluations, these methods are often impractical to deploy (a 70B parameter VLM inference at merely 8 tokens per second requires more than 160G of memory), and their monolithic network structure prohibits safety decomposition. To bridge this gap, we propose VLM-Embedded Reasoning for autonomous DrIving (VERDI), a training-time framework that distills the reasoning process and commonsense knowledge of VLMs into the AD stack. VERDI augments modular differentiable end-to-end (e2e) AD models by aligning intermediate module outputs at the perception, prediction, and planning stages with text features explaining the driving reasoning process produced by VLMs. By encouraging alignment in latent space, VERDI enables the modular AD stack to internalize structured reasoning, without incurring the inference-time costs of large VLMs. We evaluate VERDI in both open-loop and closed-loop settings. Our method outperforms existing end-to-end approaches without embedded reasoning by up to 11% in $\ell_{2}$ distance, and achieves the best overall driving performance in the closed-loop HugSim simulator, including a 10% improvement in Non-Collision Rate, while maintaining fast inference speed.
title VERDI: VLM-Embedded Reasoning for Autonomous Driving
topic Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.15925