OmniDrive-R1: Reinforcement-driven Interleaved Multi-modal Chain-of-Thought for Trustworthy Vision-Language Autonomous Driving

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Zhenguo, Zheng, Haohan, Wang, Yishen, Xu, Le, Deng, Tianchen, Chen, Xuefeng, Chen, Qu, Zhang, Bo, Huang, Wuxiong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915969497038848
author Zhang, Zhenguo
Zheng, Haohan
Wang, Yishen
Xu, Le
Deng, Tianchen
Chen, Xuefeng
Chen, Qu
Zhang, Bo
Huang, Wuxiong
author_facet Zhang, Zhenguo
Zheng, Haohan
Wang, Yishen
Xu, Le
Deng, Tianchen
Chen, Xuefeng
Chen, Qu
Zhang, Bo
Huang, Wuxiong
contents The deployment of Vision-Language Models (VLMs) in safety-critical domains like autonomous driving (AD) is critically hindered by reliability failures, most notably object hallucination. This failure stems from their reliance on ungrounded, text-based Chain-of-Thought (CoT) reasoning. While existing multi-modal CoT approaches attempt mitigation, they suffer from two fundamental flaws: (1) decoupled perception and reasoning stages that prevent end-to-end joint optimization, and (2) reliance on expensive, dense localization labels. Thus we introduce OmniDrive-R1, an end-to-end VLM framework designed for autonomous driving, which unifies perception and reasoning through an interleaved Multi-modal Chain-of-Thought (iMCoT) mechanism. Our core innovation is an Reinforcement-driven visual grounding capability, enabling the model to autonomously direct its attention and "zoom in" on critical regions for fine-grained analysis. This capability is enabled by our pure two-stage reinforcement learning training pipeline and Clip-GRPO algorithm. Crucially, Clip-GRPO introduces an annotation-free, process-based grounding reward. This reward not only eliminates the need for dense labels but also circumvents the instability of external tool calls by enforcing real-time cross-modal consistency between the visual focus and the textual reasoning. Extensive experiments on DriveLMM-o1 demonstrate our model's significant improvements. Compared to the baseline Qwen2.5VL-7B, OmniDrive-R1 improves the overall reasoning score from 51.77% to 80.35%, and the final answer accuracy from 37.81% to 73.62%.
format Preprint
id arxiv_https___arxiv_org_abs_2512_14044
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OmniDrive-R1: Reinforcement-driven Interleaved Multi-modal Chain-of-Thought for Trustworthy Vision-Language Autonomous Driving
Zhang, Zhenguo
Zheng, Haohan
Wang, Yishen
Xu, Le
Deng, Tianchen
Chen, Xuefeng
Chen, Qu
Zhang, Bo
Huang, Wuxiong
Computer Vision and Pattern Recognition
Artificial Intelligence
The deployment of Vision-Language Models (VLMs) in safety-critical domains like autonomous driving (AD) is critically hindered by reliability failures, most notably object hallucination. This failure stems from their reliance on ungrounded, text-based Chain-of-Thought (CoT) reasoning. While existing multi-modal CoT approaches attempt mitigation, they suffer from two fundamental flaws: (1) decoupled perception and reasoning stages that prevent end-to-end joint optimization, and (2) reliance on expensive, dense localization labels. Thus we introduce OmniDrive-R1, an end-to-end VLM framework designed for autonomous driving, which unifies perception and reasoning through an interleaved Multi-modal Chain-of-Thought (iMCoT) mechanism. Our core innovation is an Reinforcement-driven visual grounding capability, enabling the model to autonomously direct its attention and "zoom in" on critical regions for fine-grained analysis. This capability is enabled by our pure two-stage reinforcement learning training pipeline and Clip-GRPO algorithm. Crucially, Clip-GRPO introduces an annotation-free, process-based grounding reward. This reward not only eliminates the need for dense labels but also circumvents the instability of external tool calls by enforcing real-time cross-modal consistency between the visual focus and the textual reasoning. Extensive experiments on DriveLMM-o1 demonstrate our model's significant improvements. Compared to the baseline Qwen2.5VL-7B, OmniDrive-R1 improves the overall reasoning score from 51.77% to 80.35%, and the final answer accuracy from 37.81% to 73.62%.
title OmniDrive-R1: Reinforcement-driven Interleaved Multi-modal Chain-of-Thought for Trustworthy Vision-Language Autonomous Driving
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2512.14044