This work designs a scalable CoT curation pipeline that bootstraps rationales from a strong LLM through a two-round strategy and employs a VLM-based verifier to filter out incorrect cases, yielding a high-quality set of (CoT, answer) pairs.
Abstract
Understanding and complying with traffic regulations is a safety-critical requirement for autonomous driving, yet remains challenging due to the diversity and context dependence of traffic signage. Importantly, regulation understanding is not a simple recognition task, but a reasoning problem: whether a rule applies depends on interpreting the sign in relation to the spatial layout of lanes and scene context. To support such reasoning, MapDR provide fine-grained annotations that link each traffic sign's regulatory rules to the specific lanes they govern. Existing methods, however, largely treat this as direct sequence prediction, ignoring the underlying reasoning that connects sign semantics and map structure. To address this limitation, we explicitly incorporate reasoning into this task and propose a framework that equips vision-language models (VLMs) with chain-of-thought (CoT) capabilities. We first design a scalable CoT curation pipeline that bootstraps rationales from a strong LLM through a two-round strategy and employs a VLM-based verifier to filter out incorrect cases, yielding a high-quality set of (CoT, answer) pairs. Building on this foundation, we adopt a two-stage training scheme: supervised fine-tuning (SFT) to teach rationale-to-answer generation, followed by GRPO reinforcement learning with answer-grounded, fine-grained rewards to further improve final answer accuracy. Extensive experiments on MapDR show that our approach significantly improves both interpretability and accuracy, establishing the first reasoning-based framework for regulation-aware autonomous driving.
STGraphVQA, a framework that represents driving scenes as dynamic spatiotemporal graphs, where nodes denote traffic participants, edges encode spatial and semantic relationships, and the temporal dimension captures their evolution, provides a promising direction toward interpretable autonomous driving systems.
Dengquan Zhang, Jiayi Wu, Qingshan Wang et al.· International Conference on...· 0 citations
Driving topology reasoning is an important perception task in autonomous driving, which requires detecting road centerlines (lane) and traffic elements, further reasoning their topology relationship. However, deploying this in real-world scenarios faces two significant challenges: designing high-performance pipelines and cost-effectively annotating topological data. To overcome these obstacles, we first introduce a robust "first-detect-then-reason" framework, named TopoMLP++. The core of TopoMLP++ lies in its emphasis on designing a powerful 3D lane detector that leverages global attention modeling and geometry-aware enhancements. Additionally, it includes 2D traffic detectors augmented by YOLOv8 results. After detection, MLP-based heads are employed for lane topology prediction, where we extend traditional binary classification by integrating a geometry-aware strategy, ensuring that connected points are geometrically close. To further minimize annotation efforts and facilitate data scalability, we propose an agent-based data engine that utilizes the predictions from TopoMLP++. This framework incorporates a large language model (LLM) as an agent, which first employs TopoMLP++ to generate pseudo-labels. The agent then detects potential inconsistencies in the predictions and coordinates external tools to iteratively refine the predicted lane centerlines. This iterative process ultimately boosts the performance of TopoMLP++. Experiments on the OpenLane-V2 dataset demonstrate that TopoMLP++ achieves state-of-the-art results. Its initial version is the 1st solution for 1st OpenLane Topology in IEEE CVPR Autonomous Driving Challenge. Additionally, with just 50% labeled data, TopoMLP++ augmented by our agent-based data engine achieves 96% of the performance attained by full-data training.
Dongming Wu, Wencheng Han, Cheng-Zhong Xu et al.· IEEE Transactions on Pattern...· 0 citations
Spatial intelligence requires foundation models to maintain coherent spatial state across interactions with the physical world. However, existing data-centric approaches typically treat spatial reasoning as independent question-answer instances, enabling shortcut-based answering and providing limited supervision for persistent spatial understanding. To address this, we introduce ChainSpace, a chained-reasoning paradigm that structures spatial reasoning as a state-preserving multi-round process. In this paradigm, spatial questions are organized into logically constrained and jointly consistent chains, where later questions depend on spatial constraints established in earlier rounds. Following this principle, we instantiate ChainSpace-Bench, a manually annotated real-world multi-round benchmark with a Chain-Aware Metric, and ChainSpace-Pipeline, a simulator-based chain-structured supervision generation framework for spatial intelligence training. Experiments show that ChainSpace-Bench exposes chain-level failures that are not captured by isolated question accuracy. Additionally, with a relatively small amount of simulator-generated chained data, models trained by ChainSpace-Pipeline achieve the best performance among open-source models on ChainSpace-Bench and transfer competitively to multiple external spatial intelligence benchmarks. These results establish ChainSpace as an effective paradigm for more faithful evaluation and more data-efficient learning of spatial intelligence.
Xiaohan Zhang, Feng Gu, Xudong Rao et al.· 0 citations
WorkDrive is proposed, a framework that constructs perception-grounded causal reasoning for work zones and aligns it with trajectory prediction and achieves progressive improvement over the trajectory-only baseline.
Tianyi Jiang, Wen Zhang, Sihan Yang et al.· arXiv.org· 0 citations
Autonomous vehicles (AVs) operate in complex environments where failures are consequential. Sophisticated machine learning models for perception and planning are key to overcoming at least part of that complexity, but their black-box nature complicates validation and verification (V&V). The recent integration of Vision-Language-Action (VLA) models into AVs introduces a unique opportunity: besides generating trajectories, these models produce an explicit Chain-of-Thought (CoT) explaining their underlying rationale. This CoT provides a rich specification to cross-check model outputs and detect inconsistencies that may expose unsafe or unintended behavior. This paper assesses whether CoTs from a recent open driving VLA can support such monitoring. We curate DriveAlignBench, a specialized dataset from NVIDIA's Alpamayo 1.5 VLA for AVs containing 150 CoT-trajectory pairs, which we manually annotate for reliability, trajectory consistency, and safety. Our analysis reveals that 33.3% of CoTs are unreliable. Among reliable CoTs, the generated trajectory is consistent with the CoT in 74% of cases. Leveraging this potential, we propose integrating a CoT-trajectory consistency check into a runtime monitor. The check is nontrivial: CoTs express open-vocabulary, scene-relative driving commitments, while trajectories are low-level ego-motion sequences whose semantics depend on road geometry and motion context. To bridge this gap, we develop a family of automated consistency monitors. Our best monitor, lane-relative F-LLM with GPT-5.5, achieves F1 = 0.75, improving over the strongest raw-waypoint LLM baseline by +0.13 absolute F1 and over a rule-based monitor by +0.38. We release DriveAlignBench, the monitor implementations, and annotation tools at https://github.com/776styjsu/drive-the-thoughts.
Tianwen Yu, Lu Feng, Sebastian G. Elbaum· 0 citations
FlightLLM, a prior-guided semantic LLM-based approach for interpretable flight safety analysis that achieves competitive classification performance while generating direct and reasonable explanations for event causes is proposed.
Lu Xu, Xu Li, Linjiang Zheng et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
