Pith. sign in

REVIEW 14 cited by

Are VLMs Ready for Autonomous Driving? An Empirical Study from the Reliability, Data, and Metric Perspectives

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2501.04003 v1 pith:7HFMFSIR submitted 2025-01-07 cs.CV cs.RO

Are VLMs Ready for Autonomous Driving? An Empirical Study from the Reliability, Data, and Metric Perspectives

classification cs.CV cs.RO
keywords drivingvlmsautonomousinterpretablereliabilityvisualaddressbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Recent advancements in Vision-Language Models (VLMs) have sparked interest in their use for autonomous driving, particularly in generating interpretable driving decisions through natural language. However, the assumption that VLMs inherently provide visually grounded, reliable, and interpretable explanations for driving remains largely unexamined. To address this gap, we introduce DriveBench, a benchmark dataset designed to evaluate VLM reliability across 17 settings (clean, corrupted, and text-only inputs), encompassing 19,200 frames, 20,498 question-answer pairs, three question types, four mainstream driving tasks, and a total of 12 popular VLMs. Our findings reveal that VLMs often generate plausible responses derived from general knowledge or textual cues rather than true visual grounding, especially under degraded or missing visual inputs. This behavior, concealed by dataset imbalances and insufficient evaluation metrics, poses significant risks in safety-critical scenarios like autonomous driving. We further observe that VLMs struggle with multi-modal reasoning and display heightened sensitivity to input corruptions, leading to inconsistencies in performance. To address these challenges, we propose refined evaluation metrics that prioritize robust visual grounding and multi-modal understanding. Additionally, we highlight the potential of leveraging VLMs' awareness of corruptions to enhance their reliability, offering a roadmap for developing more trustworthy and interpretable decision-making systems in real-world autonomous driving contexts. The benchmark toolkit is publicly accessible.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 14 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. TRIP-Evaluate: An Open Multimodal Benchmark for Evaluating Large Models in Transportation

    cs.CV 2026-04 accept novelty 7.0

    TRIP-Evaluate is a new open multimodal benchmark with 837 text, image, and point-cloud items organized by a role-task-knowledge taxonomy to evaluate large models on transportation workflows.

  2. Steadily moving semi-infinite fracture in plane poroelasticity

    physics.geo-ph 2026-04 unverdicted novelty 7.0

    A new coupled boundary integral method models steadily moving semi-infinite fractures in plane poroelasticity, solving for mechanical deformation and fluid exchange with verification on analytical test cases.

  3. The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models

    cs.CV 2026-04 unverdicted novelty 7.0

    Fine-tuning VLMs for driving erodes pre-trained world knowledge, but shifting adaptation to prompt space via the Drive Expert Adapter preserves generalization while improving task performance.

  4. From Accuracy to Visual Dependence: Auditing and Filtering Modality Collapse in Traffic VideoQA

    cs.CV 2026-06 unverdicted novelty 6.0

    Audit of four VideoQA benchmarks reveals text-only shortcuts in VLMs; new diagnostics Blind Gap, Visual Gain, and Shortcut Score quantify and filter visual dependence.

  5. MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs

    cs.LG 2026-02 conditional novelty 6.0

    MapTab introduces a 328-map, 196,800-query benchmark showing that current multimodal LLMs fall far short on multi-criteria route planning from maps-plus-tables.

  6. MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs

    cs.LG 2026-02 conditional novelty 6.0

    MapTab is a new multimodal benchmark with 328 images and nearly 200k queries that shows current MLLMs have substantial difficulty with multi-criteria route planning when visual and tabular information must be combined.

  7. MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs

    cs.LG 2026-02 unverdicted novelty 6.0

    MapTab benchmark shows current MLLMs struggle with multi-criteria multimodal route planning and that combining vision and language frequently underperforms single-modality approaches.

  8. Descriptor: Distance-Annotated Traffic Perception Question Answering (DTPQA)

    cs.CV 2025-11 unverdicted novelty 6.0

    DTPQA is a new VQA benchmark consisting of synthetic and real-world traffic images with distance annotations to isolate and measure VLM perception capabilities for driving decisions.

  9. Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail

    cs.RO 2025-10 conditional novelty 6.0

    Alpamayo-R1 introduces a VLA model with a Chain of Causation dataset and multi-stage SFT-plus-RL training that reports 12% better planning accuracy and 35% fewer close encounters versus trajectory-only baselines in dr...

  10. Benchmarking the Robustness of Autonomous Driving to Environmental Illusions: A Lane Perception Perspective

    cs.CV 2026-07 conditional novelty 5.0

    Environmental illusions cause 5-7% accuracy drops in lane detection models and can trigger collisions in closed-loop simulation, with a proposed defense (MIDA) recovering ~4% robustness.

  11. Steadily moving semi-infinite fracture in plane poroelasticity

    physics.geo-ph 2026-04 conditional novelty 5.0

    XEmbodied achieves SOTA on 18 embodied VQA benchmarks by fusing 3D geometric tokens and distilled physical cues into a 30B VLM with progressive curriculum training.

  12. RoadBench: Benchmarking MLLMs on Fine-Grained Spatial Understanding and Reasoning under Urban Road Scenarios

    cs.CV 2025-11 conditional novelty 5.0

    A new 9,121-case benchmark of road-marking tasks shows most multimodal LLMs perform near or below simple rule-based baselines in fine-grained urban spatial reasoning.

  13. XEmbodied: A Foundation Model with Enhanced Geometric and Physical Cues for Large-Scale Embodied Environments

    cs.CV 2026-04 unverdicted novelty 4.0

    XEmbodied is a foundation model that integrates 3D geometric and physical signals into VLMs using a 3D Adapter and Efficient Image-Embodied Adapter, plus progressive curriculum and RL post-training, to improve spatial...

  14. Structured Labeling Enables Faster Vision-Language Models for End-to-End Autonomous Driving

    cs.CV 2025-06 unverdicted novelty 4.0

    Introduces structured NuScenes-S dataset and 0.9B FastDrive VLM claiming 20% higher decision accuracy and over 10x inference speedup versus larger unstructured VLMs.