By adding future visual state prediction and a dedicated inverse kinematics diffusion network that uses only visual boundary conditions, a 0.5B driving VLA recovers visual grounding and matches 7-8B models on NAVSIM-v2 and nuScenes.
hub Canonical reference
Vision-language-action models for autonomous driving: Past, present, and future
Canonical reference. 100% of citing Pith papers cite this work as background.
16
Pith papers citing it
Background
100% of classified citations
hub tools
citation-role summary
background 7
citation-polarity summary
years
2026 16roles
background 6polarities
background 6representative citing papers
A kinematic-to-visual lifting paradigm combined with hierarchically routed control generates action-conditioned surgical videos with better faithfulness, fidelity, and efficiency.
CAGE attack aligns perturbations with token compression to achieve lower robust accuracy on compressed LVLMs than baseline attacks across mechanisms and datasets.
DriveVer is a lightweight dual-head test-time verifier that predicts safety confidence scores and geometric refinement vectors for candidate trajectories, improving base planners on the NAVSIM benchmark.
nuReasoning is a new real-world dataset and benchmark extending nuScenes/nuPlan with 20k clips and multi-type reasoning annotations to evaluate and improve reasoning in long-tail autonomous driving.
CLAP reduces planning error on challenging driving scenarios by 24% on NAVSIM using contrastive latent-space prompt optimization on frozen VLA models with no regression on normal frames.
OneVL achieves superior accuracy to explicit chain-of-thought reasoning at answer-only latency by supervising latent tokens with a visual world model decoder that predicts future frames.
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.
LVDrive improves closed-loop driving on Bench2Drive by adding latent future scene prediction to VLA models via unified embedding space processing and two-stage trajectory decoding.
SafeAlign-VLA uses counterfactual safety pairing and anchor-based group relative policy optimization to incorporate negative data for safer VLA-based autonomous driving.
SpanVLA reduces action generation latency via flow-matching conditioned on history and improves robustness by training on negative-recovery samples with GRPO and a dedicated reasoning dataset.
DynFlowDrive models action-conditioned scene transitions via rectified flow in latent space and adds stability-aware trajectory selection, showing gains on nuScenes and NavSim without added inference cost.
A literature survey that unifies fragmented work on attacks, defenses, evaluations, and deployment challenges for Vision-Language-Action models in robotics.
L-LIO integrates audio with visual data to enhance driver safety assessment and intelligent vehicle decision-making via multimodal sensor fusion.
This review organizes literature on large multimodal models and object-centric vision into four themes—understanding, referring segmentation, editing, and generation—while summarizing paradigms, strategies, and challenges like instance permanence and consistent interaction.
citing papers explorer
-
On the Adversarial Robustness of Large Vision-Language Models under Visual Token Compression
CAGE attack aligns perturbations with token compression to achieve lower robust accuracy on compressed LVLMs than baseline attacks across mechanisms and datasets.