MDrive benchmark shows multi-agent cooperative driving systems generally outperform single-agent ones in closed-loop settings but perception sharing does not always improve planning and negotiation can harm performance in complex traffic.
SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
abstract
Vision-Language-Action (VLA) models offer a promising autonomous driving paradigm for leveraging world knowledge and reasoning capabilities, especially in long-tail scenarios. However, existing VLA models often struggle with the high latency in action generation using an autoregressive generation framework and exhibit limited robustness. In this paper, we propose SpanVLA, a novel end-to-end autonomous driving framework, integrating an autoregressive reasoning and a flow-matching action expert. First, SpanVLA introduces an efficient bridge to leverage the vision and reasoning guidance of VLM to efficiently plan future trajectories using a flow-matching policy conditioned on historical trajectory initialization, which significantly reduces inference time. Second, to further improve the performance and robustness of the SpanVLA model, we propose a GRPO-based post-training method to enable the VLA model not only to learn from positive driving samples but also to learn how to avoid the typical negative behaviors and learn recovery behaviors. We further introduce mReasoning, a new real-world driving reasoning dataset, focusing on complex, reasoning-demanding scenarios and negative-recovery samples. Extensive experiments on the NAVSIM (v1 and v2) demonstrate the competitive performance of the SpanVLA model. Additionally, the qualitative results across diverse scenarios highlight the planning performance and robustness of our model.
citation-role summary
background 1
baseline 1
citation-polarity summary
years
2026 7verdicts
UNVERDICTED 7representative citing papers
LWDrive uses future-frame supervision on VLMs to create world-model features that a multi-layer Foresight Cascade Planner refines into final trajectories, reporting 92.0 on NAVSIM and 89.6 on NAVSIM-v2.
TOAD applies test-time Cross-Entropy Method optimization to refine trajectories using the planner's scorer as a reward function, improving end-to-end autonomous driving performance without retraining.
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.
ChainFlow-VLA unifies autoregressive causal trajectory modes with VLM-conditioned diffusion refinement to reach 94.85 on NAVSIM v1, matching human performance.
DriveFuture achieves SOTA results on NAVSIM by conditioning latent world model states on future predictions to directly inform trajectory planning.
SafeAlign-VLA uses counterfactual safety pairing and anchor-based group relative policy optimization to incorporate negative data for safer VLA-based autonomous driving.
citing papers explorer
-
MDrive: Benchmarking Closed-Loop Cooperative Driving for End-to-End Multi-agent Systems
MDrive benchmark shows multi-agent cooperative driving systems generally outperform single-agent ones in closed-loop settings but perception sharing does not always improve planning and negotiation can harm performance in complex traffic.
-
LWDrive: Layer-Wise World-Model-Guided Vision-Language Model Planning for Autonomous Driving
LWDrive uses future-frame supervision on VLMs to create world-model features that a multi-layer Foresight Cascade Planner refines into final trajectories, reporting 92.0 on NAVSIM and 89.6 on NAVSIM-v2.
-
Test-Time Trajectory Optimization for Autonomous Driving
TOAD applies test-time Cross-Entropy Method optimization to refine trajectories using the planner's scorer as a reward function, improving end-to-end autonomous driving performance without retraining.
-
nuReasoning: A Reasoning-Centric Dataset and Benchmark for Long-Tail Autonomous Driving
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.
-
ChainFlow-VLA: Causal Flow Planning with Vision-Language Models
ChainFlow-VLA unifies autoregressive causal trajectory modes with VLM-conditioned diffusion refinement to reach 94.85 on NAVSIM v1, matching human performance.
-
DriveFuture: Future-Aware Latent World Models for Autonomous Driving
DriveFuture achieves SOTA results on NAVSIM by conditioning latent world model states on future predictions to directly inform trajectory planning.
-
SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving
SafeAlign-VLA uses counterfactual safety pairing and anchor-based group relative policy optimization to incorporate negative data for safer VLA-based autonomous driving.