SCORP delivers 10-28% gains in safety and 2-7% in efficiency metrics on WOMD by using dual-path scene conditioning in diffusion planning plus variance-gated group-relative policy optimization for closed-loop stability.
hub Canonical reference
AlphaDrive: Unleashing the Power of VLMs in Autonomous Driving via Reinforcement Learning and Reasoning
Canonical reference. 100% of citing Pith papers cite this work as background.
21
Pith papers citing it
Background
100% of classified citations
abstract
OpenAI o1 and DeepSeek R1 achieve or even surpass human expert-level performance in complex domains like mathematics and science, with reinforcement learning (RL) and reasoning playing a crucial role. In autonomous driving, recent end-to-end models have greatly improved planning performance but still struggle with long-tailed problems due to limited common sense and reasoning abilities. Some studies integrate vision-language models (VLMs) into autonomous driving, but they typically rely on pre-trained models with simple supervised fine-tuning (SFT) on driving data, without further exploration of training strategies or optimizations specifically tailored for planning. In this paper, we propose AlphaDrive, a RL and reasoning framework for VLMs in autonomous driving. AlphaDrive introduces four GRPO-based RL rewards tailored for planning and employs a two-stage planning reasoning training strategy that combines SFT with RL. As a result, AlphaDrive significantly improves both planning performance and training efficiency compared to using only SFT or without reasoning. Moreover, we are also excited to discover that, following RL training, AlphaDrive exhibits some emergent multimodal planning capabilities, which is critical for improving driving safety and efficiency. To the best of our knowledge, AlphaDrive is the first to integrate GRPO-based RL with planning reasoning into autonomous driving. Code will be released to facilitate future research.
hub tools
citation-role summary
background 11
citation-polarity summary
roles
background 11polarities
background 11representative citing papers
VLA-World improves autonomous driving by using action-guided future image generation followed by reflective reasoning over the imagined scene to refine trajectories.
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.
PaIR-Drive runs IL and RL in parallel branches with a tree-structured sampler to reach 91.2 PDMS and 87.9 EPDMS on NAVSIM benchmarks while outperforming sequential RL fine-tuning and correcting some human errors.
MAPLE proposes latent multi-agent rollouts with supervised fine-tuning followed by reinforcement learning using safety, progress, interaction, and diversity rewards to enable scalable closed-loop training for end-to-end autonomous driving.
FeaXDrive improves end-to-end autonomous driving by shifting diffusion planning to a trajectory-centric formulation with curvature-constrained training, drivable-area guidance, and GRPO post-training, yielding stronger closed-loop performance and feasibility on NAVSIM.
LLM-driven multi-planner scheduling framework turns open-ended passenger instructions into safe, traceable control signals for autonomous vehicles while cutting query costs and matching specialized safety levels.
VENUSS evaluates 25+ VLMs across 2600+ sequential driving scenarios and finds top models reach only 57% accuracy versus 65% for humans, with good static detection but poor performance on vehicle dynamics and temporal relations.
Sim2Real-AD enables zero-shot transfer of CARLA-trained VLM-guided RL policies to full-scale vehicles, reporting 75-90% success rates in car-following, obstacle avoidance, and stop-sign scenarios without real-world RL training data.
DVGT-2 is a streaming vision-geometry-action model that jointly reconstructs dense 3D geometry and plans trajectories online, achieving better reconstruction than prior batch methods while transferring directly to planning benchmarks without fine-tuning.
EcoGym is a new open benchmark with three economic environments that reveals no leading LLM dominates at sustained plan-and-execute decision making across scenarios.
AutoVLA unifies semantic reasoning and trajectory planning in one autoregressive VLA model for end-to-end autonomous driving by tokenizing trajectories into discrete actions and using GRPO reinforcement fine-tuning to adaptively reduce unnecessary reasoning.
CoPhy is a new RL framework that distills VLM cognition into BEV encoders, adds an auto-regressive BEV world model for action-conditioned future prediction, and optimizes policies via GRPO with dual physical-cognitive rewards, claiming SOTA on NAVSIM v1/v2.
SafeAlign-VLA uses counterfactual safety pairing and anchor-based group relative policy optimization to incorporate negative data for safer VLA-based autonomous driving.
CaAD adds ego-centric joint-causal modeling and causality-aware policy alignment to end-to-end driving, reporting Driving Score 87.53 and PDMS 91.1 on Bench2Drive and NAVSIM.
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.
RAD-2 uses a diffusion generator and RL discriminator to cut collision rates by 56% in closed-loop autonomous driving planning.
DIVER uses RL-guided diffusion to produce diverse feasible trajectories from one ground-truth path, addressing mode collapse in imitation learning for autonomous driving.
The survey frames VLA models as pipelines that generate progressively grounded action tokens and classifies those tokens into eight types to guide future development.
DeepSight uses parallel latent feature prediction in BEV for long-horizon world modeling and adaptive text reasoning to reach state-of-the-art closed-loop performance on the Bench2drive benchmark.
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 reasoning and embodied performance on 18 benchmarks.
citing papers explorer
-
SCORP: Scene-Consistent Multi-agent Diffusion Planning with Stable Online Reinforcement Post-Training for Cooperative Driving
SCORP delivers 10-28% gains in safety and 2-7% in efficiency metrics on WOMD by using dual-path scene conditioning in diffusion planning plus variance-gated group-relative policy optimization for closed-loop stability.
-
Learning Vision-Language-Action World Models for Autonomous Driving
VLA-World improves autonomous driving by using action-guided future image generation followed by reflective reasoning over the imagined scene to refine trajectories.
-
The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models
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.
-
Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving
PaIR-Drive runs IL and RL in parallel branches with a tree-structured sampler to reach 91.2 PDMS and 87.9 EPDMS on NAVSIM benchmarks while outperforming sequential RL fine-tuning and correcting some human errors.
-
MAPLE: Latent Multi-Agent Play for End-to-End Autonomous Driving
MAPLE proposes latent multi-agent rollouts with supervised fine-tuning followed by reinforcement learning using safety, progress, interaction, and diversity rewards to enable scalable closed-loop training for end-to-end autonomous driving.
-
FeaXDrive: Feasibility-aware Trajectory-Centric Diffusion Planning for End-to-End Autonomous Driving
FeaXDrive improves end-to-end autonomous driving by shifting diffusion planning to a trajectory-centric formulation with curvature-constrained training, drivable-area guidance, and GRPO post-training, yielding stronger closed-loop performance and feasibility on NAVSIM.
-
Open-Ended Instruction Realization with LLM-Enabled Multi-Planner Scheduling in Autonomous Vehicles
LLM-driven multi-planner scheduling framework turns open-ended passenger instructions into safe, traceable control signals for autonomous vehicles while cutting query costs and matching specialized safety levels.
-
How Well Do Vision-Language Models Understand Sequential Driving Scenes? A Sensitivity Study
VENUSS evaluates 25+ VLMs across 2600+ sequential driving scenarios and finds top models reach only 57% accuracy versus 65% for humans, with good static detection but poor performance on vehicle dynamics and temporal relations.
-
Sim2Real-AD: A Modular Sim-to-Real Framework for Deploying VLM-Guided Reinforcement Learning in Real-World Autonomous Driving
Sim2Real-AD enables zero-shot transfer of CARLA-trained VLM-guided RL policies to full-scale vehicles, reporting 75-90% success rates in car-following, obstacle avoidance, and stop-sign scenarios without real-world RL training data.
-
DVGT-2: Vision-Geometry-Action Model for Autonomous Driving at Scale
DVGT-2 is a streaming vision-geometry-action model that jointly reconstructs dense 3D geometry and plans trajectories online, achieving better reconstruction than prior batch methods while transferring directly to planning benchmarks without fine-tuning.
-
EcoGym: Evaluating LLMs for Long-Horizon Plan-and-Execute in Interactive Economies
EcoGym is a new open benchmark with three economic environments that reveals no leading LLM dominates at sustained plan-and-execute decision making across scenarios.
-
AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning
AutoVLA unifies semantic reasoning and trajectory planning in one autoregressive VLA model for end-to-end autonomous driving by tokenizing trajectories into discrete actions and using GRPO reinforcement fine-tuning to adaptively reduce unnecessary reasoning.
-
Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving
CoPhy is a new RL framework that distills VLM cognition into BEV encoders, adds an auto-regressive BEV world model for action-conditioned future prediction, and optimizes policies via GRPO with dual physical-cognitive rewards, claiming SOTA on NAVSIM v1/v2.
-
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.
-
Causality-Aware End-to-End Autonomous Driving via Ego-Centric Joint Scene Modeling
CaAD adds ego-centric joint-causal modeling and causality-aware policy alignment to end-to-end driving, reporting Driving Score 87.53 and PDMS 91.1 on Bench2Drive and NAVSIM.
-
SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model
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.
-
RAD-2: Scaling Reinforcement Learning in a Generator-Discriminator Framework
RAD-2 uses a diffusion generator and RL discriminator to cut collision rates by 56% in closed-loop autonomous driving planning.
-
DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving
DIVER uses RL-guided diffusion to produce diverse feasible trajectories from one ground-truth path, addressing mode collapse in imitation learning for autonomous driving.
-
A Survey on Vision-Language-Action Models: An Action Tokenization Perspective
The survey frames VLA models as pipelines that generate progressively grounded action tokens and classifies those tokens into eight types to guide future development.
-
DeepSight: Long-Horizon World Modeling via Latent States Prediction for End-to-End Autonomous Driving
DeepSight uses parallel latent feature prediction in BEV for long-horizon world modeling and adaptive text reasoning to reach state-of-the-art closed-loop performance on the Bench2drive benchmark.
-
XEmbodied: A Foundation Model with Enhanced Geometric and Physical Cues for Large-Scale Embodied Environments
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 reasoning and embodied performance on 18 benchmarks.