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 Baseline reference
Hydra-MDP: End-to-end Multimodal Planning with Multi-target Hydra-Distillation
Baseline reference. 50% of citing Pith papers use this work as a benchmark or comparison.
63
Pith papers citing it
Baseline
50% of classified citations
abstract
We propose Hydra-MDP, a novel paradigm employing multiple teachers in a teacher-student model. This approach uses knowledge distillation from both human and rule-based teachers to train the student model, which features a multi-head decoder to learn diverse trajectory candidates tailored to various evaluation metrics. With the knowledge of rule-based teachers, Hydra-MDP learns how the environment influences the planning in an end-to-end manner instead of resorting to non-differentiable post-processing. This method achieves the $1^{st}$ place in the Navsim challenge, demonstrating significant improvements in generalization across diverse driving environments and conditions. More details by visiting \url{https://github.com/NVlabs/Hydra-MDP}.
hub tools
citation-role summary
background 6
baseline 6
citation-polarity summary
representative citing papers
Bench2Drive-Robust is a new closed-loop benchmark that evaluates end-to-end autonomous driving models under deployment perturbations from camera failures, ego-state errors, and compute delays, showing substantial performance degradation beyond image-level tests.
Evasive acceleration quantifies driving risk as the minimum 2D constant relative acceleration needed to avoid collision and outperforms time-to-collision on warning timing, discrimination, and information retention across crash datasets.
The SNG framework and SNG-VLA model enable end-to-end driving systems to better incorporate global navigation for state-of-the-art route following without auxiliary perception losses.
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.
WCog-VLA couples Game-CoT semantic reasoning with an aligned decoupled diffusion transformer to generate joint multi-agent trajectories and reaches 92.9 PDMS on NAVSIM.
DriveTeach-VLA adds Driving-aware Vision Distillation pretraining and 2D Trajectory-Guided Prompts to VLA models, then reports state-of-the-art results on NAVSIM and nuScenes.
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.
UniTeD unifies perception and planning in autonomous driving via shared temporal diffusion with TTM and ARS modules, reporting SOTA results on benchmarks.
FlowR2A learns reward-conditioned action distributions via flow-matching decoder to unify dense reward supervision with dynamic proposal generation for multimodal driving planning.
The paper fine-tunes Qwen3.5-4B as a driving VLA using serialized decision traces from rule-based planners, reporting reduced ADE and miss rate on a simulator benchmark with camera inputs.
Self-play DAgger training in a batched pixel renderer produces end-to-end driving policies that reach competitive performance on HUGSIM and NAVSIM-v2 after real-world adaptation and improve with more self-play compute.
DriveJudge combines VLM reasoning with rule functions on a new 33,577-sample human-annotated dataset, outperforming EPDMS by 21.23 AUC on quality classification and DriveCritic by 6.5% on trajectory preference.
Dash2Sim recovers metric geo-referenced 4D scenes from in-the-wild monocular dashcam videos to enable the ROADWork4D benchmark, revealing that current closed-loop planners fail on work zone lane changes.
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.
Action Diffusion Transformer generates multimodal driving actions via diffusion and nearest-neighbor selection, claiming SOTA on Bench2Drive with 10x lower latency.
IDOL uses inverse dynamics on adjacent predicted latent futures to extract planning-relevant motion deltas, then optimizes trajectories with a closed-loop refinement step, reporting SOTA results on NAVSIM v1 and v2.
DriveWAM converts video generative priors into a unified video-action policy for driving, reporting strong benchmark performance and positive scaling from 4k to 100k clips.
BeyondDrive augments imitation learning with synthesized safety-critical negative trajectories and a repulsive loss to improve safety in autonomous driving, reporting 89.7 PDMS on NAVSIMv1 and generalization to other models.
CLOVER is a closed-loop generator-scorer framework that expands proposal coverage with pseudo-expert trajectories and performs conservative self-distillation to achieve state-of-the-art planning scores on NAVSIM and nuScenes.
DAWN couples a world predictor with a world-conditioned action denoiser in latent space so that each refines the other recursively, yielding strong planning and safety results on autonomous driving benchmarks.
CoWorld-VLA extracts semantic, geometric, dynamic, and trajectory expert tokens from multi-source supervision and feeds them into a diffusion-based hierarchical planner, achieving competitive collision avoidance and trajectory accuracy on the NAVSIM v1 benchmark.
DriveFuture achieves SOTA results on NAVSIM by conditioning latent world model states on future predictions to directly inform trajectory planning.
VECTOR-DRIVE uses shared self-attention with semantic-aware expert routing of tokens to VL and trajectory experts plus flow-matching action decoding to reach 88.91 driving score on Bench2Drive.
citing papers explorer
-
Grounding Driving VLA via Inverse Kinematics
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.
-
Bench2Drive-Robust: Benchmarking Closed-Loop Autonomous Driving under Deployment Perturbations
Bench2Drive-Robust is a new closed-loop benchmark that evaluates end-to-end autonomous driving models under deployment perturbations from camera failures, ego-state errors, and compute delays, showing substantial performance degradation beyond image-level tests.
-
Driving risk emerges from the required two-dimensional joint evasive acceleration
Evasive acceleration quantifies driving risk as the minimum 2D constant relative acceleration needed to avoid collision and outperforms time-to-collision on warning timing, discrimination, and information retention across crash datasets.
-
Unveiling the Surprising Efficacy of Navigation Understanding in End-to-End Autonomous Driving
The SNG framework and SNG-VLA model enable end-to-end driving systems to better incorporate global navigation for state-of-the-art route following without auxiliary perception losses.
-
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.
-
WCog-VLA: A Dual-Level World-Cognitive Vision-Language-Action Model for End-to-End Autonomous Driving
WCog-VLA couples Game-CoT semantic reasoning with an aligned decoupled diffusion transformer to generate joint multi-agent trajectories and reaches 92.9 PDMS on NAVSIM.
-
Teaching Vision-Language-Action Models What to See and Where to Look
DriveTeach-VLA adds Driving-aware Vision Distillation pretraining and 2D Trajectory-Guided Prompts to VLA models, then reports state-of-the-art results on NAVSIM and nuScenes.
-
DriveVer: Lightweight Trajectory Evaluator as Test-Time Verifier for Autonomous Driving
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.
-
UniTeD: Unified Temporal Diffusion for Joint Perception and Planning in Autonomous Driving
UniTeD unifies perception and planning in autonomous driving via shared temporal diffusion with TTM and ARS modules, reporting SOTA results on benchmarks.
-
FlowR2A: Learning Reward-to-Action Distribution for Multimodal Driving Planning
FlowR2A learns reward-conditioned action distributions via flow-matching decoder to unify dense reward supervision with dynamic proposal generation for multimodal driving planning.
-
Neuro-Symbolic Drive: Rule-Grounded Faithful Reasoning for Driving VLAs
The paper fine-tunes Qwen3.5-4B as a driving VLA using serialized decision traces from rule-based planners, reporting reduced ADE and miss rate on a simulator benchmark with camera inputs.
-
Scaling Self-Play for End-to-End Driving
Self-play DAgger training in a batched pixel renderer produces end-to-end driving policies that reach competitive performance on HUGSIM and NAVSIM-v2 after real-world adaptation and improve with more self-play compute.
-
DriveJudge: Rethinking Autonomous Driving Evaluation with Vision-Language Models
DriveJudge combines VLM reasoning with rule functions on a new 33,577-sample human-annotated dataset, outperforming EPDMS by 21.23 AUC on quality classification and DriveCritic by 6.5% on trajectory preference.
-
Dash2Sim: Closed-Loop Driving Simulation from in-the-wild Dashcam Videos
Dash2Sim recovers metric geo-referenced 4D scenes from in-the-wild monocular dashcam videos to enable the ROADWork4D benchmark, revealing that current closed-loop planners fail on work zone lane changes.
-
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.
-
Multimodal Action Diffusion for Robust End-to-End Autonomous Driving
Action Diffusion Transformer generates multimodal driving actions via diffusion and nearest-neighbor selection, claiming SOTA on Bench2Drive with 10x lower latency.
-
IDOL: Inverse-Dynamics-Guided Future Prediction for End-to-End Autonomous Driving
IDOL uses inverse dynamics on adjacent predicted latent futures to extract planning-relevant motion deltas, then optimizes trajectories with a closed-loop refinement step, reporting SOTA results on NAVSIM v1 and v2.
-
DriveWAM: Video Generative Priors Enable Scalable World-Action Modeling for Autonomous Driving
DriveWAM converts video generative priors into a unified video-action policy for driving, reporting strong benchmark performance and positive scaling from 4k to 100k clips.
-
Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negatives
BeyondDrive augments imitation learning with synthesized safety-critical negative trajectories and a repulsive loss to improve safety in autonomous driving, reporting 89.7 PDMS on NAVSIMv1 and generalization to other models.
-
CLOVER: Closed-Loop Value Estimation and Ranking for End-to-End Autonomous Driving Planning
CLOVER is a closed-loop generator-scorer framework that expands proposal coverage with pseudo-expert trajectories and performs conservative self-distillation to achieve state-of-the-art planning scores on NAVSIM and nuScenes.
-
The DAWN of World-Action Interactive Models
DAWN couples a world predictor with a world-conditioned action denoiser in latent space so that each refines the other recursively, yielding strong planning and safety results on autonomous driving benchmarks.
-
CoWorld-VLA: Thinking in a Multi-Expert World Model for Autonomous Driving
CoWorld-VLA extracts semantic, geometric, dynamic, and trajectory expert tokens from multi-source supervision and feeds them into a diffusion-based hierarchical planner, achieving competitive collision avoidance and trajectory accuracy on the NAVSIM v1 benchmark.
-
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.
-
VECTOR-Drive: Tightly Coupled Vision-Language and Trajectory Expert Routing for End-to-End Autonomous Driving
VECTOR-DRIVE uses shared self-attention with semantic-aware expert routing of tokens to VL and trajectory experts plus flow-matching action decoding to reach 88.91 driving score on Bench2Drive.
-
ReflectDrive-2: Reinforcement-Learning-Aligned Self-Editing for Discrete Diffusion Driving
ReflectDrive-2 combines masked discrete diffusion with RL-aligned self-editing to generate and refine driving trajectories, reaching 91.0 PDMS on NAVSIM camera-only and 94.8 in best-of-6.
-
Unified Map Prior Encoder for Mapping and Planning
UMPE fuses any subset of HD/SD vector maps, raster SD maps, and satellite imagery into BEV features via alignment-aware vector and raster branches, raising mapping mAP by 5.3-5.9 points and cutting planning L2 error by 0.30 m on nuScenes.
-
ProDrive: Proactive Planning for Autonomous Driving via Ego-Environment Co-Evolution
ProDrive couples a query-centric planner with a BEV world model for end-to-end ego-environment co-evolution, enabling future-outcome assessment that improves safety and efficiency over reactive baselines on NAVSIM v1.
-
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.
-
Scaling-Aware Data Selection for End-to-End Autonomous Driving Systems
MOSAIC is a scaling-aware data selection framework that outperforms baselines in training end-to-end autonomous driving planners, achieving comparable or better EPDMS scores with up to 80% less data.
-
Orion-Lite: Distilling LLM Reasoning into Efficient Vision-Only Driving Models
Orion-Lite uses latent feature distillation and trajectory supervision to create a vision-only model that surpasses its LLM-based teacher on closed-loop Bench2Drive evaluation, achieving a new SOTA driving score of 80.6.
-
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.
-
Unleashing the Potential of Diffusion Models for End-to-End Autonomous Driving
The paper introduces Hyper Diffusion Planner (HDP), a diffusion-based E2E AD framework that identifies insights on loss space, trajectory representation and data scaling, adds RL post-training, and reports 10x performance gains over 200 km of real-world testing across 6 scenarios.
-
Pseudo-Expert Regularized Offline RL for End-to-End Autonomous Driving in Photorealistic Closed-Loop Environments
Pseudo-expert regularized offline RL reduces collisions and improves route completion for camera-based driving models trained on fixed simulator datasets from nuScenes.
-
SimScale: Learning to Drive via Real-World Simulation at Scale
SimScale synthesizes unseen driving states from real logs via neural rendering and reactive environments, generates pseudo-expert trajectories, and shows that co-training on real plus simulated data improves planning robustness and generalization on real benchmarks, with gains scaling by simulation
-
PRIX: Learning to Plan from Raw Pixels for End-to-End Autonomous Driving
PRIX presents an efficient camera-only planner with a novel CaRT module that matches larger multimodal models on NavSim and nuScenes while reducing model size and inference time.
-
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.
-
PriorEye: Geospatial Visual Priors for End-to-End Autonomous Driving
PriorEye augments end-to-end driving models with a dual-memory architecture that stores and gates geospatial visual priors to improve performance and robustness to sensor corruption on NAVSIM-v2.
-
Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving
R²LPL converts recoverable policy mistakes identified in closed-loop rollouts into corrective supervised targets for lifelong policy improvement, reaching SOTA on nuPlan benchmarks with few cycles.
-
ASSCG: Just-Right Gating over Chattering for Fast-Slow LLM Planning in Autonomous Driving
ASSCG is an RWKV-based adaptive gate trained with SFT and GRPO-style RL that makes Query/Cache/Drop decisions for slow LLM guidance in fast-slow autonomous driving planners, improving scores and cutting latency on nuPlan and NAVSIM.
-
Diffusion Forcing Planner: History-Annealed Planning with Time-Dependent Guidance for Autonomous Driving
Diffusion Forcing Planner applies heterogeneous joint diffusion with time-dependent noise and classifier-free guidance on history segments to generate stable, controllable motion plans for autonomous driving on nuPlan.
-
DriveReward: A Comprehensive Dataset and Generative Vision-Language Reward Model for Autonomous Driving
Creates DriveReward dataset with counterfactual annotations and a 1B VLM reward model that outperforms larger VLMs on driving tasks and matches rule-based rewards in RL and trajectory scoring.
-
PLAN-S: Bridging Planning with Latent Style Dynamics for Autonomous Driving World Models
PLAN-S decodes a style-conditioned four-channel semantic cost map from latent representations to bridge world models and planners in autonomous driving, reporting 0.55 m average L2 and 42% collision reduction on nuScenes plus PDMS gains on NAVSIM.
-
Discrete-WAM: Unified Discrete Vision-Action Token Editing for World-Policy Learning
Discrete-WAM unifies world modeling and policy learning for autonomous driving by representing observations, states, decisions, and actions as tokens in one space and using hierarchical token editing for planning.
-
Decision-Making with Lightweight Confidence-Aware Language Model for Autonomous Driving
Lightweight confidence-aware LM distilled from multi-agent CoT demonstrations achieves SOTA success rates on nuPlan benchmark for AD decision-making with low inference latency.
-
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.
-
Towards Physically Consistent 4D Scene Reconstruction for Closed-loop Autonomous Driving Simulation
Introduces Orthogonal Projected Gradient (OPG) and a smoothness-based temporal regularization to restore spatial identifiability and ensure physically consistent 4D scene reconstruction for closed-loop autonomous driving simulation.
-
HEAT: Heterogeneous End-to-End Autonomous Driving via Trajectory-Guided World Models
HEAT uses a trajectory-driven learning paradigm and a world model predicting future latent features from ego actions to enable a single unified end-to-end autonomous driving model to perform well across heterogeneous domains on nuScenes, NAVSIM, and Waymo benchmarks.
-
DriveSafer: End-to-End Autonomous Driving with Safety Guidance
DriveSafer reduces catastrophic failures (PDMS=0) by 48% and drivable-area compliance failures by over 65% versus DiffusionDrive on the NAVSIM benchmark by combining training-time safety constraints with inference-time guidance.
-
REAP: Reinforcement-Learning End-to-End Autonomous Parking with Gaussian Splatting Simulator for Real2Sim2Real Transfer
REAP trains an end-to-end SAC policy with behavior cloning and collision penalties inside a 3DGS Real2Sim simulator and transfers it to physical vehicles, succeeding in narrow mechanical parking slots.
-
CRAFT: Counterfactual-to-Interactive Reinforcement Fine-Tuning for Driving Policies
CRAFT is an on-policy RL fine-tuning framework that decomposes closed-loop policy gradients into a group-normalized counterfactual proxy plus residual correction from interaction events, achieving top closed-loop performance on Bench2Drive across multiple driving architectures.