SOTAVerified

NavSim

Data-Driven Non-Reactive Autonomous Vehicle Benchmark.

Evaluates autonomous driving stacks that produce waypoints with a static dataset using the PDM-Score metric. The PDM-score metric performs a pseudo-simulation by rolling out the trajectory and simulating all other actors via log-replay. This results in an open-loop evaluation that correlates with closed-loop performance.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 1–26 of 26 papers

TitleStatusHype
NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and BenchmarkingCode7
VADv2: End-to-End Vectorized Autonomous Driving via Probabilistic PlanningCode5
DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous DrivingCode5
Pseudo-Simulation for Autonomous DrivingCode4
Planning-oriented Autonomous DrivingCode4
Epona: Autoregressive Diffusion World Model for Autonomous DrivingCode3
End-to-End Driving with Online Trajectory Evaluation via BEV World ModelCode3
Enhancing End-to-End Autonomous Driving with Latent World ModelCode3
Hydra-MDP: End-to-end Multimodal Planning with Multi-target Hydra-DistillationCode3
TransFuser: Imitation with Transformer-Based Sensor Fusion for Autonomous DrivingCode3
Generalized Trajectory Scoring for End-to-end Multimodal PlanningCode3
GoalFlow: Goal-Driven Flow Matching for Multimodal Trajectories Generation in End-to-End Autonomous DrivingCode3
iPad: Iterative Proposal-centric End-to-End Autonomous DrivingCode2
Hydra-NeXt: Robust Closed-Loop Driving with Open-Loop TrainingCode1
0/1 Deep Neural Networks via Block Coordinate Descent—0
Bench2Drive-R: Turning Real World Data into Reactive Closed-Loop Autonomous Driving Benchmark by Generative Model—0
Centaur: Robust End-to-End Autonomous Driving with Test-Time Training—0
DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning—0
DrivingGPT: Unifying Driving World Modeling and Planning with Multi-modal Autoregressive Transformers—0
Finetuning Generative Trajectory Model with Reinforcement Learning from Human Feedback—0
PARA-Drive: Parallelized Architecture for Real-time Autonomous Driving—0
ReCogDrive: A Reinforced Cognitive Framework for End-to-End Autonomous Driving—0
ReSim: Reliable World Simulation for Autonomous Driving—0
TransDiffuser: End-to-end Trajectory Generation with Decorrelated Multi-modal Representation for Autonomous Driving—0
GaussianFusion: Gaussian-Based Multi-Sensor Fusion for End-to-End Autonomous DrivingCode0
World4Drive: End-to-End Autonomous Driving via Intention-aware Physical Latent World ModelCode0
Show:102550

No leaderboard results yet.