Autonomous Driving Seminal Papers

Goal: build a durable corpus of high-impact autonomous driving papers, prioritizing papers with strong citation footprints, lasting conceptual importance, or clear influence on later systems.

Collection rule

Use citation count as a filter, not a definition. The corpus should include:

• papers that exceed roughly 1000 citations,
• papers that introduced a durable concept even if newer or less cited,
• benchmark or system papers that reshaped evaluation or architecture choices.

Seed list by area

Perception

• PointNet
• PointNet++
• VoxelNet
• PointPillars
• SECOND
• PV-RCNN
• CenterPoint
• Lift, Splat, Shoot
• BEVFormer
• FB-BEV
• SurroundOcc
• DETR3D
• OccFormer
• BEVNeXt

Prediction

• DESIRE
• Social LSTM
• Trajectron / Trajectron++
• CoverNet
• MultiPath
• LaneGCN
• VectorNet
• TNT
• MTR

Planning / system

• ChauffeurNet
• Conditional Imitation Learning
• Learning by Cheating
• TransFuser
• TCP
• VAD
• VADv2
• UniAD

Evaluation / benchmarks / data

• KITTI
• nuScenes
• Waymo Open Dataset
• Argoverse / Argoverse 2
• CARLA
• NAVSIM

Ingest priorities

1. Build dataset and benchmark pages first because they anchor later method comparisons.
2. Ingest one canonical paper per cluster before adding near-duplicates.
3. Maintain explicit notes on whether each paper supports modular, hybrid, or e2e interpretations.

Already seeded in batch 01

• End To End Learning For Self Driving Cars
• End To End Driving Via Conditional Imitation Learning
• Carla An Open Urban Driving Simulator
• Chauffeurnet Learning To Drive By Imitating The Best And Synthesizing The Worst
• Learning By Cheating
• Vectornet Encoding Hd Maps And Agent Dynamics From Vectorized Representation
• Learning Lane Graph Representations For Motion Forecasting
• Lift Splat Shoot Encoding Images From Arbitrary Camera Rigs By Implicitly Unprojecting To 3D
• Nuscenes A Multimodal Dataset For Autonomous Driving
• Bevformer Learning Birds Eye View Representation From Multi Camera Images Via Spatiotemporal Transformers
• Transfuser Imitation With Transformer Based Sensor Fusion For Autonomous Driving
• Planning Oriented Autonomous Driving
• Vad Vectorized Scene Representation For Efficient Autonomous Driving

Added in batch 02 (AutoVLA corpus — planning/VLA overlap)

• Textual Explanations For Self Driving — BDD-X dataset, explainability
• Talk2Car — language command grounding
• Simlingo — CARLA challenge winner, vision-only VLA
• Orion — Bench2Drive SOTA
• Emma — Waymo industry-scale model
• Alpamayo R1 — NVIDIA production VLA
• Wote Bev World Model — BEV trajectory verification
• Drivemoe — MoE for driving
• Think Twice Before Driving Towards Scalable Decoders For End To End Autonomous Driving — Cascaded decoder refinement, CARLA SOTA
• Bevformer V2 Adapting Modern Image Backbones To Birds Eye View Recognition Via Perspective Supervision — Perspective supervision for backbone-agnostic BEV perception
• Driveadapter Breaking The Coupling Barrier Of Perception And Planning In End To End Autonomous Driving — Decoupled perception-planning via adapter module (ICCV 2023)
• Fb Bev Bev Representation From Forward Backward View Transformations — Unified forward-backward view transformation for BEV (ICCV 2023)
• Vadv2 End To End Vectorized Autonomous Driving Via Probabilistic Planning — Probabilistic planning via action vocabulary, successor to VAD
• Navsim Data Driven Non Reactive Autonomous Vehicle Simulation — Non-reactive simulation benchmark bridging open-loop and closed-loop evaluation (NeurIPS 2024)