Driving on Registers

TL;DR

DrivoR is a transformer-based end-to-end autonomous driving system that uses camera-aware tokens for scene representation, enabling accurate, efficient, and interpretable driving performance.

Contribution

Introducing a novel transformer architecture with camera-aware register tokens for compact scene encoding in autonomous driving.

Findings

Outperforms or matches strong baselines on multiple benchmarks.

Reduces downstream computation significantly.

Provides interpretable sub-scores for driving aspects.

Abstract

We present DrivoR, a simple and efficient transformer-based architecture for end-to-end autonomous driving. Our approach builds on pretrained Vision Transformers (ViTs) and introduces camera-aware register tokens that compress multi-camera features into a compact scene representation, significantly reducing downstream computation without sacrificing accuracy. These tokens drive two lightweight transformer decoders that generate and then score candidate trajectories. The scoring decoder learns to mimic an oracle and predicts interpretable sub-scores representing aspects such as safety, comfort, and efficiency, enabling behavior-conditioned driving at inference. Despite its minimal design, DrivoR outperforms or matches strong contemporary baselines across NAVSIM-v1, NAVSIM-v2, and the photorealistic closed-loop HUGSIM benchmark. Our results show that a pure-transformer architecture, combined with targeted token compression, is sufficient for accurate, efficient, and adaptive end-to-end driving. Code and checkpoints will be made available via the project page.