SMART: Advancing Scalable Map Priors for Driving Topology Reasoning

TL;DR

SMART introduces a scalable map prior model for driving topology reasoning that leverages standard maps and satellite data, improving lane understanding and reasoning performance without relying on sensor-specific training data.

Contribution

The paper presents SMART, a novel scalable approach that uses standard and satellite maps to learn map priors, enhancing topology reasoning independently of sensor configurations.

Findings

Achieves superior offline lane topology understanding using SD and satellite data.

Improves online topology reasoning accuracy by up to 28% on OpenLane-V2.

Demonstrates seamless integration with existing topology reasoning methods.

Abstract

Topology reasoning is crucial for autonomous driving as it enables comprehensive understanding of connectivity and relationships between lanes and traffic elements. While recent approaches have shown success in perceiving driving topology using vehicle-mounted sensors, their scalability is hindered by the reliance on training data captured by consistent sensor configurations. We identify that the key factor in scalable lane perception and topology reasoning is the elimination of this sensor-dependent feature. To address this, we propose SMART, a scalable solution that leverages easily available standard-definition (SD) and satellite maps to learn a map prior model, supervised by large-scale geo-referenced high-definition (HD) maps independent of sensor settings. Attributed to scaled training, SMART alone achieves superior offline lane topology understanding using only SD and satellite inputs. Extensive experiments further demonstrate that SMART can be seamlessly integrated into any online topology reasoning methods, yielding significant improvements of up to 28% on the OpenLane-V2 benchmark.