FASTopoWM: Fast-Slow Lane Segment Topology Reasoning with Latent World Models
This work addresses a key perception challenge for autonomous driving systems by improving lane topology reasoning, though it appears incremental as it builds on existing temporal propagation methods.
The paper tackles the problem of lane segment topology reasoning for autonomous driving by proposing FASTopoWM, a framework that integrates latent world models to enhance temporal perception, resulting in improved performance on the OpenLane-V2 benchmark with a lane segment detection mAP of 37.4% versus 33.6% and centerline perception OLS of 46.3% versus 41.5% compared to state-of-the-art methods.
Lane segment topology reasoning provides comprehensive bird's-eye view (BEV) road scene understanding, which can serve as a key perception module in planning-oriented end-to-end autonomous driving systems. Existing lane topology reasoning methods often fall short in effectively leveraging temporal information to enhance detection and reasoning performance. Recently, stream-based temporal propagation method has demonstrated promising results by incorporating temporal cues at both the query and BEV levels. However, it remains limited by over-reliance on historical queries, vulnerability to pose estimation failures, and insufficient temporal propagation. To overcome these limitations, we propose FASTopoWM, a novel fast-slow lane segment topology reasoning framework augmented with latent world models. To reduce the impact of pose estimation failures, this unified framework enables parallel supervision of both historical and newly initialized queries, facilitating mutual reinforcement between the fast and slow systems. Furthermore, we introduce latent query and BEV world models conditioned on the action latent to propagate the state representations from past observations to the current timestep. This design substantially improves the performance of temporal perception within the slow pipeline. Extensive experiments on the OpenLane-V2 benchmark demonstrate that FASTopoWM outperforms state-of-the-art methods in both lane segment detection (37.4% v.s. 33.6% on mAP) and centerline perception (46.3% v.s. 41.5% on OLS).