Coherent Online Road Topology Estimation and Reasoning with Standard-Definition Maps
This work addresses the reliance on HD maps for autonomous driving, offering a more accessible solution using SD maps, though it appears incremental in improving existing prediction methods.
The paper tackles the challenge of coherent online construction of high-definition maps for autonomous vehicles by predicting lane segments, topology, and road boundaries using standard-definition maps, achieving performance that significantly outperforms previous methods.
Most autonomous cars rely on the availability of high-definition (HD) maps. Current research aims to address this constraint by directly predicting HD map elements from onboard sensors and reasoning about the relationships between the predicted map and traffic elements. Despite recent advancements, the coherent online construction of HD maps remains a challenging endeavor, as it necessitates modeling the high complexity of road topologies in a unified and consistent manner. To address this challenge, we propose a coherent approach to predict lane segments and their corresponding topology, as well as road boundaries, all by leveraging prior map information represented by commonly available standard-definition (SD) maps. We propose a network architecture, which leverages hybrid lane segment encodings comprising prior information and denoising techniques to enhance training stability and performance. Furthermore, we facilitate past frames for temporal consistency. Our experimental evaluation demonstrates that our approach outperforms previous methods by a large margin, highlighting the benefits of our modeling scheme.