发表机构
Korea Advanced Institute of Science and Technology (KAIST)（韩国科学技术院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对动态城市环境下在线高清地图构建的时序稳定性问题，提出即插即用模块MapTCL，通过双向矢量一致性学习与栅格地图一致性学习联合训练，在两个基准上显著提升基线模型性能且无额外推理开销。
AI 中文摘要
在动态城市环境中，由于移动物体和遮挡，构建可靠的在线高清地图仍然具有挑战性。尽管近期研究采用特征级时序融合来解决该问题，但这些方法仅依赖逐帧的真值监督，因此缺乏明确的目标来直接惩罚连续在线高清地图之间的几何噪声和时序抖动。为解决这一问题，我们提出MapTCL，一种辅助训练策略，通过双向对齐构建当前帧与过去帧之间的时序一致性损失。具体而言，双向矢量一致性学习（BVCL）将关联的过去和当前矢量实例之间的几何与语义差异建模为辅助损失；我们还采用栅格地图一致性学习（RCL）作为额外损失，以稳定密集的鸟瞰图（BEV）特征。通过联合使用这两种损失进行训练，MapTCL提升了生成高清地图的时序稳定性。在两个标准基准上的大量实验证明了该方法的有效性。作为一种通用的即插即用模块，MapTCL可持续增强现有基线模型，在nuScenes上实现+3.7 mAP和+2.8 C-mAP的提升，在Argoverse 2上实现+3.1 mAP和+2.5 C-mAP的提升，且无额外推理开销。
英文摘要
Constructing reliable online HD maps remains challenging in dynamic urban environments due to moving objects and occlusions. While recent works employ feature-level temporal fusion to address this, they rely solely on per-frame ground truth supervision. Consequently, they lack an explicit objective to directly penalize the geometric noise and temporal jitter between consecutive online HD maps. To address this, we propose MapTCL, an auxiliary training strategy that formulates temporal consistency loss between current and past frames via bidirectional alignment. Specifically, Bidirectional Vector Consistency Learning (BVCL) models the geometric and semantic discrepancies between associated past and current vector instances as an auxiliary loss. We also employ Raster map Consistency Learning (RCL) as an additional loss to stabilize dense BEV features. By jointly training with these dual losses, MapTCL improves the temporal stability of generated HD maps. Extensive experiments on two standard benchmarks demonstrate the effectiveness of our approach. As a versatile plug-and-play module, MapTCL consistently enhances existing baseline models, achieving gains of +3.7 mAP & +2.8 C-mAP on nuScenes and +3.1 mAP & +2.5 C-mAP on Argoverse 2 without additional inference overhead.
CommentsAccepted at 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
