面向探索与导航的语义感知预测建图
Semantic-Aware Predictive Mapping for Exploration and Navigation
Kenneth J. K. Ong · William W. J. Teo
中文摘要
预测建图可通过从部分占据观测推断未观测到的几何布局,辅助机器人探索与导航。然而,仅依赖占据信息的表示难以区分几何形态相近但语义不同的结构。室内门便是典型例子:门在占据图上呈现为被占据的单元格,形态与墙体相似,却暗示观测区域之外可能连接的房间或走廊。本工作研究语义门线索是否能改善此类模糊区域附近的预测性几何占据建图。研究者在 CogniPlan 数据集子集上修改部分占据图,插入由门引发的歧义,同时保持真值布局不变。在此修改数据集上,对比了一个仅使用几何信息的基线模型与一个语义线索模型,后者额外接收门通道输入。评估指标包括全图与 10 像素门区域掩膜上的 L1 误差、F1 分数和交并比(IoU)。全图性能两模型总体相近,但在门局部区域,语义线索模型呈现明显定性提升:L1 从 0.004342 降至 0.000025,F1 从 0.031311 升至 1.000000,IoU 从 0.015905 升至 1.000000。结果表明,语义线索可改善几何观测本身存在歧义的区域内的预测性占据补全效果。
关键要点
- 01问题:仅用占据表示无法区分室内门等几何相似但语义不同的结构,导致观测区域外的连通空间被遗漏
- 02方法:在 CogniPlan 数据集部分占据图中注入门引起的歧义,并为语义模型增加独立的门通道输入
- 03结果:全图指标两模型接近,但门局部区域 L1 由 0.004342 降至 0.000025,F1 与 IoU 均由约 0.02 升至 1.0
- 04局限:仅基于 CogniPlan 数据集子集进行修改与评估,门区域掩膜范围较窄,泛化性有待进一步验证
解读
尚无解读。
原始英文摘要
arXiv:2610.10382v1 Announce Type: new Abstract: Predictive mapping can support robotic exploration and navigation by estimating unseen geometric layouts from partial occupancy observations. However, occupancy-only representations may fail to distinguish semantically different structures with similar geometry. This is particularly relevant for indoor doors, which may appear as occupied cells like walls but indicate possible connected rooms or corridors beyond the observed region. This work investigates whether semantic door cues improve predictive geometric occupancy mapping around such ambiguous regions. We modify a subset of the CogniPlan dataset by inserting door-induced ambiguities into partial occupancy maps while keeping the ground-truth layouts unchanged. We compare a geometry-only control model with a semantic-cued model trained on the same modified dataset, where the semantic-cued model receives an additional door channel. Evaluation uses L1 error, F1 score, and Intersection over Union (IoU) over both the full map and a 10-pixel door-region mask. Full-map performance remains broadly similar between models, but localized door-region results show a clear qualitative improvement: L1 decreases from 0.004342 to 0.000025, while F1 and IoU improve from 0.031311 and 0.015905 to 1.000000 and 1.000000, respectively. These results suggest that semantic cues can improve predictive occupancy completion in regions where geometric observations alone are ambiguous.