Mapping indoor environments based on human activity

摘要: We present a novel approach to build approximate maps of structured environments utilizing human motion and activity. Our uses data recorded with suit which is equipped several IMUs detect movements person door opening closing events. In our we interpret the as constraints handling events landmark detections in graph-based SLAM framework. As cannot distinguish between individual doors, employ multi-hypothesis on top system deal high data-association uncertainty. result, able accurately robustly recover trajectory person. additionally take advantage fact that people traverse free space doors separate rooms geometric structure environment after graph optimization. evaluate experiments carried out different users types.