ezNavi: An Easy-to-Operate Indoor Navigation System Based on Pedestrian Dead Reckoning and Crowdsourced User Trajectories

摘要

Recently, researchers have paid attention to designing indoor navigation services for smartphone users. Conventional indoor navigation systems highly rely on well-known indoor information and prior training phase for localization, and thus it is time- and labor-consuming to bootstrap indoor navigation services. Without too much prior configurations, the proposed indoor navigation system, called ezNavi, utilizes trajectory information (reported by users) to generate indoor pathway and point of interests (POIs). The proposed system consists of a front-end mobile application (APP) and a back-end server. The mobile APP infers users' walking trajectories according to sensory values from smartphones. The back-end server processes crowdsourced trajectories with the help of deployed Bluetooth low energy beacon devices, and then produces pathways of the indoor environment. After obtaining more trajectories, the ezNavi can further refine the derived pathways to provide more efficient guidance services for users. Our experiment and prototyping results reveal that the ezNavi can effectively derive users' walking trajectories, produces indoor pathways, and indicates directions for users.