2013-微软印度研究院-Zee-Zer-Effrt-Crwdsurcing-fr-Indr-Lcalizatin.pdfVIP

Zee: Zero-Effort Crowdsourcing for Indoor Localization † † † ‡∗ Anshul Rai , Krishna Kant Chintalapudi , Venkata N. Padmanabhan , Rijurekha Sen † Microsoft Research India ‡ Indian Institute of Technology, Bombay ABSTRACT 1. INTRODUCTION Radio Frequency (RF) fingerprinting, based on WiFi or cellular sig- RF fingerprinting of WiFi signals is a popular approach to in- nals, has been a popular approach to indoor localization. However, door localization. Typically, there is an initial training or calibra- its adoption in the real world has been stymied by the need for site- tion phase during which received signal strength (RSS) measure- specific calibration, i.e., the creation of a training data set compris- ments from multiple WiFi access points are recorded at known lo- ing WiFi measurements at known locations in the space of interest. cations. Then, when a device is to be located, RSS measurements While efforts have been made to reduce this calibration effort using from proximate APs are matched against the training data, either modeling, the need for measurements from known locations still deterministically [3] or probabilistically [32], to estimate location. remains a bottleneck. In this paper, we present Zee – a system that The need for calibration is a key bottleneck since it is labour- makes the calibration zero-effort, by enabling training data to be intensive. Further, it needs to be repeated for each new space and crowdsourced without any explicit effort on the part of users.