PiLoc: A self-calibrating participatory indoor localization system

Abstract

Abstract-While location is one of the most important context information in mobile and ubiquitous computing, large-scale deployment of indoor localization system remains elusive.In this work, we propose PiLoc, an indoor localization system that utilizes opportunistically sensed data contributed by users. Our system does not require manual calibration, prior knowledge and infrastructure support. The key novelty of PiLoc is that it merges walking segments annotated with displacement and signal strength information from users to derive a map of walking paths annotated with radio signal strengths.We evaluate PiLoc over 4 different indoor areas. Evaluation shows that our system can achieve an average localization error of 1.5m.

iMap: Automatic inference of indoor semantics exploiting opportunistic smartphone sensing

“…Automatic inference of indoor maps has gained popularity recently [8], [15]. While Jigsaw [8] requires the manual cooperation of users to take photos, the approach used in PiLoc [15] further automates the mapping process by matching WiFi signal spectrum of user contributed walking trajectories and eliminates the need of manual assistance from participating users.…”

“…While Jigsaw [8] requires the manual cooperation of users to take photos, the approach used in PiLoc [15] further automates the mapping process by matching WiFi signal spectrum of user contributed walking trajectories and eliminates the need of manual assistance from participating users. The map construction part of iMap leverages the idea of WiFi spectrum matching proposed by our prior work [15], and we introduce the crowdsourced indoor map construction in this section.…”

“…That is, to capture each step taking event using accelerometer and to estimate the heading direction of each step through sensor fusion of all IMU sensors. While these walking trajectories alone suffer from cumulated error, the error can be corrected by merging different user walking trajectories [15]. Most importantly, merging walking trajectories of different users also automatically generates the map of walking paths of the indoor environment.…”

“…One important observation highlighted by [15] is, when smartphone users walk along different indoor walking paths, different signal evolution trends of WiFi access points can be observed as shown in Figure 1(a). These signal evolution trends can be stable and unique for the same walking paths even over different time periods and different phone models.…”

“…By merging crowdsourced walking trajectories captured through smartphone inertial sensing, the indoor map of walking paths can be automatically constructed. This part of the system is based on our prior work on indoor localization [15]. Generated map forms the first-level semantic of the indoor environment and is used as the basis for the next level annotation.…”