Loading [a11y]/accessibility-menu.js
Crowdsourcing-Based Magnetic Map Generation for Indoor Localization | IEEE Conference Publication | IEEE Xplore

Crowdsourcing-Based Magnetic Map Generation for Indoor Localization


Abstract:

We propose an automatic map construction algorithm that assembles many users' trajectories to a FingerPrinting (FP) map of an indoor environment. Our crowdsourcing approa...Show More

Abstract:

We propose an automatic map construction algorithm that assembles many users' trajectories to a FingerPrinting (FP) map of an indoor environment. Our crowdsourcing approach uses only the inertial sensors of mobile devices, neither dedicated infrastructure nor floor plan information is required. We identify overlapping segments of the trajectories, which are obtained by a simple pedestrian dead reckoning algorithm, by clustering magnetic and heading features. These clusters are used as landmarks in a novel geometric map construction process. Finally, the FP map is derived from the geometric map by exploiting the recorded sensor data of the individual trajectories. Initial results of recordings in a typical office environment show that the algorithm is able to construct a magnetic FP map which closely matches a manually generated magnetic map.
Date of Conference: 24-27 September 2018
Date Added to IEEE Xplore: 15 November 2018
ISBN Information:

ISSN Information:

Conference Location: Nantes, France

References

References is not available for this document.