DISCOVER >>>

Sunday, 6 October 2019

Smartphone-based Activity Recognition and Multi-sensor Fusion based Indoor Positioning System


IoT2US Lab won 3rd place in the premier competition of the 10th International Conference on Indoor Positioning and Indoor Navigation (IPIN). There are fifteen teams in all from top universities (e.g. Ghent University), institutes or high tech companies (e.g. Intel, Tencent, LINE) and indoor positioning professional companies (e.g. Xihe Tech, Fineway, AraraDS) all over the world. This is one of the most world-famous two indoor positioning competitions, IPIN and Microsoft Indoor Localization Competition (IPSN) [2].


Team information

The work is a product of IoT2US Lab, in the School of (EECS), QMUL, in collaboration with UCL, Electronic Engineering department. Our team members are Bang Wu (QMUL), Chengqi Ma (UCL), Stefan Poslad (supervisor, QMUL), David Selviah (supervisor, UCL), Wei Wu (WHU), Xiaoshuai Zhang (QMUL) , Guangyuan Zhang (QMUL) , Zixiang Ma (QMUL).

Competition Goal

The goal of this competition track is to evaluate the performance of different indoor localization solutions based on the signals available to a smartphone (such as WiFi readings, inertial measurements, etc…) and received while a person is walking along several regular unmodified multi-floor buildings. This track is done off-site, so all data for calibration and evaluation is provided by competition organizers before the celebration of the IPIN conference. The competition teams can calibrate their algorithmic models with several databases containing readings from sensors typically found in modern mobile phones and some ground-truth positions. Finally, each team will compete using additional database files, but in this case, the ground-truth reference is not given and must be estimated by the competitors. This is an off-line competition where all competitors have the same data of the testing environment, so custom on-site calibration is not allowed.

Our Solution: Smartphone-based Activity Recognition and Multi-sensor Fusion based  Indoor Positioning System


Performance


Poster Display



No comments:

Post a Comment