6-Axis Sensor Assisted Low Complexity High Accuracy-Visible Light Communication Based Indoor Positioning System

Chinnapat SERTTHIN  Tomoaki OHTSUKI  Masao NAKAGAWA  

Publication
IEICE TRANSACTIONS on Communications   Vol.E93-B   No.11   pp.2879-2891
Publication Date: 2010/11/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.E93.B.2879
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on Fundamental Issues on Deployment of Ubiquitous Sensor Networks)
Category: 
Keyword: 
visible light communication (VLC),  visible light communication identification (VLID),  field-of-view (FOV),  6-axis sensor,  root means square of error distance (RMSED),  switching estimated receiver position (SwERP),  

Full Text: PDF>>
Buy this Article




Summary: 
The authors focus on the improvement of Visible Light Communication Identification (VLID) system that provides positioning information via LED light bulb, which is a part of Visible Light Communication (VLC) system. The conventional VLID system provides very low positioning estimation accuracy at room level. In our approach, neither additional infrastructure nor modification is required on the transmitter side. On the receiver side, 6-axis sensor is embedded to provide 3-axis of Azimuth and 3-axis of Tilt angulations information to perform positioning estimation. We verify the proposed system characteristics by making both empirical and numerical analysis, to confirm the effectiveness of proposed system. We define two words to justify the characteristic of the proposed system, which are Field-of-View (FOV: ψc) Limit and Sensitivity (RXS) Limit. Both FOV and Sensitivity Limits have crucial impact on positioning estimation accuracy. Intuitively, higher positioning accuracy can be achieved with smaller FOV configuration in any system that has FOV. Conversely, based on system characteristics of VLID, we propose a positioning estimation scheme, namely Switching Estimated Receiver Position (SwERP) yields high accuracy even with wide FOV configuration. Cumulative Distribution Function (CDF) of error distance and Root Mean Square of Error Distance (RMSED) between experimental positions and estimated receiver positions are used to indicate the system performance. We collected 440 samples from 3 receivers' FOV configurations altogether 1320 samples within the experimental area of 1200 mm5000 mm2050 mm. The results show that with the proposed scheme, the achievable RMSEDs are in the range of 298 and 463 mm under different FOV configurations, which attained the maximum accuracy improvement over 80% comparing to the one without positioning estimation scheme. The proposed system's achievable accuracy does not depend on transmitters' orientation; only one transmitter is required to perform positioning estimation.