A Power-Gap Elimination and Unsupervised Training (PEUT) Algorithm For Indoor LocaliSation (15208N)
For more information on this technology contact:
To achieve accurate location estimation in a 2D space a mobile device must obtain RSSs from at least 3 access points (APs). Most of the existing WLAN infrastructures, however, are deployed for best communication coverage and, therefore, one mobile device may receive signals from only 1 or 2 APs in many locations. As a result, accurate localisation cannot be achieved due to lack of information.
This invention addresses this problem providing a localisation solution for mobile devices based on two algorithms (PE - power-gap elimination and UT - unsupervised training) to eliminate training costs using historical RSS records on mobile phones and to deal with mobile phone diversity.
The performance of this invention is around 5 meters across conditions tested. Accuracy is, therefore, acceptable in certain applications when considering its zero training cost and ease of implementation in existing WLAN infrastructures.