The Hong Kong University of Science and Technology

Suining He

Shueng Han Gary Chan

When access point signals are altered, conventional fingerprint-based indoor localization techniques are note accuracy. Localization with altered access point and fingerprint updating can achieves accurate indoor localization and automatically update a fingerprint database with altered access points. Using subset sampling, the system detect the altered access points, filter them out by a received signal strength vector and find the location of a client. Given the received signal strength vectors received and the estimated location, the system can update a fingerprint database with the signal changes by applying a non-parametric Gaussian process regression method.

Facilitation of indoor localization and fingerprint updates of altered access point signals

Locating mobile devices based on a junction of signal tiles is disclosed herein. Mobile devices are mapped to a convex hull within which the mobile device is likely to be located in based on a received signal strength indicator (RSSI) associated with an access point (AP). Using the RSSI from a set of APs, the mobile device can be located within a junction of a set of convex hulls, or tiles using a linear programming function. To mitigate signal noise and improve the accuracy of the location determination, the localization system can use the RSSIs of the APs that have signals with a predetermined signal dynamic range that leads to smaller tiles. Additionally, the localization system can reduce the search complexity of matching fingerprints to APs by partitioning the space into clusters with reference points that have similar signal vectors.

Growing community of inventors

Hong Kong, China

Suining He