Multi Fingerprint Map for Indoor Localisation

Fingerprinting is one of the location estimation technique used in indoor applications. It maps information about wireless signals (e.g. the RSS value) into spatial coordinates. Because WiFi is an ubiquitous communication technology, supported by smartphones, Fingerprinting-based localisation algorithms that use WiFi signals are suitable for LBS applications. Although good results can be achieved using Fingerprinting, this is not an error free localisation technique. The end-user of the LBS application can interfere with these algorithms. If a user that was facing an Access Point rotates 180, the received RSS from that Access Point will decrease (and vice-versa). Although the user did not move, this RSS variation might be interpreted as "the user moved". A possible solution to cope with this problem is to acquire data at different directions, at each spatial point, during the off-line phase. Multiple Fingerprint Maps, that also include direction information, can therefore be built. The correct map (or combination of maps) can then be chosen during the on-line phase. Because most smartphones have a magnetic sensor, it is possible to know which direction the user is facing, therefore this information can be used to select the best Fingerprint Map to be used by the Location Estimation Algorithm. With this approach it was possible to improve up to 10.3% the location estimation precision, when compared with the use of an FM generated by averaging data collected in all directions.