A Novel Approach for Wi-Fi Fingerprinting Using Logical Sequences of Intelligent Checkpoints

G. Retscher, H. Hofer:
"A Novel Approach for Wi-Fi Fingerprinting Using Logical Sequences of Intelligent Checkpoints";
Vortrag: IGNSS 2015 Conference, Surfers Paradise, Gold Coast, Australien (eingeladen); 14.07.2015 - 16.07.2015; in:"IGNSS 2015", (2015), Paper-Nr. 3, 16 S.

[ Publication Database ]

Abstract:


Commonly location fingerprinting is employed for Wi-Fi positioning. In the training phase RSSI scans have to be measured at known reference points (RPs) which are usually distributed in a grid within the area of interest. The grid has to be rather dense and therefore RSSI measurements are very labour consuming. The newly developed approach aims at a significant reduction of the number of required RPs. For that purpose particular points are selected which we call intelligent checkpoints (iCPs). To navigate a user to a certain room in a building certain iCPs have to passed on the way. Doors, stairs and corridors can serve as iCPs. Hence, a logical sequence can be derived due to the interdependence of these points along the way. iCPs are twofold intelligent, i.e., because of the fact that they depend on the selection of the points for the RSSI scans and because of the logical sequence in their correct order along the path depending on the building. While navigating then always the following iCP is known due to the vector graph allocation in the fingerprinting database. Further constraints between the iCPs are definable, e.g. the heading of the moving user can be logically defined while walking along a corridor or even the step length while climbing stairs. Thus only a small limited number of iCPs needs to be tested when matching the current RSSI values. The required processing time is significantly reduced. Furthermore the scanned Wi-Fi Access Points can be weighted on the RSSI value level. From field tests in an office building it could be seen that the iCP approach achieves a higher success rate for correct matching of the RSSI fingerprints than conventional approaches. In average correct matching results of 90.0 % were achieved using a joint Wi-Fi database including training measurements of all employed smartphones. An even higher success rate is achieved if the same mobile device is used in both phases. Thus the positioning accuracy can be crucially increased.