| Popis: |
Recently, there has been much interest in positioning based on the widespreading WLAN technology, notably observed in the increasing number of hotspots and mobile devices equipped with WLAN interface. One technique to use WLAN for positioning is location fingerprinting, where WLAN networks in preselected sample locations are collected and used as fingerprints for those locations. However, to collect such fingerprints, existing services typically need to employ many skilled wardrivers who scan networks in the streets. This approach turns out to be very costly, especially when a large scale system coverage with acceptable positioning accuracy is required. Therefore, we propose WBroximity as a novel solution for the aforementioned concerns. With WBroximity, not only WLAN but also Bluetooth fingerprints are collected, therefore benefitting from the short range of Bluetooth to enable more precise positioning. Furthermore, such hybrid fingerprints are collected by using the paradigm of participatory sensing, thus cutting the extra costs needed to employ special personnel for this task, and allowing the system coverage to expand to wherever participants reach. In this paper, we present the technical details of realizing WBroximity as a location provider and its usage for collecting real fingerprint datasets. We evaluate the achieved accuracy in light of combining WLAN and Bluetooth, and the inherent aspects of participatory sensing, like number of participants and quality of participation. We give also an initial design and evaluation of a countermeasure to mitigate the effects of malicious participation. |
