User localisation using visual sensing and RF signal strength

In this paper, we simulate a museum scenario where we are concerned with identifying the exhibit that a person is observing. Many different technologies exist for determining the location of a user, including GPS, GSM and RF-id tags. Due to their low-power and passive nature, along with their relatively good accuracy, we examine the use of image-based localisation, alongside RF-based localisation for this task. In image-based localisation, we investigate if it is possible to determine a user's location given an image captured at their current location. In RF-based localisation, we can attempt localisation using a set of signal strength readings from detected wireless networks. As the two sources of data are complementary, we investigate different fusion strategies and measure the resulting increase in performance from using both systems.