Employing User Feedback for Fast, Accurate, Low-Maintenance Geolocationing

  • Authors:
  • Ezekiel S. Bhasker;Steven W. Brown;William G. Griswold

  • Affiliations:
  • -;-;-

  • Venue:
  • PERCOM '04 Proceedings of the Second IEEE International Conference on Pervasive Computing and Communications (PerCom'04)
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

One way to improve inferences on sensor data is to tunethe algorithms through a time-consuming offline procedure.A less expensive, and potentially more accurate method isto use an online procedure based on feedback from users,who often know best what the data means to them. Wepresent a method for user-assisted location inference basedon 802.11b wireless signal strengths. A user 'corrects' systemgeolocations by clicking on a map, recording a 'virtualaccess point' (VAP) at the selected point for future inferences.A best VAP is selected using simple criteria, includingthe VAP's creator. This permits using other's VAPswhile getting their own if one exists, capturing user-specificbehavior. The system is also self-maintaining with respectto changing access point deployments. Indoor experimentsshow very good accuracy for this simple method.