BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Challenges and approaches to improving the accuracy of indoor posi
 tioning systems - Niki Trigoni\, Oxford University Department of Computer 
 Science
DTSTART:20121024T131500Z
DTEND:20121024T141500Z
UID:TALK40689@talks.cam.ac.uk
CONTACT:Stephen Clark
DESCRIPTION:While GPS has become the de facto standard as an outdoor posit
 ioning\ninfrastructure\, researchers are still actively working on the pro
 blem\nof indoor positioning. This is an exciting area with many practical\
 napplications from indoor navigation to location-based advertising.\n\nThe
 re are currently three key problems that challenge the adoption and\nwides
 pread use of indoor positioning systems. 1) non-line-of-sight\nsignal prop
 agation due to the cluttered indoor environment\, which\ncorrupts radio di
 stance measurements\, and leads to significant\nposition errors\; 2) spars
 ity of reference points\, which results in\nlarge uncertainty in position 
 estimation\; and 3) large spatial\nvariation in the accuracy of location s
 ensors\, which is difficult to\nmeasure empirically.\n\nIn this talk\, I w
 ill report on recent research carried out in the\nOxford Sensor Networks G
 roup to address these challenges. First\, I\nwill present two distinct app
 roaches to addressing the\nnon-line-of-sight (NLOS) problem\, an unsupervi
 sed approach\, based on\nthe theory of compressed sensing\, and a supervis
 ed learning approach\nbased on features of the signal distribution. Second
 \, I will show how\nit is possible to overcome the problem of infrastructu
 re sparsity by\nexploiting encounters (wireless contacts) between moving n
 odes.\nFinally\, I will present a novel Expectation Maximization algorithm
 \nthat can be used to learn the accuracy of different location sensors\nin
  different parts of a building\, without having to actually measure\nit em
 pirically.\n
LOCATION:Lecture Theatre 1\, Computer Laboratory
END:VEVENT
END:VCALENDAR