Abstract

Indoor object tracking using low-cost optical sensors, Agata Brajdic

The purpose of this talk is to demonstrate applicability of using low-cost optical sensor technology for indoor object tracking. The tracking is performed by processing data from object-mounted electronic unit with optical sensor with the particle filter technique. Although inexpensive, such a system is expected to sustain object’s alternating velocities whilst maintaining an adequate level of positioning accuracy. Experience with the prototype implementation on a simple use case scenario will be reported. Finally, directions for extending this work will be discussed, which would serve as a ground for my future PhD research.

Modelling uncertainty in sensor data, Mattias Linnap

All sensor data has uncertainty. I will present a survey of recent techniques that embrace this fact using probabilistic models instead of trying to clean up the data. I will also consider issues and solutions in data provenance: tracking the sources and reliability of scientific data.