PIMAN
Title:
PIMAN ( Pôle de compétence en Inspection
et Maintenance Assistée
par langage Naturel)
Partner: Voice-Insight S.A., VRContext, Elsys
Funding: Région Bruxelles-Capitale
Project number:
Duration: 2007-2008
Project overview:
This project aims at assisting maintenance operators in industrial
environments by providing them, by the means of a wireless interface,
with an augmented reality of their surroundings. The three main
features of the interface are:
- A speech recognition
module, allowing speech based interaction with the interface
(Voice Insight).
- A localization service,
allowing to provide in real time the position of the operator
within its environment (MLG).
- A 3D visual representation
of the environment, featuring an avatar of the operator (VRContext,
Elsys).
Machine Learning Group
contribution:
The main contribution of the MLG will lie in the design of the
localization service. The localization system will rely on a
set of wireless emitters/transmitters (WET) embedded in the
environment, that can communicate with the interface carried
by the operator. The goal is to use the network formed by the
WETs as a GPS, where each WET acts as a local satellite from
which distances with the operator have to be inferred on the
basis of radio indicators. In a first stage, the MLG will carry
out real-world deployments to characterize the radio communication
profiles in industrial environments, and to identify indicators
that can be used to infer distances. On the basis of these results,
the MLG's role will be to assess and to determine the localization
technique (trilateration, multi-dimensional scaling, bounding
box) that provides the best position estimates given the distances
inferred. An overview of the research on wireless sensor networks in the Machine Learning Group is available on this poster.
MLG researcher involved:
Yann-Aël
Le Borgne
Mathieu
Van Der Haegen
Jean-Michel Dricot
|