Wednesday, 3rd of March

10h (room R2014, 660 building) (see location)

Sarra Houiddi and Dominique Fourer

(IBISC, Université d'Évry Val d'Essonne)

Home Electrical Appliances Recognition using Relevant Features and Deep Neural Networks

Non-Intrusive Load Monitoring (NILM) refers to the analysis of the
aggregated current and voltage measurements
of Home Electrical Appliances (HEAs) recorded at the house electrical
panel. Such methods aim at identifying each HEA for
a better control of the energy consumption and for future smart grid

Here, we focus on the selection of relevant features for discriminating
HEAs. Our contributions are

First, we introduce a new publicly available dataset of individual HEAs
described by a large set of electrical features
computed from current and voltage measurements in steady-state conditions.
Second, we investigate five feature selection methods, including two new
ones: a proposed method based on a heuristic
approach and another one based on a deep neural network.

The impact of the feature selection methods is empirically investigated
using our proprietary dataset together with the publicly available
dataset PLAID, considering several classification algorithms.

All TAU seminars: here

Collaborateur(s) de cette page: guillaume .
Page dernièrement modifiée le Lundi 02 mars 2020 21:57:04 CET par guillaume.