| KEYWORD |
Deep Learning techniques for Non-Intrusive Load Monitoring
Tesi esterna in azienda
Parole chiave ARTIFICIAL INTELLIGENCE, ARTIFICIAL NEURAL NETWORKS, DEEP NEURAL NETWORKS, SMART GRIDS
Riferimenti EDOARDO PATTI
Riferimenti esterni Marco Castangia (marco.castangia@polito.it), Christian Camarda (christian@midorisrl.eu)
Gruppi di ricerca DAUIN - GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA, ELECTRONIC DESIGN AUTOMATION - EDA, Energy Center Lab, GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA, ICT4SS - ICT FOR SMART SOCIETIES
Tipo tesi SPERIMENTALE
Descrizione Non-Intrusive Load Monitoring (NILM) is a technique that enables to separate the electrical loads of individual appliances starting from the total electrical load of the household.
In the past decade, deep learning techniques proved particularly effective in detecting and extracting the power consumption of several devices from the aggregate power consumptions.
However, due to the costs and labor efforts involved in collecting device-specific power consumptions, researchers experienced a widespread lack of annotated data to train deep learning models.
This thesis is aimed at developing innovative data augmentation techniques that can improve the performance of deep learning models in the context of NILM.
For this purpose, the student will also implement state-of-the-art deep learning models in order to test and evaluate the proposed data augmentation methodology.
Conoscenze richieste Python
Scadenza validita proposta 07/08/2025
PROPONI LA TUA CANDIDATURA