logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-01302020-194247


Tipo di tesi
Tesi di laurea magistrale
Autore
SCIVA, ALESSANDRO
Indirizzo email
a.sciva@studenti.unipi.it, alessandro.sciva@gmail.com
URN
etd-01302020-194247
Titolo
Deep Learning-based MIMO Indoor Positioning
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
COMPUTER ENGINEERING
Relatori
relatore Prof. Luise, Marco
relatore Prof. Sanguinetti, Luca
Parole chiave
  • Channel
  • CNN
  • Colab
  • Convolutional
  • Deep
  • Indoor
  • Learning
  • MIMO
  • Neural
  • Positioning
  • Tensorflow
  • Wireless
Data inizio appello
21/02/2020
Consultabilità
Completa
Riassunto
In this work a deep convolutional neural network was trained, with the purpose of indoor positioning. Starting from a MIMO propagation model, expressing wireless channel response as function of 3-dimensional position coordinates, a specific CNN have been trained. This was then tested with data coming from a real scenario. Furthermore, a dedicated CNN was trained considering this dataset, in order to compare perfomances with the previous one. Finally, an hybrid approach was provided: the amount of data coming from real scenario was expanded with gradually increasing portions of data generated from propagation model. A dedicated CNN was trained also in this case to evaluate overall performances.
File