Politecnico di Torino (logo)

Artificial Neural Networks applied to Quality Prediction of a Wi-Fi link

Francesco Xia

Artificial Neural Networks applied to Quality Prediction of a Wi-Fi link.

Rel. Stefano Scanzio, Gianluca Cena. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2021

PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (3MB) | Preview

Artificial intelligence, and in particular machine learning, is one of the enabling technologies of Industry 4.0. It is successfully used in many application contexts and with different scopes, including, for example, in preventive maintenance, automated inspections, and optimization of communication processes. This thesis aims to evaluate whether and to what extent artificial neural networks (ANN), a particular application of machine learning, can be profitably used to predict the quality of a Wi-Fi channel in terms of frame delivery ratio. Specifically, we defined two approaches for this purpose: one based on ANNs and the other based on a more traditional approach that mimics current adaptive solutions. Then we tested the ability of each solution to predict the values of a target, which represents the state of a channel over time. Both mechanisms try to infer the future state of a wireless channel by analyzing its conditions in its recent past. For this purpose, data streams transmitted over multiple Wi-Fi channels have been sampled periodically, and these represent the starting data shared by both methods. The results obtained are encouraging, with the ANN models showing superior performance to the traditional method in almost all performance indicators examined. Our analysis also shows that the current methodology has ample room for improvement, especially in terms of performance achievable by ANN models. From our point of view, this approach has vast real-world applicability, with the primary goal of improving the reliability and resilience of a wireless system.

Relators: Stefano Scanzio, Gianluca Cena
Academic year: 2021/22
Publication type: Electronic
Number of Pages: 91
Corso di laurea: Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering)
Classe di laurea: New organization > Master science > LM-32 - COMPUTER SYSTEMS ENGINEERING
Aziende collaboratrici: UNSPECIFIED
URI: http://webthesis.biblio.polito.it/id/eprint/21312
Modify record (reserved for operators) Modify record (reserved for operators)