This thesis explores reinforcement learning approaches for the antenna selection problem in Wi-Fi devices, aiming to provide adaptive solutions to the dynamic nature of wireless communi- cation channels. The problem is mathematically modeled using Markov Decision Processes (MDPs), and a reinforcement learn- ing framework is implemented to address this challenge. The study evaluates the performance of a Proximal Policy Optimiza- tion (PPO) algorithm, comparing it with traditional approaches. A key objective of this research is to apply reinforcement learn- ing and machine learning to real-world problems, demonstrating their advantages in optimizing antenna selection in dynamic com- munication scenarios.
Questa tesi esplora l’applicazione del Reinforcement Learning al problema della selezione delle antenne nei dispositivi Wi-Fi, con l’obiettivo di fornire soluzioni adattive alla natura dinamica dei canali di comunicazione wireless. Il problema viene modellato matematicamente attraverso i Processi Decisionali di Markov (MDP) e affrontato con un framework di apprendimento per rinforzo. In particolare, viene valutata la performance dell’algoritmo Proximal Policy Optimization (PPO), confrontandola con approcci tradizionali. L’obiettivo principale di questa ricerca è dimostrare come l’intelligenza artificiale e il reinforcement learning possano essere applicati a problemi reali, evidenziando i vantaggi nella selezione ottimale delle antenne in scenari di comunicazione dinamici.
Reinforcement learning based algorithm for antenna selection
DANESIN, FEDERICO
2024/2025
Abstract
This thesis explores reinforcement learning approaches for the antenna selection problem in Wi-Fi devices, aiming to provide adaptive solutions to the dynamic nature of wireless communi- cation channels. The problem is mathematically modeled using Markov Decision Processes (MDPs), and a reinforcement learn- ing framework is implemented to address this challenge. The study evaluates the performance of a Proximal Policy Optimiza- tion (PPO) algorithm, comparing it with traditional approaches. A key objective of this research is to apply reinforcement learn- ing and machine learning to real-world problems, demonstrating their advantages in optimizing antenna selection in dynamic com- munication scenarios.File | Dimensione | Formato | |
---|---|---|---|
Danesin_Federico.pdf
accesso aperto
Dimensione
12.14 MB
Formato
Adobe PDF
|
12.14 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/83735