Application of cognitive radio frameworks to satellite systems
This project was developed for my undergraduate thesis in Computer Engineering at "Roma Tre" University.
It involved simulating a message exchange between a primary user and a satellite, with all studies centered around
Signal Detection Probability. The project was entirely implemented in Java following an analysis and design phase involving UML diagrams.
The development process encompassed a range of subjects encountered throughout my undergraduate studies, including fundamentals of telecommunications (the focus of the thesis itself), concurrent programming,
software analysis and design, and object-oriented programming.
The entire project can be found on GitHub at the following link: https://github.com/alegio98/TLCproject
Pokémon - AI challenge
The objective of this AI-themed Pokemon project/challenge was to introduce a new functionality to the Pokédex that allows us to predict the winner of a 1:1 battle. In practical terms, the goal was to "Develop a classifier capable of predicting the outcome of a battle." We utilized datasets containing comprehensive statistics of Pokémon, including attack, defense, abilities, attributes, and more.
The entire project can be found on GitHub at the following link: https://github.com/alegio98/Pokemon-ChallengeIA-Giugno21
IMPORTANT NOTE :
In my GitHub repository, you will find several projects following a similar structure, all related to AI challenges. These projects were undertaken during the university course : "Artificial Intelligence." I deliberately chose to keep some of the mistakes I made during the projects' development (at that time). Notably, a significant improvement in data analysis and training is evident throughout the course's progression. Eventually, I achieved a highly satisfactory score in the exam.
Fire Detection with ML
Project carried out for the university courses "Machine Learning & Intelligent Systems for the Internet," in collaboration with a colleague (information available on GitHub). The goal was to build a network capable of recognizing fires/incidents in outdoor environments. The performances of a network, whose architecture was designed for the purpose of fire detection within images and trained from scratch, are compared with those of a network where the transfer learning technique was employed.
The entire project can be found on GitHub at the following link: https://github.com/ErGren/FireDetection-ML
Collection.jl
The project was developed for the university course "Parallel and Distributed Computing," using the Julia programming language. It is entirely centered around the Linear Algebraic Representation (LAR) library, pioneered by Professor Alberto Paoluzzi.
The LAR library aims to provide a minimal characterization of the geometry and topology of a cellular complex by reconstructing the primary chain (also referred to as the 3D pipeline) for generating vertices, edges, faces, and cells of any solid object.
In my specific case, I had the opportunity to explore the evolution of 15 different types of complexes.
The chosen representation for mathematical description within LAR involves a collection of sparse integer matrices. These matrices constitute the core data structure employed by the library. The "Collection" project's code primarily revolves around
the development of solid (and non-solid) geometric figures such as cubes, squares, spheres, helicoids, toroids, and more.
The main objective of this endeavor was to optimize the code, enhancing its stability and reducing ambiguity. Additionally, significant effort was invested in addressing errors associated with other existing libraries.
This journey exemplifies my commitment to refining code to ensure both robustness and clarity, all while contributing to the advancement of LAR-based geometrical modeling.
The entire project can be found on GitHub at the following link: https://github.com/alegio98/Collection.jl