Embarking on the Autonomous Journey - A Strikingly Engineered Car Control System Design

Aug 2, 2026

Upload by Anda

The thesis develops an autonomous car control system using the Raspberry Pi as the core processing unit, aiming to combine innovation with sustainability in modern transportation. Two predictive components are integrated: a Convolutional Neural Network for lane detection and a traffic signs detector. Real-time decision needs are addressed with a Proportional Integral Derivative controller and a web server to process outputs from trained neural networks. Unity and the Meta Quest 2 VR headset support visual testing via AR/VR-enhanced evaluation.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages96

Reading Time4 h 2 min

AccessFree to download

Introduction

Frame of reference

Our Proposal

Related Work

Sensor Integration in Autonomous Driving Systems

Visualization and AR Integration

Hardware Setup

The car components

Software Architecture

System Design and Implementation

Data Collection and Processing

Data Collection and Processing for Lane Detection model

Data Collection and Processing for Traffic Signs Recognition model

Deep Learning Model

Lane Detection model

Traffic Signs Detection model

GPS Localization

Implementation

System Control

The principle of PID control

Implementing PID on Raspberry Pi

AR & VR Headset

Implementation

Conclusions and Future Works

Master Thesis in ICT for Internet & Multimedia Embarking on the Autonomous Journey: A Strikingly Engineered Car Control System Design Master Candidate Kristel Çoçoli Supervisor Leonardo Badia Date : 10 October Academic Year 2022/2023 I dedicate this to my most inspirational professor, Leonardo, to my love, Henri, to my strength, my mommy, to my family, Enxhi, Alisa, Oriana. To the infinity and beyond. Abstract The evolution of autonomous cars represents a significant stride in modern transportation, combining innovation with sustainability. Central to this thesis is the development of an autonomous car control system with the Raspberry Pi as its primary processing unit. Integrated into this system are two predictive frameworks: a Convolutional Neural Network (CNN) for lane detection, and a traffic signs detector. Within this construct, the Raspberry Pi module acts as the center for a single car. To cater to its immediate decision-making needs, a Proportional Integral Derivative controller and web server were instituted to process data through the trained neural networks. To enhance it visually, Unity and the Meta Quest 2 VR headset have been used as a testing method. This study not only illuminates the engineering dynamics behind autonomous vehicles but also underscores the potential of augmented reality in automotive testing. ITALIANO: L’evoluzione dei veicoli autonomi rappresenta un significativo passo avanti nel trasporto moderno, combinando l’innovazione con la sostenibilità . Al centro di questa tesi vi è lo sviluppo di un sistema di controllo di auto autonome con il Raspberry Pi come unità di elaborazione principale. In questo sistema sono integrati due framework predittivi: una CNN per il rilevamento delle corsie e un rilevatore di segnali stradali. All’interno di questa struttura, il modulo Raspberry Pi agisce come centro per un’unica auto. Per soddisfare le sue immediate esigenze decisionali, è stato istituito un controllore Proportional Integral Derivative (PID) e un server web per elaborarei dati attraverso le reti neurali addestrate. Per migliorarne l’aspetto visivo, sonostati utilizzati Unity e il visore perlarealtà virtuale Meta Quest 2, come strumenti per il testing. Questo studio non solo esplora le dinamiche ingegneristiche delle auto autonome, ma sottolinea anche il potenziale della realtà aumentata nei test automobilistici. Contents List of Figures xi List of Tables xiii List of Algorithms xvii List of Code Snippets xvii List of Acronyms xix 1 Introduction 1 1. 1 Frame of reference . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1. 1 Autonomous Vehicles . . . . . . . . . . . . . . . . . . . . . 1 1.1.2 Machine Learning Methods .................. 2 1.1.3 Convolutional Neural Networks for Self-Driving Car Systems ............................... 4 1.2 Our Proposal ............................... 5 2 Related Work 7 2.1 Sensor Integration in Autonomous Driving Systems ........ 7 2.2 Object Detection and Behavioral Impacts .............. 8 2.3 Visualization and AR Integration ................... 10 2.4 Challenges in Autonomous Driving Systems ............ 10 3 Hardware Setup 13 3. 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.2 The car components ........................... 15 3.2.1 Raspberry Pi ........................... 15 3.2.2 The motors . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 CONTENTS 3.2.3 The camera module . . . . . . . . . . . . . . . . . . . . . . 16 3.2.4 The ultrasonic sensor . . . . . . . . . . . . . . . . . . . . . . 16 3.2.5 The GPS module . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2.6 The IMU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2.7 The assembled car . . . . . . . . . . . . . . . . . . . . . . . 20 4 Software Architecture 23 4.1 System Design and Implementation ................. 24 4.1.1 Technology Stack ..

FAQs

What are the main components of the proposed autonomous car control system?

The system centers on a Raspberry Pi and integrates a CNN for lane detection and a traffic signs detector. A PID controller and a web server handle immediate decision-making and data processing.

How does the thesis handle lane and traffic sign perception?

Lane detection uses a Convolutional Neural Network, supported by a data collection and augmentation workflow. Traffic sign recognition follows a dedicated dataset collection, preprocessing, training, and evaluation pipeline.

How are testing and visualization performed in the project?

Unity and the Meta Quest 2 VR headset are used to create a visual testing approach. The study also highlights the role of augmented reality in automotive testing.

Aug 2, 2026

Upload by Anda