Driving simulators have become indispensable tools in modern automotive engineering, as they allow for the safe and repeatable testing of vehicle dynamics, human-machine in- teractions, and Advanced Driver Assistance Systems (ADAS). The effectiveness of these simulators relies heavily on the realism of their virtual environments. In particular, the topological and semantic accuracy and surface detailing of the simulated tracks directly influence the sensations transmitted to the driver. This thesis focuses on the generation and virtual reconstruction of real-world racetracks, aiming to define, compare, and val- idate methodologies for creating high-fidelity driving scenarios. The initial phase of the research addresses the software compatibility and environmental realism required to con- vert a Blender model of the Vairano circuit, originally generated from vehicle telemetry data, into a functional driving simulator scenario. Furthermore, to improve the level of fidelity, a dedicated Simulink model was developed to simulate the presence of the kerbs and the characteristic vertical oscillations. The second part of the study proposes novel methodologies for generating driving scenarios. Specifically, it examines two distinct ap- proaches based on the use of web-mapping software and image processing, evaluating their respective advantages and limitations. A fundamental principle underlying the pro- posed methods is the strict separation of the scenario’s graphical representation from its logical and topological structure, a concept already established in the first phase of the research. Furthermore, a dedicated section explores techniques for creating static visual elements, which are essential for the overall realism of the simulation. To demonstrate the robustness and adaptability of the proposed methodologies, they are applied not only to the Vairano circuit but also to the Tazio Nuvolari circuit. This second application proves the validity of the methods on tracks with different layouts. Finally, a comparison of the evaluated approaches is presented, using on-field measurements.
I simulatori di guida sono diventati strumenti indispensabili nella moderna ingegneria automobilistica, poiché consentono di testare in modo sicuro e ripetibile la dinamica del veicolo, le interazioni uomo-macchina e i sistemi avanzati di assistenza alla guida (ADAS). L’efficacia di questi simulatori dipende fortemente dal realismo dei loro ambienti virtu- ali. In particolare, l’accuratezza topologica e i dettagli superficiali dei tracciati simu- lati influenzano direttamente le sensazioni trasmesse al pilota. Questa tesi si concentra sulla generazione e sulla ricostruzione virtuale di circuiti reali, con l’obiettivo di definire, confrontare e validare metodologie per la creazione di scenari di guida ad alta fedeltà. La fase iniziale della ricerca affronta la compatibilità software e il realismo ambientale necessari per convertire un modello Blender del circuito di Vairano, generato originar- iamente a partire dai dati di telemetria del veicolo, in uno scenario di simulazione di guida funzionale. Per migliorare il livello di fedeltà, un modello Simulink dedicato è stato sviluppato per simulare la presenza dei cordoli e delle caratteristiche vibrazioni verticali. La seconda parte dello studio propone nuove metodologie per la generazione di scenari di guida. Nello specifico, esamina due approcci distinti basati sull’utilizzo di software di web-mapping e sull’elaborazione delle immagini, valutandone i rispettivi vantaggi e limiti. Un principio fondamentale alla base dei metodi proposti è la rigorosa separazione della rappresentazione grafica dello scenario dalla sua struttura logica e topologica, un concetto consolidato già nella prima fase della ricerca. Inoltre, una sezione dedicata esplora le tec- niche per la creazione degli elementi visivi statici, essenziali per il realismo complessivo della simulazione. Per dimostrare la robustezza e la generalizzabilità delle metodologie proposte, queste vengono applicate non solo al circuito di Vairano, ma anche al circuito Tazio Nuvolari. Questa seconda applicazione dimostra la validità dei metodi su tracciati con conformazioni differenti. Infine, viene presentato un confronto degli approcci valutati, utilizzando misure sul campo.
Reconstructing real world race tracks for driving simulators : methodologies, tools and comparative evaluation
Luzzi, Leonardo Shuhei
2025/2026
Abstract
Driving simulators have become indispensable tools in modern automotive engineering, as they allow for the safe and repeatable testing of vehicle dynamics, human-machine in- teractions, and Advanced Driver Assistance Systems (ADAS). The effectiveness of these simulators relies heavily on the realism of their virtual environments. In particular, the topological and semantic accuracy and surface detailing of the simulated tracks directly influence the sensations transmitted to the driver. This thesis focuses on the generation and virtual reconstruction of real-world racetracks, aiming to define, compare, and val- idate methodologies for creating high-fidelity driving scenarios. The initial phase of the research addresses the software compatibility and environmental realism required to con- vert a Blender model of the Vairano circuit, originally generated from vehicle telemetry data, into a functional driving simulator scenario. Furthermore, to improve the level of fidelity, a dedicated Simulink model was developed to simulate the presence of the kerbs and the characteristic vertical oscillations. The second part of the study proposes novel methodologies for generating driving scenarios. Specifically, it examines two distinct ap- proaches based on the use of web-mapping software and image processing, evaluating their respective advantages and limitations. A fundamental principle underlying the pro- posed methods is the strict separation of the scenario’s graphical representation from its logical and topological structure, a concept already established in the first phase of the research. Furthermore, a dedicated section explores techniques for creating static visual elements, which are essential for the overall realism of the simulation. To demonstrate the robustness and adaptability of the proposed methodologies, they are applied not only to the Vairano circuit but also to the Tazio Nuvolari circuit. This second application proves the validity of the methods on tracks with different layouts. Finally, a comparison of the evaluated approaches is presented, using on-field measurements.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_07_Luzzi_Tesi.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Testo della tesi
Dimensione
69.57 MB
Formato
Adobe PDF
|
69.57 MB | Adobe PDF | Visualizza/Apri |
|
2026_07_Luzzi_Executive Summary.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Executive Summary
Dimensione
481.39 kB
Formato
Adobe PDF
|
481.39 kB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/260321