The Head-Up Display (HUD) is a transparent display that allows to see information without changing viewpoint or taking the eyes off the outside world. In vehicles, it is implemented by projecting the information directly onto the windshield, allowing the driver to stay focused on driving, always keeping his eyes on the road. Nowadays, HUDs are used in cars to show speed, fuel level, navigation directions and other information already shown on the dashboard or by the infotainment system. The technological potential of windshield as a tool for augmented reality applications is much more interesting. Augmented reality is an interactive experience that brings components from the digital world into the real world. The objects of the real world are augmented through the integration of information generated by the computer, as to be perceived by the observer as a natural part of the environment. Hence, the user is informed intuitively and can stay focused completely on driving, without striving for understanding the content of the message. The goal of this thesis work, carried out in collaboration with Ferrari S.p.A., is to exploit HUDs in racing conditions. Showing driving information in augmented reality, the driver is able to intuitively understand which driving styles and trajectories must be followed during the lap to improve performances and driving skills on the track. Therefore, we implemented all the algorithms needed to correctly project information on the HUD. To generate the features to be reprojected, we developed a method to estimate the position of the car in the carriageway, based on computer vision algorithms and GPS data. Through the pinhole camera model, we derived the perspective equations for reprojection. Furthermore, we developed a calibration tool for the HUD to adapt the vision to the driver's height and position, as these are variables that influence correctness of reprojection. In conclusion, all the algorithmic was enclosed in a single software package.
L'Head-Up Display (HUD) è un display trasparente che permette la visualizzazione di dati senza distogliere lo sguardo dal mondo esterno. Nelle automobili viene implementato proiettando l’informazione direttamente sul parabrezza, permettendo al conducente di rimanere concentrato sulla guida, tenendo sempre lo sguardo sulla strada. Ad oggi, gli HUD vengono utilizzati nelle automobili per mostrare velocità, livello di carburante, indicazioni di navigazione e altre informazioni già mostrate nel cruscotto della vettura o sullo schermo dell’infotainment. Il potenziale tecnologico del parabrezza come strumento per applicazioni di realtà aumentata è molto più interessante. La realtà aumentata è un'esperienza interattiva che porta componenti del mondo digitale nel mondo reale. Gli oggetti del mondo reale sono aumentati attraverso l'integrazione di informazioni generate dal computer, tali da essere percepite dall’osservatore come parte naturale dell’ambiente. In questo modo l’utente viene informato intuitivamente, senza doversi impegnare per comprendere il contenuto del messaggio, concentrandosi completamente sulla guida. L’obiettivo di questa tesi, svolta in collaborazione con Ferrari S.p.A., è quello di sfruttare gli HUD in condizioni di guida sportiva. Mostrando informazioni di guida in realtà aumentata, il pilota riesce a comprendere in modo intuitivo quali stili di guida e quali traiettorie mantenere durante il giro in pista, migliorando la performance sul giro e le capacità di guida in circuito. A tal fine, abbiamo implementato l'algoritmica necessaria a riproiettare correttamente le informazioni su HUD. Per generare le feature da riproiettare abbiamo messo a punto un metodo per stimare la posizione della vettura nella carreggiata, basato su algoritmi di computer vision e dati GPS. Tramite il modello camera pinhole, abbiamo derivato le relazioni prospettiche per la riproiezione. Inoltre, abbiamo sviluppato un tool di calibrazione dell’HUD per adattare la visione alla statura e alla posizione del conducente, trattandosi di variabili che influenzano la correttezza della riproiezione. In conclusione, tutta l’algoritmica è stata racchiusa in un unico pacchetto software.
Analisi e sviluppo di un head-up display per applicazioni di guida sportiva
FRANCESCHETTI, LUCA
2017/2018
Abstract
The Head-Up Display (HUD) is a transparent display that allows to see information without changing viewpoint or taking the eyes off the outside world. In vehicles, it is implemented by projecting the information directly onto the windshield, allowing the driver to stay focused on driving, always keeping his eyes on the road. Nowadays, HUDs are used in cars to show speed, fuel level, navigation directions and other information already shown on the dashboard or by the infotainment system. The technological potential of windshield as a tool for augmented reality applications is much more interesting. Augmented reality is an interactive experience that brings components from the digital world into the real world. The objects of the real world are augmented through the integration of information generated by the computer, as to be perceived by the observer as a natural part of the environment. Hence, the user is informed intuitively and can stay focused completely on driving, without striving for understanding the content of the message. The goal of this thesis work, carried out in collaboration with Ferrari S.p.A., is to exploit HUDs in racing conditions. Showing driving information in augmented reality, the driver is able to intuitively understand which driving styles and trajectories must be followed during the lap to improve performances and driving skills on the track. Therefore, we implemented all the algorithms needed to correctly project information on the HUD. To generate the features to be reprojected, we developed a method to estimate the position of the car in the carriageway, based on computer vision algorithms and GPS data. Through the pinhole camera model, we derived the perspective equations for reprojection. Furthermore, we developed a calibration tool for the HUD to adapt the vision to the driver's height and position, as these are variables that influence correctness of reprojection. In conclusion, all the algorithmic was enclosed in a single software package.| File | Dimensione | Formato | |
|---|---|---|---|
|
2018_12_Franceschetti.pdf
non accessibile
Descrizione: Testo della tesi (aggiornato)
Dimensione
66.36 MB
Formato
Adobe PDF
|
66.36 MB | 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/10589/144865