Model-Based validation of Driver Drowsiness Detection System for ADAS

The work described in this Master’s Degree thesis was born after the collaboration with the company Maserati S.p.a, an Italian luxury car maker with its headquarters located in Modena, in the heart of the Italian Motor Valley, where I worked as a stagiaire in the Virtual Engineering team between September 2021 and February 2022. This work proposes the validation using real-world ECUs of a Driver Drowsiness Detection (DDD) system prototype based on different detection methods with the goal to overcome input signal losses and system failures. Detection methods of different categories have been chosen from literature and merged with the goal of utilizing the benefits of each of them, overcoming their limitations and limiting as much as possible their degree of intrusiveness to prevent any kind of driving distraction: an image processing-based technique for human physical signals detection as well as methods based on driver-vehicle interaction are used. A Driver-In-the-Loop simulator is used to gather real data on which a Machine Learning-based algorithm will be trained and validated. These data come from the tests that the company conducts in its daily activities so confidential information about the simulator and the drivers will be omitted. Although the impact of the proposed system is not remarkable and there is still work to do in all its elements, the results indicate the main advantages of the system in terms of robustness against subsystem failures and signal losses.

Analisi del tempo di reazione dei conducenti dei veicoli con metodologie innovative: l'influenza del sistema Adaptive Cruise Control

La presente tesi si pone l’obiettivo di studiare e comprendere l’influenza che i sistemi di assistenza alla guida (ADAS – Advanced Driver Assistance Systems), installati negli autoveicoli di nuova generazione, hanno sulla condotta di guida degli automobilisti, con particolare attenzione alla distrazione ed al workload che essi provocano. Punto centrale dell’analisi è il sistema Adaptive Cruise Control (ACC) che permette al guidatore del veicolo sia di mantenere una velocità di marcia costante sia di rilevare, tramite sensoristica, i veicoli che lo precedono, intervenendo sul sistema frenante e sulla centralina di controllo del motore, così da garantire il mantenimento della distanza di sicurezza selezionata. Lo studio, attraverso l’utilizzo di tecniche innovative, si sofferma, in particolare, sull’analisi del comportamento di guida dei conducenti a bordo di un veicolo dotato di Adaptive Cruise Control. Il fine della ricerca è quello di determinare se e quanto il sistema ACC influisca sul conducente in termini di carico di lavoro cognitivo e fisico e di livelli d’attenzione, concentrandosi sulla valutazione del tempo di reazione con sistema acceso o spento.

Rapid Control Prototyping of an L2+ Highway Assist System

Advanced Driver Assistance Systems (ADAS) are proving to have huge potential in road safety, comfort, and efficiency. In recent years, car manufacturers have equipped their high-end vehicles with Level 2 ADAS, which are, according to SAE International, systems that combine both longitudinal and lateral active motion control. These automated driving features, while only available in highway scenarios, appear to be very promising towards the introduction of hands-free driving. However, as they rely only on an on-board sensor suite, their continuative operation may be affected by the current environmental conditions: this prevents certain functionalities such as the automated lane change, other than requiring the driver to keep constantly the hands on the steering wheel. The enabling factor for hands-free highway driving proposed by Mobileye is the integration of high-definition maps, thus leading to the so-called Level 2+. This thesis was carried out during an internship in Maserati's Virtual Engineering team. The activity consisted of the design of an L2+ Highway Assist System following the Rapid Control Prototyping approach, starting from the definition of the requirements up to the real-time implementation and testing on a simulator of the brand new compact SUV Maserati Grecale. The objective was to enhance the current Level 2 highway driving assistance system with hands-free driving capability; for this purpose an Autonomous Lane Change functionality has been designed, proposing a Model Predictive Control-based decision-maker, in charge of assessing both the feasibility and convenience of performing a lane-change maneuver. The result is a Highway Assist System capable of driving the vehicle in a traffic scenario safely and efficiently, never requiring driver intervention.

Registratore di dati di evento (EDR). Regolamento europeo 2019/2144 e caso studio.

Il numero dei decessi per incidenti stradali nel mondo continua ad aumentare secondo ritmi preoccupanti, così come anche il numero dei feriti e i conseguenti costi sociali legati al fenomeno. In Europa, e più in generale nei paesi ad alto reddito, il progresso tecnologico e le azioni legislative hanno permesso di raggiungere importanti successi nel miglioramento della sicurezza stradale con conseguente riduzione del numero dei decessi. Purtroppo però, i numeri testimoniano come ci sia ancora molta strada da percorrere. In questa tesi vengono presentate numerose statistiche aggiornate sul fenomeno dell’incidentalità stradale e vengono analizzate le azioni intraprese a livello europeo volte al raggiungimento degli obbiettivi internazionali ed europei fissati per il 2030 ed il 2050. Viene poi analizzato nel dettaglio il recente Regolamento Europeo 2019/2144 con particolare riferimento al Registratore di Dati di Evento (EDR) rientrante tra gli Advanced Driver Assistance Systems (ADAS) resi obbligatori dallo stesso Regolamento Europeo 2019/2144. A tal proposito viene illustrato, tramite un caso studio verificatosi in Emilia-Romagna, come un dispositivo EDR possa essere utilizzato efficacemente per ricostruire la dinamica di un sinistro stradale.

An Embedded In-Cabin Lightweight High-Performance 3D Gaze Estimation System

Gaze estimation has gained interest in recent years for being an important cue to obtain information about the internal cognitive state of humans. Regardless of whether it is the 3D gaze vector or the point of gaze (PoG), gaze estimation has been applied in various fields, such as: human robot interaction, augmented reality, medicine, aviation and automotive. In the latter field, as part of Advanced Driver-Assistance Systems (ADAS), it allows the development of cutting-edge systems capable of mitigating road accidents by monitoring driver distraction. Gaze estimation can be also used to enhance the driving experience, for instance, autonomous driving. It also can improve comfort with augmented reality components capable of being commanded by the driver's eyes. Although, several high-performance real-time inference works already exist, just a few are capable of working with only a RGB camera on computationally constrained devices, such as a microcontroller. This work aims to develop a low-cost, efficient and high-performance embedded system capable of estimating the driver's gaze using deep learning and a RGB camera. The proposed system has achieved near-SOTA performances with about 90% less memory footprint. The capabilities to generalize in unseen environments have been evaluated through a live demonstration, where high performance and near real-time inference were obtained using a webcam and a Raspberry Pi4.