The evolution toward Cooperative, Connected and Automated Mobility (CCAM) is transforming road transportation into a cooperative system in which vehicles, Vulnerable Road Users (VRUs), (including pedestrians, cyclists, etc,) and infrastructure exchange information through heterogeneous wireless technologies and integrate Artificial Intelligence (AI) algorithms for prediction and decision support. This paradigm introduces several challenges while enabling safety-oriented services such as cooperative awareness, cooperative perception, and remote assistance. In particular, these services impose performance, interoperability, and context-awareness requirement, which must be met even in the presence of heterogeneous (and potentially resource-constrained) platforms in terms of computational capabilities, communication interfaces, and supported messaging protocols. In this context, the thesis addresses the design and validation of integrated sensing, communication, perception, and control solutions to enable safe and efficient cooperative mobility services. In particular, the work combines theoretical investigation, architectural design, and experimentation on real prototypes for vehicles and electric-bicycles (e-bike). The first research line focuses on VRUs, with particular interest on e-bikes. Through the development of on-board hardware prototypes for e-bike, data acquisition and the analysis of mobility patterns jointly with the dynamics of VRU Awareness Message (VAM) generation, new message triggering mechanisms were proposed. These triggers rely on cyclist stability, in line with standardization bodies recommendations, with the objective of increasing road safety in mixed traffic scenarios (vehicles and e-bikes) that characterize CCAM settings. The thesis also includes the design, implementation, and validation of an open sensing, perception, and communication platform for connected vehicles. The architecture combines low-cost devices for embedded sensing and perception (e.g., cameras) with a multi-radio On-Board Unit (OBU) that supports short-range connectivity via ITS-G5 and long-range connectivity via 5G. Machine learning (ML) models are appropriately adapted to the operating context to improve surrounding environment detection and to estimate the driver’s attentional state, ultimately enhancing driving safety. The final part of the thesis considers Teleoperated Driving, a crucial service in the CCAM ecosystem, given the pressure it puts on the network infrastructure, the tight delivery constraints and the presence of human-in-the-loop. The thesis proposes and validates a solution for remote vehicle control that leverages vehicle proximity to dynamically form a platoon and exploits the contextual perception of adjacent vehicles, achieving benefits in terms of reduced radio resources required for video streaming toward the remote operator. Overall, the thesis shows that context-driven solutions, embedded perception, and communication are a key enabler to achieve scalable CCAM services, particularly when human-in-the-loop operation and high bandwidth demanding services must be supported under realistic network constraints.
L’evoluzione verso la mobilità cooperativa, connessa e automatizzata (CCAM) sta trasformando il trasporto stradale in un sistema cooperativo in cui veicoli, Utenti Vulnerabili della Strada (VRU), (tra cui pedoni, ciclisti, etc.), e infrastrutture scambiano informazioni tramite tecnologie wireless eterogenee e integrano algoritmi di Intelligenza Artificiale (IA) per la previsione e il supporto decisionale. Questo paradigma introduce diverse sfide, consentendo al contempo servizi orientati alla sicurezza quali la consapevolezza cooperativa, la percezione cooperativa e l’assistenza remota. In particolare, questi servizi impongono requisiti di prestazione, interoperabilità e consapevolezza del contesto, che devono essere soddisfatti anche in presenza di piattaforme eterogenee (e potenzialmente con risorse limitate) in termini di capacità computazionali, interfacce di comunicazione e protocolli di messaggistica supportati. In questo contesto, la tesi affronta la progettazione e la validazione di soluzioni integrate di rilevamento, comunicazione, percezione e controllo per consentire servizi di mobilità cooperativa sicuri ed efficienti. In particolare, il lavoro combina indagine teorica, progettazione architettonica e sperimentazione su prototipi reali per veicoli e biciclette elettriche (e-bike). La prima linea di ricerca è dedicata agli utenti vulnerabili della strada (VRU), con particolare focus sulle e-bike. Tramite lo sviluppo di prototipi hardware a bordo e-bike per l’acquisizione dati e l’analisi dei pattern di mobilità congiuntamente alla dinamica di generazione dei messaggi VRU Awareness Message (VAM), sono stati proposti nuovi meccanismi di trigger per l’invio dei messaggi. Tali trigger si basano sulla stabilità del ciclista, in linea con le raccomandazioni degli enti di standardizzazione, con l’obiettivo di aumentare la sicurezza stradale in scenari di traffico misto (veicoli ed e-bike) tipici dei contesti CCAM. La tesi include inoltre la progettazione, implementazione e validazione di una piattaforma aperta di sensing, percezione e comunicazione per veicoli connessi. L’architettura combina dispositivi a basso costo per sensing e percezione integrata, (ad esempio videocamere), e una On-Board Unit (OBU) multi-radio che supporta la connettività a corto raggio tramite ITS-G5 e a lungo raggio tramite 5G. Modelli di Machine Learning (ML) sono opportunamente adattati al contesto operativo per migliorare la rilevazione dell’ambiente circostante e stimare lo stato di attenzione del conducente al fine di migliorare la sicurezza di guida. La parte finale della tesi prende in esame la guida teleoperata (ToD), un servizio cruciale nell’ecosistema CCAM, data la pressione che essa esercita sull’infrastruttura di rete, i rigidi vincoli di consegna e la presenza dell’uomo nel loop di controllo (human-in-the-loop). E’ quindi proposta e validata una soluzione per il controllo remoto dei veicoli che sfrutta la prossimità tra veicoli per formare dinamicamente un convoglio, sfruttando anche la percezione del contesto dei veicoli adiacenti, con benefici in termini di riduzione delle risorse radio utilizzate per lo streaming video all’operatore remoto. Nel complesso, la tesi mostra che soluzioni context-driven, percezione embedded e comunicazione rappresentano un abilitatore chiave per ottenere servizi CCAM scalabili, in particolare quando operazioni human-in-the-loop e servizi ad alta richiesta di banda devono essere supportati sotto vincoli di rete realistici.
Advanced Solutions for the CCAM ecosystem: Design, Prototyping and Assessment / Zappalà, Domenico Mario. - (2026 Apr 17).
Advanced Solutions for the CCAM ecosystem: Design, Prototyping and Assessment
Zappalà, Domenico Mario
2026-04-17
Abstract
The evolution toward Cooperative, Connected and Automated Mobility (CCAM) is transforming road transportation into a cooperative system in which vehicles, Vulnerable Road Users (VRUs), (including pedestrians, cyclists, etc,) and infrastructure exchange information through heterogeneous wireless technologies and integrate Artificial Intelligence (AI) algorithms for prediction and decision support. This paradigm introduces several challenges while enabling safety-oriented services such as cooperative awareness, cooperative perception, and remote assistance. In particular, these services impose performance, interoperability, and context-awareness requirement, which must be met even in the presence of heterogeneous (and potentially resource-constrained) platforms in terms of computational capabilities, communication interfaces, and supported messaging protocols. In this context, the thesis addresses the design and validation of integrated sensing, communication, perception, and control solutions to enable safe and efficient cooperative mobility services. In particular, the work combines theoretical investigation, architectural design, and experimentation on real prototypes for vehicles and electric-bicycles (e-bike). The first research line focuses on VRUs, with particular interest on e-bikes. Through the development of on-board hardware prototypes for e-bike, data acquisition and the analysis of mobility patterns jointly with the dynamics of VRU Awareness Message (VAM) generation, new message triggering mechanisms were proposed. These triggers rely on cyclist stability, in line with standardization bodies recommendations, with the objective of increasing road safety in mixed traffic scenarios (vehicles and e-bikes) that characterize CCAM settings. The thesis also includes the design, implementation, and validation of an open sensing, perception, and communication platform for connected vehicles. The architecture combines low-cost devices for embedded sensing and perception (e.g., cameras) with a multi-radio On-Board Unit (OBU) that supports short-range connectivity via ITS-G5 and long-range connectivity via 5G. Machine learning (ML) models are appropriately adapted to the operating context to improve surrounding environment detection and to estimate the driver’s attentional state, ultimately enhancing driving safety. The final part of the thesis considers Teleoperated Driving, a crucial service in the CCAM ecosystem, given the pressure it puts on the network infrastructure, the tight delivery constraints and the presence of human-in-the-loop. The thesis proposes and validates a solution for remote vehicle control that leverages vehicle proximity to dynamically form a platoon and exploits the contextual perception of adjacent vehicles, achieving benefits in terms of reduced radio resources required for video streaming toward the remote operator. Overall, the thesis shows that context-driven solutions, embedded perception, and communication are a key enabler to achieve scalable CCAM services, particularly when human-in-the-loop operation and high bandwidth demanding services must be supported under realistic network constraints.| File | Dimensione | Formato | |
|---|---|---|---|
|
PhD Thesis_Zappalà Domenico Mario_XXXVIII ciclo.pdf
accesso aperto
Tipologia:
Tesi di dottorato
Licenza:
Creative commons
Dimensione
10.11 MB
Formato
Adobe PDF
|
10.11 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


