
Francesco Maria Ferrante
End-to-end Real-Time Simulation and Validation Platform for Vehicular Applications.
Rel. Angelo Bonfitto. Politecnico di Torino, Corso di laurea magistrale in Mechatronic Engineering (Ingegneria Meccatronica), 2025
Abstract: |
Nowadays, the automotive industry is going through a digital transformation: vehicles are evolving into software-defined systems with over-the-air (OTA) updates, feature-on-demand and rapid release cycles. But, still, the development pipeline, from desktop simulation tools to hardware- in-the-loop (HiL) rigs and in-vehicle ECUs, remains fragmented. Simulation frameworks, mid- dleware stacks and real-time test benches often do not interoperate seamlessly, in this way the end-to-end validation is delayed and the time-to-market increases. To address this gap, this the- sis proposes a solution for an end-to-end real-time automotive platform that decouples functional logic from both execution environment and specific use cases. By abstracting transport, timing and hardware layers, the platform can host different driver-assistance or vehicle-management module, such as mission optimization or driver-state monitoring. The platform is validated in two comple- mentary phases. In Phase 1 a ROS 2–based Driver-in-the-Loop chain integrates CARLA, Simulink vehicle dynamics, and the computation of some off-board modules, that could be run on cloud, and on-board (online) modules that can run in real-time. In Phase 2 the same online software modules are deployed on a dSPACE SCALEXIO real-time target and subsequently they are built as C code; a lightweight UDP/CAN link replaces ROS 2 to guarantee 10 Hz hard-real-time deadlines without changing the data model or the internal logic. A bidirectional cloud–edge pipeline under- pins both phases, enabling OTA feature updates, telemetry streaming and adaptive services driven by real driver behavior. Benchmarks across various hardware and communication configurations demonstrate that execution times consistently remain within the specified period, validating that the architecture operates within its deterministic time limit. Finally, only the transport and exe- cution layers swap between phases, the data schema and application modules remain unchanged, proving the platform to be hardware- and use-case agnostic with minimal reconfiguration effort. This flexible, modular architecture closes the loop between virtual prototyping and real-world de- ployment, marking a way for rapid in-vehicle experimentation and continuous OTA evolution of next-generation driver-assistance systems. |
---|---|
Relatori: | Angelo Bonfitto |
Anno accademico: | 2024/25 |
Tipo di pubblicazione: | Elettronica |
Numero di pagine: | 97 |
Informazioni aggiuntive: | Tesi secretata. Fulltext non presente |
Soggetti: | |
Corso di laurea: | Corso di laurea magistrale in Mechatronic Engineering (Ingegneria Meccatronica) |
Classe di laurea: | Nuovo ordinamento > Laurea magistrale > LM-25 - INGEGNERIA DELL'AUTOMAZIONE |
Aziende collaboratrici: | SENSOR REPLY S.R.L. CON UNICO SOCIO |
URI: | http://webthesis.biblio.polito.it/id/eprint/36508 |
![]() |
Modifica (riservato agli operatori) |