Search

Browse

• Communities

My Space

• Login

Development of an application framework for digital twin-driven intelligent autonomous vehicles

Sarantinoudis Nikolaos

Full record

---

Summary

The integration of digital twins with autonomous vehicles is a promising yet insufficiently explored research domain. Digital twins provide virtual representations of physical systems that enable real-time synchronisation, monitoring, simulation, and optimisation throughout the system lifecycle. Although extensively studied and applied in industrial and manufacturing contexts, the adoption of digital twins in autonomous vehicles remains fragmented, lacking a unified methodological framework. This doctoral dissertation addresses this gap by systematically investigating existing applications, proposing a structured conceptual framework for digital twin deployment in autonomous vehicle domain and validating the framework experimentally through representative case studies.The research is structured in three main phases. The first phase comprises of a comprehensive literature review and bibliometric analysis, mapping the research and academic landscapes of digital twin applications in autonomous vehicles across land, aerial, surface, and underwater domains. Quantitative and qualitative analyses identify key trends, application areas, methodologies, and technological enablers. The study highlights an exponential growth of publications in the last five years, underlining both the increasing relevance of the field and the absence of standardised approaches that could support replication and scalability.Building upon these insights, the second phase formulates a conceptual application framework. The framework is based on four main pillars: (i) system requirements and definition, (ii) communication middleware and data flow, (iii) digital modelling and (iv) physical-digital integration and validation. It provides a formalised methodology for embedding digital twins into autonomous vehicles, from design and development to testing, deployment, and operation. The framework emphasises interoperability, scalability, and replicability, positioning digital twins as a bridging mechanism between theoretical models and real-world implementations.The third phase focuses on validation use cases. Three tests were conducted; two of them using scaled autonomous vehicle platforms and the third one using a high-fidelity simulator representing the physical system. The first validation employed a custom-made testing platform, while the second utilised a RoboRacer vehicle and extended this setup through hardware and sensor upgrades, enabling the evaluation of modularity and robustness. The third validation constitutes a scale-up by applying the framework to a full-scale simulated vehicle. Open-source ecosystems, such as ROS/ROS2, were used to establish bidirectional synchronisation between physical and digital environments, simulation and visualisation. The use cases demonstrated the framework’s capacity to support algorithm testing, perception and control integration, and performance evaluation under diverse operational conditions. Results confirmed the feasibility of reducing development costs, mitigating risks, and facilitating experimentation in scenarios that would otherwise be unsafe or impractical in real-world testing.The contributions of this dissertation are threefold. First, it provides a systematic mapping of the research domain, offering a consolidated reference for scholars and practitioners. Second, it introduces a novel conceptual framework that formalises the application of digital twins in autonomous vehicles, addressing a critical gap in the literature and leveraging the paradigm of digital twins from Industry 4.0, a well-established domain for digital twins applications. Third, it presents experimental evidence of the framework’s applicability and makes openly available resources to enable replication and extension by the research community. In conclusion, this dissertation advances the state of the art by establishing a structured approach to digital twin integration in autonomous vehicles. By combining theoretical analysis, methodological innovation, and empirical validation, it contributes both to academic knowledge and to practical implementation pathways. The proposed framework lays the foundation for safer, more reliable, and cost-efficient development and operation of intelligent autonomous systems, while also creates opportunities for future extensions in cross-domain applications and real-world deployments.

Available Files

Services