As of September 2025, Work Package 5 (WP5) has become one of the project’s main “translation layers”: turning iDriving’s research ideas into working prototypes, integrated services, and testable workflows that can be demonstrated and iteratively improved in real pilot contexts. Over the first half of the project, WP5 has progressed from early conceptualisation and requirements gathering to concrete deployment, integration, and early validation across its five tasks.
Building the backbone for the iDriving digital ecosystem
A major focus in the period leading up to September 2025 has been putting in place the practical infrastructure that allows iDriving components to work together as a single ecosystem. WP5 deployed core backbone services and operational tooling—such as the message and integration stack, CI/CD and DevOps foundations, secure access mechanisms, and the project’s shared development environment—so that tools developed across the consortium can be deployed, updated, and tested in a coordinated way. This work directly supported the preparation and submission of Deliverable D5.1 (First iDriving Prototype) and enabled end-to-end integration testing across components.
Just as importantly, WP5 has already used this backbone in practice: end-to-end integration tests were successfully executed, including a live rehearsal in Nevers that helped validate how multiple elements of the iDriving stack behave together under pilot-relevant conditions.
From indicators to intelligence: the Safety Criteria Catalogue becomes operational
WP5 is also the work package where safety “measurement” evolves into safety “intelligence”. In the earlier months (M01–M06), partners focused on defining and refining KPIs for traffic, safety, and maintenance, and on laying the groundwork for dynamic updates to the Safety Criteria Catalogue (SCC).
By September, this work had moved beyond definitions: the SCC advanced into a functional toolchain, including KPI computation templates, a KPI manager, and mapping/visualisation tools that can be used and tested within simulated environments.
In parallel, WP5’s scientific activity on SCC evolution has been pushing toward real-time predictive capability: an end-to-end workflow (from data ingestion to indicator computation, prediction and risk alert generation) was established, enriched with surrogate safety metrics (such as PET, TT, DRAC) to support mode-aware risk estimation for different road user categories.
Digital twins and simulation pipelines: testing safely, before testing on real roads
A defining WP5 contribution is the way it uses digital twins and simulation to accelerate development and validation. By September 2025, WP5 had advanced multi-scale simulation models and co-simulation pipelines (including SUMO-based models and CARLA–SUMO workflows), supporting synthetic dataset generation for surrogate safety analysis and predictive warning development across multiple pilot contexts.
One concrete example is the SUMO-based digital twin of Nevers, deployed to stream high-resolution vehicle trajectories and enable automated conflict detection—creating training data for SCC updates and predictive models.
Smarter maintenance: risk frameworks and AI-supported planning foundations
WP5 is not only about detecting risk; it is also about enabling proactive action. In September 2025, the work on maintenance-oriented intelligence had progressed through refinement of a Failure Mode and Effect Analysis (FMEA) framework, alignment of historical and contextual data for specific pilot contexts (including Karlovac), and structuring of core modules for an AI-optimised maintenance tool (risk assessment, health/logistics, scheduling).
The XR-enabled control centre: from requirements to early on-site validation
A particularly visible strand of WP5 is the Digital Twin-based control centre with XR features. During the first half of the project, requirements were gathered in close collaboration with pilot sites and end users.
By September 2025, this had progressed into interface mock-ups and early implementation work, including an initial REST API communication module that was integrated with other components and tested remotely and on premises in Nevers in September 2025. Feedback gathered from these tests is being used to refine and further develop the module and the control centre functionality.
What this means for iDriving’s pilots
WP5’s progress up to September 2025 positions the project for the next phase: expanding digital twin capabilities, strengthening AI-enabled safety and maintenance services, and deploying/validating integrated solutions across iDriving’s pilot use cases.
