The project developed a comprehensive digital twin framework for semi‑trailer vehicles, integrating real‑time data acquisition, model‑based estimation, and cloud‑based data management. A central measurement computer, an online twin running on a Speedgoat real‑time platform, and an offline twin hosted in the cloud were connected through the OKIT CareLAN gateway. The communication was validated by a test drive that confirmed reliable data exchange between the nodes. A trigger manager was implemented to detect relevant driving manoeuvres from the online twin’s individual models—vertical excitation, suspension state monitoring, and dynamic vehicle model—by aggregating triggers from each sub‑model. When a trigger was activated, a recorder started logging time series from the measurement computer and model outputs, including estimated parameters. Deactivation of the trigger stopped recording and the data were transmitted via CareLAN to the offline twin. The resulting packet was assembled into a single .inp file that served as input for subsequent identification routines and road‑profile extraction. The trigger manager and recording process were first deployed in a laboratory environment, where early‑stage bugs were identified and corrected, before being used in on‑road tests that demonstrated a robust and fault‑tolerant operation of the full data‑capture chain.

For visualisation, an agile Grafana dashboard was created that displayed live sensor streams, vehicle position on a map, and stored data for post‑flight analysis. Result artefacts from the offline twin’s computations were made available immediately on the OKIT host “Geminae” and streamed to the cloud, where they could be accessed through a web browser for further inspection. The OKIT‑Cloud platform also served as a secure repository for measurement files, diagrams, and collaborative office documents, enabling efficient data sharing among consortium members.

The road‑damage detection component employed a YOLO‑based neural network trained to distinguish asphalt cracks from surface contamination. Tests on both test tracks and real‑world routes showed a high detection rate and reliable adaptation to unseen scenarios. The system could differentiate between cracks and debris such as tire wear, as illustrated in the test‑track results. In parallel, a real‑time 3D scene‑reconstruction pipeline was integrated into the sensor network to estimate the position and orientation of the trailer. A pose‑quality framework measured the relative pose error against a high‑accuracy ADMA sensor, confirming the accuracy of the estimated pose for use in autonomous driving safety assessments.

The digital online twin was realised as a Rapid Control Prototyping (RCP) unit built in MATLAB/Simulink. A modular template system allowed individual sub‑models to be inserted, with all input, output, and parameter signals automatically linked. This workflow enabled independent verification of each sub‑model and facilitated offline testing with synthetic data. The final integrated Simulink model represented the complete online twin, providing realistic real‑time simulation of the vehicle dynamics and control logic.

Collaboration was organised around a consortium of industry and research partners. OKIT and imes jointly defined the interface for episode data transfer and parameter exchange, with OKIT’s cloud acting as the central hub. The BPW (Betriebs‑ und Produktionswerk) contributed the test vehicle and measurement hardware, while the imes team supplied the data‑recording and trigger‑management logic. The project was funded by the German Federal Ministry of Economics and Technology (BMWK) under contract 3831/03.07_5, and spanned a multi‑year period during which regular consortium meetings, technical workshops, and public outreach events were held. The consortium also disseminated results through industry magazines, the BPW online portal, and presentations at the 7th International Commercial Vehicle Technology Symposium, thereby ensuring that the developed digital twin technology reached a broad audience of researchers and practitioners.