TwinShip Innovation 4
🚢 TwinShip Innovation Spotlight: DT Model Evaluation Methodology & Standards We're excited to highlight Innovation 4 of the TwinShip project – a robust methodology for evaluating Digital Twin (DT) models in the maritime sector. 🔍 Current Challenge: Most evaluation methods focus on the accuracy of regression or classification models but often overlook how structural changes in the DT model impact real-world ship performance. This can reduce trust and slow adoption. 🚀 TwinShip’s Breakthrough: Building on learnings from the COGNITWIN project, TwinShip introduces a comprehensive evaluation framework that includes hardware-in-the-loop (HIL) testing, internal vs. external model performance assessment, and a focus on trustworthiness and explainability, especially for AI/ML-based DT models. The initiative also contributes to international standardization through collaboration with the ISO/IEC JTC 1/SC 41 Digital Twin committee, helping shape future guidelines for trustworthy maritime DTs. |
![]() |
Â