De-risk assessment: Gaussian Processes for AI-based Antenna Data Analysis

The project aims to develop a meta-modelling framework integrated into TICRA Tools to address computationally expensive antenna design tasks. To this end, the activity focuses on leveraging Gaussian Process (GP) models to reduce computational costs while maintaining accuracy, offering a more efficient solution for complex design optimizations. 

Achievements and status: Prototype implemented in TICRA Tools for antenna design optimization using meta-models.Demonstrated significant computational efficiency compared to gradient-based and global search methods, offering faster convergence for computationally expensive problems.Automatic selection of the appropriate meta-model, making it accessible for engineers without expertise in machine learning.Achievements and status: Prototype implemented in TICRA Tools for antenna design optimization using meta-models.Demonstrated significant computational efficiency compared to gradient-based and global search methods, offering faster convergence for computationally expensive problems.Automatic selection of the appropriate meta-model, making it accessible for engineers without expertise in machine learning.