Comparison between neural network and kriging meta-model techniques on a crosstalk problem based on Kron’s model

In electronic device design, it is crucial to establish rules to address electromagnetic compatibility (EMC) considerations upfront. Systems under consideration can be modeled to simulate physical phenomena prior to their integration into an electronic device. This paper explores and evaluates two meta-modeling approaches: neural networks and kriging. We examine a crosstalk issue utilizing Kron’s method, known for its ability to deliver precise results, making it suitable for practical applications.