Explainable Artificial Intelligence for Antenna Sensor Modeling

The extensive application of Artificial Intelligence and Machine Learning is notable in additive manufacturing realms such as manufacturing supervision and error identification. Nonetheless, the escalating employment of AI in vital processes during production necessitates its decision-making and prediction abilities to display interpretability. Consequently, the need for explainable Artificial Intelligence (xAI) in such manufacturing procedures becomes paramount.In this paper, we develop an explainable AI framework for antenna sensor modeling during additive manufacturing. An explainable machine learning method is used in this framework for modeling based on the frequency response data from the antenna sensor. In this approach, the explainable dimension reduction method can interpret the chosen components using projection matrix. The decision tree regressor exhibits proficiency in efficiently handling small datasets and retaining commendable explainability.Experiments demonstrate that the S-parameter prediction generated by our method are more accurate and reliable. This method can be generalized to a new dataset.Our work could provide a useful tool for antenna sensor modeling and signal processing fields during additive manufacturing.