Fast RCS Evaluation of Cone-Cylinder-Flare Shape Using a Machine Learning Model

A shape of medium geometric complexity initially designed for computational fluid dynamics engineering is eval-uated for RCS. The RCS is expressed as a function of both geometric and EM wave parameters. A boosted decision-tree (AdaBoost) model is then trained for regression of the RCS function using a limited number of exact RCS computations as training data. The context of this study is to integrate such models into more complex multidisciplinary workflows.