Skip to Main Content
Directional couplers with planar geometry were analyzed and designed successfully and efficiently by means of neural networks. The considered material was the alloy of Aluminum, Gallium and Arsenic represented by AlxGa1-xAs. In this work, the needed data for training the neural networks were obtained through analytical solutions. Satisfactory results were obtained by the neural network, finally, a tool for the design and analysis of such couplers has been implemented in C/C++.
Date of Conference: Oct. 29 2011-Nov. 1 2011