Artificial Neural Networks for RF/Microwave Modelling — System Theory Approach

The goal of most artificial neural network applications in RF/microwave design is to obtain accurate and efficient models. In the majority of these applications, neural networks are treated as black boxes. System theory also provides different approaches to modelling problems. Reviewing basic problems in system theory may enrichen the applications of artificial networks in RF/microwave design. Space mapping technique provides an accurate and efficient approach for the solution of design problems. This technique is also used for modelling with artificial neural networks. As it is shown to be effective in various RF/microwave design problems, space mapping will be investigated using a system theory approach.