Neural Network Field Programmable Gate Array (FPGA) Controllers for Reconfigurable Antennas

Advantages of reconfigurable antennas are numerous, but limited by the method of controlling their configuration. This dissertation proposes to utilize the advantages of both Neural Networks (NN) and Field Programmable Gate Arrays (FPGAs) to overcome this dilemma.;In this work, the methodology of modeling of reconfigurable antennas using neural network embedded on an FPGA board is presented. This work shows a new approach of modeling reconfigurable antennas using neural networks in Matlab, a code is written to generate a NN for any antenna (or any reconfigurable system in general) by providing input/output data of the antenna.;An HDL code is generated using Xilinx System Generator and sent to an FPGA board using Xilinx ISE. With a NN embedded on the FPGA board, we have a highly reconfigurable system in real time controller that thinks exactly as the system it models. This “brain” is connected to the antenna and becomes the decision maker in antenna switching or reconfiguration. Also, with the new approach of using Matlab to generate HDL code; this work opens the door to those who are interested in implementing designs on FPGAs without having enough knowledge in HDL programming.;Different types of reconfigurable antennas with different way of reconfigurability are modeled and simulated. NN models show great match with measured antennas data. NN_FPGA controller is built for each antenna.