Neural Networks for RF and Microwave Design

This chapter presents examples of ANN-model based analysis and optimization of RF and microwave designs. Examples of component optimization, CPW circuits (a power divider and a folded double-stub filter), optimization of a CPW patch antenna, optimization of multilayer band pass filters, and yield optimization of a three-stage MMIC MESFET amplifier are presented.
When ANN models for various RF and microwave components have been developed, they can be used in the design and optimization of circuits, antennas, integrated circuit-antenna modules, and other subsystems and systems. The role of models in the design process was discussed in Chapter 2, "Modeling and Optimization for Design."
For RF and microwave circuit design, ANN models are linked to the circuit simulator (such as HP-EEsof's MDS, ADS). Integration of ANN models with circuit simulators was discussed in Section 5.1. Once linked with the simulator, these models are used just as any other model available within the network simulator. The models can be used with other ANN models or combined with many other models available in the simulator to form a complete circuit.
Circuit and/or component optimization, where the gradient of a function is needed, can also be performed using the routines available within the circuit simulator. Two-layer networks with sigmoidal hidden units have the ability to simultaneously approximate both a function and its derivative [1]. This isdue to having continuous hidden layer and output layer neuron activation functions (sigmoids), which are differentiable everywhere.
Several examples of the use of ANN...