Modeling and optimization of microwave circuits and devices using wavelet neural networks

In this paper, a new neural network modeling approach based on the wavelet transform is developed, which can be employed as a useful tool for learning a mapping between an input and an output of nonlinear microwave circuit. As to an optimization problem in microwave circuits and devices, the wavelet neural network can establish an exact model with it through a self-adaptive procedure by learning the input and output data of microwave circuits. The activation function of the proposed network is the compact supported nonorthogonal function which is the inverse Mexican-hat function. A gradient-based training algorithm is developed. The structure of the wavelet networks are feedforward networks with one hidden layer. We have applied it to the modeling of GaAs MESFET and optimization design of a band-pass filter for DSRC system and obtained a satisfactory result.