Small-signal and noise modeling of HBTs based on neural network approach

Heterojunction bipolar transistor (HBT) technology is very attractive for microwave wireless communications. A small-signal and noise model of HBTs is presented in this paper. Modeling procedure is based on the artificial neural network (ANN) approach, which enables high accuracy together with the efficiency and simplicity commonly requested from CAD techniques. The prediction of device S- and noise parameters over the whole frequency range and over the broad ranges of operating conditions is possible by the developed ANN model.