Reverse modeling of microwave circuits with bidirectional neural network models
Document Type
Article
Publication Date
12-1-1998
Abstract
Neural networks have been developed into an alternative modeling approach for the microwave circuit-design process. In this paper, we describe a neural network-based microwave circuit-design approach that implements the solution-searching optimization routine by a modified neural network learning process. Both the development of a microwave circuit model and the searching of a design solution can thus take advantage of a hardware neural network processor, which is significantly faster than a software simulation. In addition, a systematic simulation-based approach to convert conventional circuit models into neural network models for this design process will be described. The development of a heterojunction bipolar transistor (HBT) amplifier model and its applications are demonstrated. © 1998 IEEE.
Publication Source (Journal or Book title)
IEEE Transactions on Microwave Theory and Techniques
First Page
1492
Last Page
1494
Recommended Citation
Vai, M., Wu, S., Li, B., & Prasad, S. (1998). Reverse modeling of microwave circuits with bidirectional neural network models. IEEE Transactions on Microwave Theory and Techniques, 46 (10 PART 1), 1492-1494. https://doi.org/10.1109/22.721152
