Simple and Effective Extrapolation Technique for Neural-Based Microwave Modeling

This letter presents a simple and effective technique allowing neural network based microwave models to be used outside their training range. Our technique detects necessity of extrapolation inside each hidden neuron and performs extrapolation using a modified neuron activation function. Compared to existing techniques, the proposed technique is easy to implement and the extrapolation result is equally good or better. Examples of extrapolating neural models of a high electron mobility transistor and a power amplifier are presented.

[1]  Qi-Jun Zhang,et al.  Neural Networks for RF and Microwave Design , 2000 .

[2]  Claude Brezinski,et al.  Extrapolation methods - theory and practice , 1993, Studies in computational mathematics.

[3]  Chu Kiong Loo,et al.  Extrapolation detection and novelty-based node insertion for sequential growing multi-experts network , 2003, Appl. Soft Comput..

[4]  M.C.E. Yagoub,et al.  Robust neural based microwave modelling and design using advanced model extrapolation , 2004, 2004 IEEE MTT-S International Microwave Symposium Digest (IEEE Cat. No.04CH37535).

[5]  Jianjun Xu,et al.  Neural based dynamic modeling of nonlinear microwave circuits , 2002, IMS 2002.

[6]  Qi-Jun Zhang,et al.  Artificial neural networks for RF and microwave design - from theory to practice , 2003 .

[7]  Diego Masotti,et al.  A new family of neural network-based bidirectional and dispersive behavioral models for nonlinear RF/microwave subsystems(invited paper) , 2002 .

[8]  J.E. Rayas-Sanchez,et al.  EM-based optimization of microwave circuits using artificial neural networks: the state-of-the-art , 2003, IEEE Transactions on Microwave Theory and Techniques.

[9]  Mauro Mongiardo,et al.  A review of artificial neural networks applications in microwave computer-aided design , 1999 .

[10]  T. Jan,et al.  Robust short term prediction using combination of linear regression and modified probabilistic neural network model , 2003, Proceedings of the International Joint Conference on Neural Networks, 2003..