Neural network and memory polynomial methodologies for PA modeling

This paper attempts to present the performance of an artificial neural network (ANN) model and a memory polynomial (IMP) model for power amplifier (PA) modeling, which exhibits memory effects. The ANN model was based on time delay neural network (TDNN) and the memory polynomial model was developed using analytical polynomial function. Both models were developed to fit the dynamic AM-AM and AM-PM conversions of the PA obtained from QPSK digital modulated signal. The comparison results show that both models are applicable to model the PA, however, the TDNN model, compared to the memory polynomial model, can give better modeling results.

[1]  Arif Ahmed,et al.  Efficient PA modeling using neural network and measurement setup for memory effect characterization in the power device , 2005, IMS 2005.

[2]  F.M. Ghannouchi,et al.  Dynamic behavioral modeling of 3G power amplifiers using real-valued time-delay neural networks , 2004, IEEE Transactions on Microwave Theory and Techniques.

[3]  G. Kompa,et al.  Power amplifier modeling using memory polynomial with non-uniform delay taps , 2004, 34th European Microwave Conference, 2004..

[4]  M. Ibnkahla,et al.  Vector neural networks for digital satellite communications , 1995, Proceedings IEEE International Conference on Communications ICC '95.

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

[6]  J.S. Kenney,et al.  Behavioral modeling of RF power amplifiers considering IMD and spectral regrowth asymmetries , 2003, IEEE MTT-S International Microwave Symposium Digest, 2003.

[7]  Raviv Raich,et al.  Orthogonal polynomials for power amplifier modeling and predistorter design , 2004, IEEE Transactions on Vehicular Technology.

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

[9]  G. Kompa,et al.  Power amplifier linearization using memory polynomial predistorter with non-uniform delay taps , 2004, 2004 IEEE MTT-S International Microwave Symposium Digest (IEEE Cat. No.04CH37535).