Volterra NN-based behavioral model for new wireless communications devices

GEORGINA STEGMAYER,; CHIOTTI, OMAR

Neural Computing & Applications

Springer

Año: 2008 vol. 18 p. 283 - 291

0941-0643

This paper presents a neural network-based behavioral model for accurately reproducing the nonlinear and dynamic behavior of new wireless communications devices. Moreover, an efficient procedure to extract a behavioral Volterra model from the parameters of the NN-based model is explained, thus providing a simple way to construct very compact and accurate models, which may provide open information about device performance. Experimental tests try to demonstrate the validity of the proposed approach, for characterization of RF transistors and power amplifiers, showing strong nonlinearities and memory effects in the analyzed bandwidth.