NEURAL PREСOMPENSATORS AGAINST NONLINEAR DISTORTIONS IN HIGHFREQUENCY POWER AMPLIFIER
The cascade structure of nonlinear digital precompensator synthesized by the direct learning algorithm is proposed. Functional link and polynomial perceptron neural networks are represented. It is shown that for the power amplifier described by Wiener-Hammerstein model the most accuracy of nonlinear distortion cancelling is provided with the cascade precompensator including the polynomial perceptron network and redially pruned Volterra model.
Authors: E. B. Solovyeva
Direction: Electrical Engineering
Keywords: Nonlinear compensation, nonlinear model, power amplifier, nonlinear distortion
View full article