A compensating method based on SOM for nonlinear distortion in 64-QAM-OFDM system <Abstract> In this paper, we propose a novel compensator that uses a neural network (SOM algorithm) to correct the non-linearity introduced by the HPA. In the proposed scheme, a twodimensional SOM topology structure is designed to process the real and imaginary parts concurrently and the Euclidean distance between the signal points and the winning neuron in the ideal 64 QAM signal constellation is used as a stationary radius to define the neighborhood of the winning neuron. Because the SOM is inherently nonlinear, it may thus be viewed as a nonlinear generalization of principal components analysis. Simulation results show the effectiveness of the proposed scheme in 64 QAM-OFDM systems under a nonfading AWGN channel and traveling-wave tube amplifier (TWTA) model. |