Neural-network-based predistortion method for high-power amplifiers with memory

This paper presents a novel predistorter architecture based on Generalized Radial Basis Function (GRBF) neural network for high-power amplifier (HPA) with memory in an orthogonal frequency division multiplexing (OFDM) system. The predistorter is implemented using an indirect learning architecture. An efficient algorithm to update the neural network weight matrices is derived. Simulation results show that the proposed neural network predistorter can effectively reduce the nonlinear distortion of HPA and produce a faster convergence speed than the conventional backpropagation algorithm.