Tensor-based methods for high-power amplifier identification and predistortion linearization

Abstract

In this paper, we consider the problem of identification and predistortion of nonlinear high-power amplifier (HPA) using tensor-based methods. The HPA is modeled by a Wiener system structured as a linear time invariant system followed by memoryless nonlinearity. From a third-order Volterra kernel, we show that the linear subsystem of Wiener system can be estimated by means of the singular value decomposition algorithm. Then, the nonlinear subsystem is estimated by means least square algorithm. The identified Wiener PA model will be used to estimate a Hammerstein based predistorter using an adaptive algorithm in order to linearize the HPA. The proposed identification and predistortion methods are illustrated by means of simulation results.