Identification of the Volterra kernels of nonlinear power amplifiers

Abstract
A method for identifying the Volterra model of a nonlinear power amplifier with memory is given. It first requires the estimation of the Volterra structure. The coefficients of the discrete-time series can then be calculated using a least-squares criterion, based on the known input and measured output samples. In the measurement part, an identified Volterra model is compared to the simpler quasi-memoryless model, described by a Taylor series. Second, two power amplifiers are compared by simulations and verified by measurements.

This publication has 10 references indexed in Scilit: