Spline-Based Model for Digital Predistortion of Wide-Band Signals for High Power Amplifier Linearization

Different linearization techniques have been proposed to linearize Power Amplifiers (PAs) and improve the PA efficiency for non-constant envelope modulation schemes. Among all the linearization techniques, Digital Predistortion (DPD) is one of the most cost effective. For wider bandwidth applications, PA memory effects can no longer be ignored. Therefore in order to achieve a higher linearization performance, the Predistorter (PD) requires to compensate for memory nonlinearities. In this paper a new PD model based on piecewise polynomial approximation in the form of splines is described and compared with memory polynomial PD. The purpose of the spline PD is to increase the approximation accuracy and/or reduce the complexity. Simulation results with 16 Quadrature Amplitude Modulation (QAM) show a significant improvement in Adjacent Channel Power Ratio (ACPR) for a Wiener-Hammerstein power amplifier model. A set-up to perform the measurements on a 100 watt High Power Amplifier (HPA) is also developed. The experiments are carried out on both an uplink WCDMA and a crest-reduced two-carrier WCDMA signal. The results show that a higher linearization performance is obtained by using spline basis functions as compared to conventional memory polynomial PD.