An algorithm for adaptive predistortion of certain time-varying nonlinear high-power amplifiers

An extended Kalman filter (EKF) is proposed for adaptive precompensation of time-varying nonlinear distortion caused by high-power amplifiers (HPA) such as traveling wave tube (TWT) amplifiers frequently used in for instance, radar transmitters. In many applications, a means for adaptive predistortion is essential in order to reduce out-of band transmission. The proposed scheme is based on a nonlinear Wiener model for identification of the HPA characteristics and a predistorter with a Hammerstein-type structure. The usefulness of the proposed scheme is illustrated by means of a numerical example. In particular, it is shown that suppression of nonlinear distortion, introduced by a typical Ku-band TWT amplifier, in the order of 25-30 dB can be achieved. Furthermore, the performance of the EKF is compared to that of the recursive prediction error method (RPEM) and the stochastic gradient algorithm (LMS).