A modified decomposed vector rotation behavioral model for concurrent dual-band envelope tracking RF power amplifiers

This paper proposes a novel behavioral modeling technique for concurrent dual-band (DB) envelope tracking (ET) power amplifiers (PAs) by borrowing the concept of generalized memory polynomial (GMP) model into a decomposed vector rotation (DVR) structure. By employing the proposed method, fewer coefficients, low computation complexity can be achieved with comparable performance of Volterra-based models. An experiment is carried out by employing a concurrent dual-band signal with each of 5 MHz OFDM signals and the center frequency of 2.0 GHz and 2.6 GHz, respectively. The modeling performance was evaluated with an accuracy of around 1% for normalized root mean square error (NRMSE) in both bands.