Behavioral Modeling and Design of Predistortion for Wireless Transmitters with Multi-Branch RF Power Amplifier

In this paper, behavioral modeling approach and predistortion design methods are proposed for wireless transmitters with multi-branch radio frequency (RF) power amplifiers (PAs). Multi -branch RF PA architectures are studied such as balanced RF PA systems and linear amplification with nonlinear component (LINC) systems. Conventional behavioral modeling using single input and single output data cannot extract the behaviors of RF PAs in each branch. The behavioral modeling method using separated input signals and weighted input identification is proposed to extract accurate behavioral models for each branch. A decentralized predistortion structure that consists of multi-branch memory polynomial architecture predistortion components is proposed. The proposed behavioral modeling method is used for the design of decentralized predistortion using indirect learning method. The performances of proposed modeling and predistortion method are evaluated by comparing the normalized mean square error (NMSE) values with conventional single-branch behavioral modeling and predistortion. The input signal is 10MHz long-term evolution (LTE) signal. The results show that the proposed behavioral modeling and decentralized predistortion architecture are suitable for wireless transmitters with multi-branch RF PAs.