Improvements on long term memory modeling in power amplifiers.

Abstract

This paper addresses the behavioral modeling for radio frequency amplifiers in communication systems design. The RF front-end in new communication systems is faced with variable envelope and large bandwidth signals (including desired signals and large interferers), which require the behavioral model of the amplifying unit (either in TX or RX) to accurately account for nonlinear memory effects. A tremendous amount of work has been devoted to the subject; nevertheless critical issues remain to accomplish a model that performs equally well with all possible input stimuli. This communication focuses on the particular point of long term memory identification. A new model equation is proposed that improves the previous modeling approaches.