Sparse Identification of Memory Effects and Nonlinear Dynamics for Developing Parsimonious Behavioral Models of RF Power Amplifiers

This article, deals with the sparse identification of memory effects and nonlinear dynamics for accurate and efficient behavioral modeling of RF Power Amplifiers (PAs). Here, we use sparse regression using a sequential thresholded least-squares algorithm to determine the fewest relevant terms from a large set of available terms required to accurately represent the dynamics of RF PAs. The proposed approach develops a framework for behavioral modeling of RF PAs, taking into advantage, the advances in sparsity techniques which balances the model accuracy with complexity. We show that, for similar modeling performance, the proposed method requires fewer coefficients than the standard memory polynomial model and simplified Volterra based models.