The Early Years
Digital Predistortion has been around to some extent for the last 20 years. In this chapter, we will walk through all of the key papers and their contributions.
Creation of the Indirect Learning Architecture
In 1988, A multilayer neural network controller introduced the indirect learning architecture. However, it would not be used for DPD for another 9 years. This paper is a good reminder of the previous wave of attention that neural networks recieved.
Introduction of the Indirect Learning Architecture to DPD
Changsoo Eun, and E.J. Powers introduced the indirect learning architecture to DPD in their 1997 paper in IEEE Transactions on Signal Processing, “A New Volterra Predistorter Based onthe Indirect Learning Architecture.”
The Generalized Memory Polynomial
The generalized memory polynomial (GMP) is likely the most widely used model in the world. Morgan et. al. introduced the extension to the memory polynomial in their IEEE Transactions on Signal Processing paper, “A Generalized Memory Polynomial Model for Digital Predistortion of RF Power Amplifiers”