This is the source code and manuscript for my masters thesis on adaptive digital predistortion (DPD).
This work was completed while at The University of New Haven.
Radio Frequency (RF) Power Amplifiers (PA) are fundamentally non-linear due to their design and environmental effects, this non-linearity reduces performance and introduces distortion to the system. To improve the linear region of operation an adaptive Digital Predistortion Process (DPD) is used consisting of an adaptive Finite Impulse Response (FIR) filter utilizing the Least Means Square (LMS) algorithm.
This process pre distorts the PA input to be the inverse of the amplifier distortion. As the pre distorted input passes though the PA and the distortion is added the output becomes increasingly linear as the distortion has been accounted for.
The two source codes in this repository are DPD_Heavy and DPD_Slim. The source code was developed in MatLAB R2022b.
DPD_Heavy contains all of the test cases and results provided in the manuscript, it is also not very organized.
DPD_Slim contains one test case, is well organized, and easier to understand.