Application of Improved AHTVSSLMS Algorithm in Digital Predistortion System

In modern wireless communication systems, the nonlinear region of power amplifiers severely affects communication quality. Digital predistorters have become vitally important at the front end of power amplifiers to counteract this nonlinear distortion and improve communication quality. We introduce a variable step-size LMS algorithm founded on the absolute inverse hyperbolic tangent to identify system parameters for the indirect learning structure of a predistortion model. We also compared the performance in terms of convergence time and output signal quality among the least mean square (LMS) algorithm, variable step-size least mean square (VSSLMS), and other algorithms. In the experiments, we utilized a predistortion model with a 64-QAM signal and memory polynomial. Experimental results showed that, compared to the conventional LMS and VSSLMS, the improved AHTVSSLMS algorithm significantly enhances the convergence, tracking performance, ability to suppress spectral spreading, and linearization effect in predistortion identification.