Improved Variable Step-size LMS for Digital Predistortion in Wideband Power Amplifiers

Abstract

Digital predistortion (DPD) is one of the most promising linearisation techniques in the field of power amplifier (PA) linearization. The adaptive algorithm plays an important part in the DPD system. This paper proposes an improved variable step-size LMS algorithm based on error function and sine function, which can provide both more stable convergence and lower steady-state error. Simulation results show that the proposed adaptive algorithm can effectively suppress the spectral regrowth for the wideband PA with nonlinearity and memory effect.