Study On Digital Baseband Predistortion Of Power Amplifiers

Nonlinear distortion of mobile communication systems mainly comes from power amplifiers (PA). The nonlinear distortion impacts systems severely with linear modulations and broadband signals applying. As a result, linearization of power amplifiers attaches more important. Moreover, of linearizations, digital baseband predistortion is most excellent, easiest to implenment and has greatest development potential. This paper pays attention on LUT index technique and adaptative algorithm of digital baseband predistortion. Cavers derivated the formula of predistorter (PD) and PA cascaded system. Based on the formula, Cavers converts adaptation issiue of PD to a root finding problem of a complex equation and proposed secant based adaptative method. Due to complex signals, secant has enormous computational load. This paper presents an improved discrete Newton's method based adaptative method. It simplifies and converts a complex equation to a system of two element real equations on mathematics as to avoid complex multiplications and reduce computational load. Further more, MATLAB simulation results prove the proposed method not only revises nonlinear distortion as well as secant but also has less convergency time and less runtime.On the other hand, memory effect of PA in virtue of broadband signals such as OFDM is not neglected. Simulation results prove the characteristics affected by memory effect of PA not be a clear point-to-point curve but a bundle of stochastic discrete points. So tranditional memoryless predistortion can not compensate memory nonlinear distortion of PA. And simulation results also prove secant and discrete Newton's method based adaptative methods application to memoryless PD have a certain role in modifying PD's parameters, but do not meet the requirement. Therefore, memory predistortion technique need further research.