Adaptive Predistortion Linearization of Nonlinear Power Amplifiers

by

Bakri Aboukarr

M. Eng., 1990

Abstract

The radio spectrum is a scarce resource, which necessitates the use of highly efficient modulation schemes in digital transmission systems. Although, theoretically such modulation methods achieve a high spectrum efficiency, they tend to suffer from the nonlinearities of high power amplifiers and as a result they spread the spectrum of the filtered signal, causing some adjacent channel interference. This thesis addresses the problem of adaptive linearization of high power amplifiers. The technique considered here is independent of the modulation scheme and is suitable for narrow band signals. In practice, it is applied at base band. In this technique, the low pass equivalent envelope of the modulated signal is distorted prior to amplification (i.e., predistorted), such that the amplified signal is a linear version of the modulated signal. The predistortion is achieved by mapping the modulated signal to a distored signal using table lookup functions. The mapping is done using the polar representation of the low pass complex envelope of the modulated signal since the amplifier's nonlinearities are a function of the amplitude. The adaptation of the table lookup functions are based on the Least Mean Square (LMS) algorithm. The adaptation technique is analyzed and tested on a Quadrature Phase Shift Keying (QPSK) modulation system using computer simulation. The results obtained indicate that the technique provides excellent performance. In addition, the technique enjoys the advantages of implementation simplicity, low memory and processing requirements and fast convergence.

Back to the list of graduate students