Predistortion of quadrature amplitude modulation signals using Volterra series approximation

Abstract

Modern digital communication systems are being called upon to move ever increasing amounts of information over decreasingly available bandwidth. This requires that communication systems employ bandwidth-efficient modulation schemes to conserve bandwidth while moving the information at higher data rates. A major stumbling block to using higher order modulation schemes in long haul communication is the distortion caused by high power amplifiers. These high power amplifiers are required to amplify the signal power to a level that will allow distant receivers to correctly demodulate and decode the information. The distortion caused by the high power amplifiers can render a modulation scheme unusable due to the high symbol error rates which result from the extensive skewing of the modulation scheme's signal constellation. This thesis details a predistortion technique using Volterra series approximation techniques to model the inverse of the high power amplifier's distortion characteristics. A 64 Quadrature Amplitude Modulation (64-QAM) system incorporating a predistorter is used to demonstrate the ability to achieve acceptable bit error rates. The implementation of the inverse model and the communication system is performed in MATLAB. The results show the viability of predistortion of digital data to allow the higher order modulation schemes to be incorporated into communication schemes, increasing the overall data rate while conserving bandwidth.