GCMAC based predistortion architectures for personal mobile systems

Predistortion in personal mobile telephony systems has been proposed to deal with saturation in amplification at the transmission stage. This paper discusses the elements and architectures in signal and data predistortion for GSM and UMTS. The GCMAC neural network has been introduced as a predistorter to provide the design with adaptive, digital and practical features. Modulation aspects of the technologies involved define the models to substitute the amplifiers in the proposed schemes. The results included show how the predistortion architectures developed allow working linearly in saturation regimen. This point is important since it provides base transceiver stations (BTS), repeaters or even terminals with a higher gain. A performance evaluation including convergence and spectrum characteristics is carried out.