Turbo equalization for EDGE system with DDF-SOVA

As one of the 3G wireless communication standards, EDGE (enhanced data rates for global evolution) adopts 8PSK modulation to provide higher data rates using the existing TDMA system infrastructure of GSM. We study a receiver scheme which improves the detection performance through iterative equalization and decoding. Due to the 8-ary 8PSK alphabet of EDGE modulation, the number of states in a trellis for a full turbo equalizer is prohibitively high. Hence, a delayed decision feedback (DDF) approach is used to reduce the number of trellis states. A non-binary soft output Viterbi algorithm (SOVA) is developed to calculate the soft value of 8PSK channel input symbols. Iterative channel estimation is also applied to improve the accuracy of channel estimation. In addition, a prefilter is designed to "shorten" the ISI (intersymbol interference) channel length. Simulation results show our turbo equalization approach can achieve about 2 dB BER (FER) performance gain compared with non-turbo detection after 5 detection iterations.