A novel DFE using a priori information in turbo equalization

Original turbo equalization, using a trellis-based channel equalizer and channel decoder, improves the bit error rate performance significantly. However, given that large alphabet modulation is employed in the system with multipath channels causing significant inter-symbol interference (ISI), the optimal maximum a posteriori probability (MAP) equalizer is prohibitively complex. Therefore, sub-optimum equalizers such as decision feedback equalizer (DFE) have to be considered. The gain in BER offered by an iterative receiver when using DFE is, however, limited by error propagation. In this paper, we propose a novel low complexity DFE algorithm to replace the BER-optimal MAP equalizer in turbo equalization without loss in BER performance. This is accomplished by extracting and delivering the extrinsic information as a priori information for the detection and decoding step. According to both analytical and simulation results, we show that the performance of the proposed modified DFE algorithm significantly improves BER performance over conventional DFE based turbo equalization.