Iterative source-coded equalization: turbo error concealment for ISI channels

In this paper, we analyze the behavior of soft decision source decoding for ISI channels and present an iterative scheme similar to Turbo equalization allowing to combat efficiently intersymbol interference effects. In a first step, we investigate the behavior of different index assignments in the quantizer and show that gains are possible by iterative processing without any channel (de)coding. In a second step, we use short block codes resulting in redundant index assignments at the transmitter. The receiver does not perform any explicit channel decoding but exploits the additional redundancy of the index assignment in the soft decision source decoder. Furthermore, we show that a very simple channel precoder permits even larger gains without any additional decoding complexity. Moreover, the extension towards a flexible multi-mode system allowing a trade-off between channel quality and quantization quality over a wide range of channel conditions is presented.