Bit-Wise Partial Noise Elimination in Cooperative Decode-Amplify-Forward Relay Node

This paper deals with an adaptive bit-wise switching strategy for hybrid decode-forward (DF) and decode-amplify-forward (DAF) relaying and proposes a novel bit-wise partial noise elimination technique. In relay nodes of DAF relaying, bit log likelihood ratios (LLR)s with high reliability is proactively hard-decoded to cut noise propagation. With this method, we have a difficulty to calculate LLRs in the destination node, due to multi-modality of probability density function (PDF) caused by nonlinear operations of the noise elimination. In order to resolve this issue, a classification problem is formulated in this paper. We will demonstrate that DAF with the bit-wise noise elimination functionality is capable of improving frame error rate (FER) performance in the destination node.