Adaptive Network Coding Based on the Mutual-Information Model

In this paper, we consider cooperative communication in a two-way relay network and propose a spectrally efficient adaptive partial repeated network coding scheme. This scheme avoids error propagation of network coding. In the scheme, the relay node is assumed to have perfect channel state information, and the mutual information (MI) model is used to adaptively optimize the amount of data that need to be network encoded and forwarded by the relay node under a given block-error-rate constraint. Detailed implementations of the scheme are presented. Simulation results show that the MI model can accurately predict the optimal amount of information that needs to be forwarded by the relay node, and the proposed scheme can substantially increase the spectral efficiency and throughput while maintaining a good decoding performance.