Node selection and LLR damping for channel equalization with belief propagation

In recent years, several progresses have been made in belief propagation (BP) based signal detection with large scale factor graphs. When we apply the BP algorithm to equalization in a SIMO multipath channel, the factor graph has many short loops where the edge strengths have a fixed pattern. Thus, the proper convergence is not always expected. In general, the LLR oscillates in ill-converged cases. Therefore, preventing the LLR oscillation becomes important for the BP-based equalization. In the paper, we propose applications of node selection (NS) and LLR damping to prevent the LLR from oscillating. The NS extends the loop length virtually by a serial LLR update whereas the LLR damping suppresses the oscillation directly. The simulation results show that the error floor is significantly reduced with the LLR damping and NS in the uncoded case and that the NS works very well in the coded case.