Superposed 16-QAM Signal Detection Using GaBP in a Massive MIMO System

Gaussian belief propagation (GaBP) has been proposed to detect a large number of signals. It is based on message passing of the second-order complexity. In general, the GaBP uses bit-wise reliability. Thus, it is not straightforward to apply the GaBP to the multi-level-modulated signals. In this paper, we propose to compose a 16-QAM symbol from two QPSK symbols using superposition modulation and evaluate the detection performance of the GaBP for superposed 16-QAM signals in a massive MIMO system. In the uncoded case, the uniform and Gray-mapped 16-QAM provides the best performance. However, in the coded case, the performance of superposed 16-QAM becomes much better and outperforms the uniform and Gray-mapped 16-QAM. Thus, the potential capability of superposition modulation in the GaBP detection has been confirmed.