A Channel-Blind Detection for SCMA Based on Image Processing Techniques
Chao Yang, Weihong Xu, Zaichen Zhang, Xiaohu You, Chuan Zhang · 2018
Sparse-code multiple-access (SCMA) is an effective non-orthogonal multiple-access (NOMA) technique. Existing detectors such as deterministic message passing algorithm (DMPA) are one-dimensional and require precise channel estimation. This paper proposes a blind detector from a two-dimensional perspective. Main work involves pattern construction and pre-filtering with different image techniques. In this paper, a 4 × 4 Sudoku template is applied for the pattern construction of one-dimensional SCMA signals. Total variation based on first order differential operator is adopted for global pre-filtering of DMPA. Image training is adopted in DMPA to further reduce the environment noise. The output signal of both pre-filtering methods are detect through DMPA with constant noise density N0. Numerical results show that two-dimensional blind detection can well compensate the performance when channel estimation of DMPA is not perfect. A general hardware architecture of the detecting method is also proposed in this paper.