Simplified sparse code multiple access receiver by using truncated messages

Low complexity receiver of sparse code multiple access based on dynamic trellis

“…The discussions in Section 2.2 show that the complexity in the MPA algorithm mainly comes from RNs update, which can be represented by trellis representation [11]. The trellis of the kth RN contains d c columns and M rows, which indicate the number of users who have occupied the kth physical resource and the size of SCMA codebook, respectively.…”

“…It can be considered that those symbols with high confidence in each column are more worthy of being involved in iterations, while other symbol nodes with less confidence could be truncated. Regarding this, an EML-MPA and a CAD-EML-MPA are proposed in [11]. The basic idea of EML-MPA is to select only the most reliable m c (m c < M) symbols in each column of the trellis graph to be involved in subsequent iterations.…”

“…Therefore, low complexity SCMA detection algorithms have become a research hot-spot in recent years [11][12][13][14][15][16]. Mu et al [12] proposed a partial marginalisation based message passing algorithm (PM-MPA).…”

“…To further reduce the detection complexity of single antenna SCMA, Lai et al [11] proposed the extended Max-Log based MPA (EML-MPA) detection method by reducing the number of symbols M in each iteration of MPA. The complexity of EML-MPA is reduced to…”

“…where m c denotes the number of symbols actually being involved in the iterations. Besides, to make the method adapt to the dynamic channel, Lai et al [11] also proposed a modified version of EML-MPA, i.e. the channeladaptive extended Max-Log based MPA (CAD-EML-MPA), which has a similar complexity of EML-MPA.…”