A Low Complexity Detection Algorithm for SCMA

TL;DR

This paper introduces a low-complexity detection algorithm for SCMA that significantly reduces computational load while maintaining high accuracy, enabling more efficient massive connectivity in 5G networks.

Contribution

A novel detection algorithm for SCMA that discretizes PDFs to lower complexity from exponential to logarithmic scale, improving efficiency without sacrificing accuracy.

Findings

Detection complexity reduced from exponential to logarithmic scale.

Detection accuracy approaches that of existing algorithms with finer discretization.

Supports massive connectivity in 5G with lower computational resources.

Abstract

Sparse code multiple access (SCMA) is a new multiple access technique which supports massive connectivity. Compared with the current Long Term Evolution (LTE) system, it enables the overloading of active users on limited orthogonal resources and thus meets the requirement of the fifth generation (5G) wireless networks. However, the computation complexity of existing detection algorithms increases exponentially with $d_f$ (the degree of the resource nodes). Although the codebooks are designed to have low density, the detection still takes considerable time. The parameter $d_f$ must be designed to be very small, which largely limits the choice of codebooks. In this paper, a new detection algorithm is proposed by discretizing the probability distribution functions (PDFs) in the layer nodes (variable nodes). Given $M$ as the size of one codebook, the detection complexity of each resource node (function node) is reduced from $O(d_f M^{d_f})$ to $O(d_f^3 \ln (d_f))$. Its detection accuracy can quickly approach that of the previous detection algorithms with the decrease of sampling interval in discretization.