Sparsity-Aware Ordered Successive Interference Cancellation for Massive Machine-Type Communications

Detection Techniques for Massive Machine-Type Communications: Challenges and Solutions

“…The well-known K-Best detector was modified in [14], rewriting the signal model as a sparse model suitable for CS approaches. [15] reports a detection order method based on the activity probability of devices and the sorted QR decomposition (SQRD), in order to increase the efficiency of SA-SIC. The work in [16] investigates the reliability of each soft-estimate obtained by a regularized MMSE-SIC detector while [17] proposed an algorithm based on the direction method of multipliers (ADMM).…”

“…Considering the same assumptions of the SA-SIC, the authors of [15] proposed a permutation of columns of channel matrix H based on channel gains. The idea of SA-SIC with A-SQRD is to replace the l 0 -norm with the l 2 -norm in (10), incorporate the regularization factor into the channel matrix • Lower complexity alternative to the S-MAP detectors;…”

“…SA-SIC with A-SQRD [15] • Replaces the l 0 -norm with the l 2 -norm, incorporated into the channel H.…”

“…x /σ 2 v , the linear mean squared error (LMMSE), unsorted SA-SIC [14], SA-SIC with A-SQRD [15], AA-MF-SIC [16], AA-RLS [43] and AA-RLS-DF [43] are compared. As a lower bound, the Oracle LMMSE detector, which has the knowledge of the index of nonzero entries, is considered.…”

“…As OMP has a better performance, a lot of improvements of it are available in the literature and some of them were specifically designed for mMTC. SA-SIC A-SQRD [15] AA-MF-SIC [16] AA-RLS [45] AA-RLS-DF [45] Oracle MMSE wGOMP Drawing inspiration from the block-sparse variant of OMP, the GOMP [99], the authors in [33] propose an improvement where they refine the activity detection, exploiting the channel code. Considering perfect CSI and that the BS has the knowledge of the number of active devices, structuring the receiver in an iterative feedback approach, the main idea of wGOMP is to pass weights w based on the channel decoding output to the multi-user detection.…”