Skip to content

Factor-Graph Aided Compressed Sensing for Grant-Free SCMA

2026 · IEEE Wireless Communications Letters · Vol 15, pp. 5547-5551 · 0 citations · 20 references

Abstract

Grant-free access enables IoT devices to transmit without prior scheduling in next-generation networks. Devices send short pilot sequences to keep overhead low, and the base station is required to jointly recover user activity and channel coefficients from the superimposed pilot observations. We propose a factor-graph compressed sensing framework for joint user activity detection and channel estimation in grant-free sparse code multiple access (SCMA) systems. By mapping the SCMA bipartite factor graph onto per-resource compressed sensing sub-problems, the required pilot overhead is governed by the per-resource active-user load rather than the global number of active users. Each sub-problem is solved by a factor-graph-aided Bernoulli–Gaussian approximate message passing (FG-BG-AMP) algorithm. Expectation-maximization (EM) learning jointly estimates the noise variance and the activity probability online. We further derive a practical pilot length design criterion based on the compressed sensing recovery condition. Simulations show that FG-BG-AMP approaches within 0.7 dB of the oracle linear minimum mean-square error (LMMSE) benchmark at $P=8$ . It retains a 6.9 dB advantage over the strongest advanced structured baseline at high SNR, and stays robust to deterministic pilots and correlated traffic.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.