A Hybrid Approach To Coded Compressed Sensing Where Coupling Takes Place Via The Outer Code

This article seeks to advance coded compressed sensing (CCS) as a practical scheme for unsourced random access. The CCS algorithm features a concatenated structure where an inner code is tasked with support recovery and an outer code conducts message disambiguation. Recently, the CCS scheme was improved through the use of approximate message passing (AMP) with a dynamic denoiser that shares soft information between the inner and outer decoders. This significantly improves performance at the cost of additional complexity. This work shows how the spatial coupling generated by the outer code is sufficiently strong to justify relaxing certain constraints on the inner code. It is shown that a block diagonal sensing matrix with the aforementioned dynamic denoiser forms an effective means to get good performance at reduced complexity. This novel architecture can be used to scale CCS to dimensions that were previously impractical. Findings are supported by numerical simulations.