A CS-Based Grant-Free Media Access Scheme for NOMA-Based Industrial IoTs with Location Awareness

Grant-free media access is vital for applications in Industrial IoTs (IIoTs), where stringent delays are required. Recently, due to the capability of supporting parallel receptions, Non-Orthogonal Multiple Access (NOMA) has gained research interests in IIoTs. Obviously, combining them organically is beneficial for enhancing the delay performances. In this paper, for a typical convergecast wireless network where its data sink is NOMA-based, we propose a grant-free MAC (Media Access Contention) scheme based on Compressive Sensing in Busy Tone Channel (CSiBTC), by exploiting the transmission sparsity in IIoTs. First, a to-be transmitter acquires the identities of active transmitters with the proposed CSiBTC scheme completely by itself. Two construction methods for CSiBTC are proposed for two distinct application scenarios respectively. Then, given the locations of all wireless sensors and the data sink in the network, the to-be transmitter can find out if it is eligible for starting its transmission without impairing the on-going transmissions. The scheme is grant-free and makes the most use of the parallel reception capability of NOMA, and therefore both the delay performance and throughput performance can be improved with respect to the general CS-based MAC. Performance evaluations also strongly support the above conclusions.