Theoretical Throughput Analysis for Massive Random Access With Spatial Successive Decoding

A key challenge of enabling massive machine type communications (mMTC) is to support infrequent small data transmission from a large number of machine type devices (MTDs). Random access with spatial successive decoding allows a packet to be received by multiple receivers and decoded with successive interference cancellation (SIC) across receivers, increasing not only throughput but also reliability, which is a promising random access protocol for mMTC. However, its theoretical throughput analysis is still an open issue. To address this issue, an innovative bipartite graph based throughput analysis framework is proposed in this paper, in which multiple receivers serve as sum nodes, active users serve as burst nodes, independent on-off fading channels are used to model edges connecting sum nodes and burst nodes. With this bipartite graph representation, an and-or tree iteration with binomial distributed node perspective degree distribution is formalized to evaluate the asymptotic throughput of random access with spatial successive decoding. Moreover, as a byproduct theorem, an alternative simplified proof is given to derive the analytical throughput of random access with spatial diversity but without successive decoding. Comprehensive numerical results are provided to illustrate their critical performance characteristics and verify the advantages of including spatial diversity and successive decoding for random access.