Compressed Sensing Based Multi-User Detection with Modified Sphere Detection in Machine-to-Machine Communications

Abstract:
Recently, Compressed Sensing has been proposed as a promising physical layer technique for Multi-User Detection in Massive Machine Communication (MMC). MMC is characterized by low data rates, low control signaling overhead and different traffic models compared to human-oriented communication. In this context, Compressed Sensing based Multi-User Detection (CS-MUD) enables efficient direct random access as a potential solution for the massive access problem. Previous work has not considered realistic traffic characteristic. In this work, we consider a typical Poisson traffic model to formulate a MAP detection problem exploiting user sparsity.