Achieving gaussian multiple access channel capacity with spatially coupled sparse graph multi-user modulation

Communication over a Gaussian multiple access channel (GMAC) is considered. Each user modulates his data as a superposition of redundant data streams. The interconnection of the modulated data symbols can be described by means of a sparse graph. The transmitted modulation graphs couple at the receiver, which performs joint detection based on iterative interference cancellation. We prove that the GMAC capacity can be achieved by such a system even for the case when all the users access the channel with equal power and transmit with equal rate.