Gülbahar, Burhan2017-10-122017-10-122017-060090-6778http://hdl.handle.net/10679/5658https://doi.org/10.1109/TCOMM.2017.2669995Due to copyright restrictions, the access to the full text of this article is only available via subscription.Magneto-inductive (MI) wireless communications is an emerging subject with a rich set of applications, including local area networks for the Internet-of-Things, wireless body area networks, in-body and on-chip communications, and underwater and underground sensor networks as a low-cost alternative to radio frequency, acoustic or optical methods. Practical MI networks include multiple access channel (MAC) mechanisms for connecting a random number of coils without any specific topology or coil orientation assumptions covering both short and long ranges. However, there is not any information theoretical modeling of MI MAC (MIMAC) capacity of such universal networks with fully coupled frequency selective channel models and exact 3-D coupling model of circular coils instead of long range dipole approximations. In this paper, K-user MIMAC capacity is information theoretically modeled and analyzed, and two-user MIMACs are modeled with explicitly detailed channel responses, bandwidths and coupled thermal noise. K-user MIMAC capacity is achieved through Lagrangian solution with K-user water-filling optimization. Optimum orientations maximizing capacity and received power are theoretically analyzed, and numerically simulated for two-user MIMACs. Constructive gain and destructive interference mechanisms on MIMACs are introduced in comparison with the classical interference based approaches. The theoretical basis promises the utilization of MIMACs in 5G architectures.engrestrictedAccessarticle6562594260700040381930002310.1109/TCOMM.2017.2669995Magneto-inductive communicationsMultipleaccessCapacityTopology managementWater-filling2-s2.0-85025707542