Policy-based multiple access for decentralized low power systems

In this paper, we present an information theoretic study of a decentralized multiple access channel (MAC), where users are allowed to autonomously change their data rates, transmission powers, and channel codes, independently from other users in the system. We introduce a policy-based access mechanism, where each user is allocated a set of code-books, with various data rates and powers. Successful reception of all users is guaranteed, provided that each of the users adheres to its assigned access policy, when selecting its date rate and power. We completely characterize the capacity region for such policy-based access in the low power regime, for both single and multi-antenna (MIMO) systems, and we show that nonorthogonal policies outperform orthogonal ones. We also show that known MIMO capacity scaling laws, which apply in single user systems, carry over into framework of decentralized MAC.