Interference-Cooperation in Multi-User/Multi-Operator Receivers
In a multi-user scenario where users belong to different operators, any interference mitigation method needs unavoidably some degree of cooperation among service providers. In this paper we propose a cooperation strategy based on the exchange of mutual interference among operators, rather than of decoded data, to let every operator to recover an augmented degree of diversity either for channel estimation and multi-user detection. In xDSL scenario where multiple operators share the same cable binder the interference-cooperation (IC) approach outperforms data-exchange methods and preserves to certain degree the privacy of the users as signals can be tailored to prevent each operator to infer parameters (channel and data) of the users from the other operators. The IC method is based on Expectation Maximization estimation shaped to account for the degree of information that each operator can exchange with the others during the two steps of multi-user channel estimation and multi-user detection. Convergence of IC is guaranteed into few iterations and it does not depend on the structure of the interference. IC performance attains those of centralized receivers (i.e., one fusion-center that collects all the received signals from all the users/operators), with some loss when in heavily interfered multi-user channel such as in twisted-pair communications allocated beyond 50-100MHz spectrum.