Generalized dynamic spectrum access: An order statistics design perspective

Abstract

A new adaptive multiple access scheme based on the theory of order statistics is introduced. The proposed algorithm employs a semi-random allocation mechanism that limits subchannel selection to the highest-gain subchannels. Because the proposed low-complexity method requires relatively small channel information overhead and processing delays, it can be feasible for higher data rates where the bit duration approaches the channel coherence time. Since by its nature this method is non-iterative, it is suitable for networks with large number of users. Numerical results reveal significant system performance improvement over conventional approaches.