Decentralized dynamic channel allocation in correlated Nakagami fading channels: an order statistics analysis

In many future wireless applications and ad-hoc networks a central processor is not available. It is desirable to adopt distributed methods with minimum processing requirements that can efficiently and fairly allocate resources among multiple users. This paper considers a new fair decentralized multiple access channel allocation scheme. The proposed algorithm employs a semi-random selection mechanism that limits subchannel selection to the highest-gain subchannels. Hence, the analytical foundation for the proposed method is the theory of order statistics. Since by its nature this algorithm is non-iterative, it requires relatively small complexity, channel information overhead, and processing delays. We derive results for the bit error rate (BER) over a set of correlated and not necessarily exchangeable Nak-agami fading subchannels. Numerical results reveal significant system performance improvement over a conventional random allocation approach. The performance of this new algorithm can be close to the highly complex optimal search method.