Distributed Optimization of Cognitive Radios Employed in Dynamic Spectrum Access Networks

We propose a novel distributed cross-layer optimization approach for cognitive radio wireless networks. A cognitive radio is a wireless device that is capable of sensing the prevailing environmental conditions and automatically adapting its operating parameters in order to enhance the performance of both the system and potentially the whole network. Given the lack of a centralized intelligence to control the operating parameters of all the wireless devices in a distributed network, the proposed approach employs cross-layer optimization of individual wireless devices using partial operating parameter and environmental information from adjacent devices. Given that all wireless links employ non-contiguous orthogonal frequency division multiplexing (NC-OFDM) and assuming stationary nodes, the proposed distributed optimization approach will attempt to minimize bit error rate (BER), minimize out-of-band (OOB) interference, maximize overall throughput, and minimize power consumption using a multi-objective fitness function derived in this work. The optimization implementation is based on several variants of genetic algorithms.