Robust End-to-End QoS Maintenance in Non-Contiguous OFDM Based Cognitive Radios

A recent development in wireless communication is Cognitive Radio (CR) technology, an innovative radio design approach which allows the realization of intelligent allocation of the scarce radio resources such as spectrum. In this paper, we attempt to exploit past channel information and the flexibility of non-contiguous orthogonal frequency division multiplexing (NC- OFDM) based CRs to maintain end-to-end QoS performance under dynamic spectrum sharing environments. So far, most research works in resource allocation in CRs have mainly concentrated on the spectrum opportunity discovery aspect while the robust QoS performance problem has remained largely unexplored. In this work, we use the concept of portfolio optimization to achieve QoS maintenance in NC-OFDM CR systems. The problem of allocating power to maintain throughput is cast into a channel gain variance minimization and mean- variance maximization frameworks to achieve a given throughput performance under fixed BER and power limitation constraints. Numerical results are presented to demonstrate the QoS maintenance performance in various wireless channel settings.