UBC Theses and Dissertations
Dynamic resource allocation for OFDM-based cognitive radio systems Bansal, Gaurav
Cognitive radio (CR) is an emerging technology that would improve spectrum utilization by exploiting unused spectrum in dynamically changing environments. We investigate the design of link adaptation algorithms (e.g., adaptive power and bit loading) for orthogonal frequency division multiplexing (OFDM)-based CR systems. Different power and bit loading schemes can be designed for CR users which exploits the time varying nature of fading gains across the OFDM subcarriers. However, one of the challenges here is to ensure that the interference caused to the primary users (PUs) remains below the target interference threshold. Therefore, not only do we need to consider the fading gains, but also the spectral distance of the subcarriers from the PU's band. In this thesis, we propose an optimal power loading algorithm, assuming that the rate can be varied continuously, for an OFDM-based CR system. The downlink transmission capacity of the CR user is thereby maximized, while the interference introduced to the PU remains within a tolerable range. We investigate the case of discrete (or integer) rate adaptation. A sub-optimal scheme for integer bit loading is presented which approximates the optimal continuous rate value to a nearest integer. Next, we propose schemes that maximize the capacity of CR users when only imperfect channel state information (CSI) is available at the CR transmitter while guaranteeing the statistical interference constraint. Further, we propose resource allocation schemes for a multiuser scenario where power is loaded for CR users not only in the subcarriers where PU is not present (overlay fashion) but also in the subcarriers where PU is present (underlay fashion). Finally, for the scenarios where the link between CR source and destination might be weak and not reliable for communication, we employ relays and propose relay and power allocation schemes. Numerical results have been presented for all the proposed algorithms.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International