Distributed Power Control and User Selection Mechanism for Cognitive Radio Networks.

Cognitive radios (CRs) has been proposed as a viable solution to the impending spectrum underutilization problem, albeit, comes the risk of excessive interference to the licensed user of the spectrum called the incumbent user (IU). The research demonstrated in this thesis therefore formulates Distributed Power Control algorithms for CRs as a way of curtailing excessive interference to IUs owing to spatial coexistence. Duo quality of service (QoS) objectives is imposed for CR operation: (a) protection of incumbent users (b) support of CR terminals. Proposed therefore is a power control framework for cognitive radio networks (CRNs) exploiting spectrum sense and radio environment knowledge called distributed power control with incumbent protection via spectrum sensing (DPC-IPSS). The proposed DPC-IPSS algorithm is formulated for a worst case (WC) and non-worst case (NWC) scenario. Under WC scenario, the DPC-IPSS algorithm simultaneously fulfils QoS objectives by protecting incumbents in the outermost service area. However, in reality incumbents may not always be in worst case. Therefore, further proposed is a stochastic power control scheme for CRs under a NWC regime which allows the CRN to access extra capacity based on incumbent outage information. In addition, this research considered mobility patterns within the CRN. The DPC-IPSS algorithms fail in mobility driven environments. This is because time dependent channel variations were not taken into account. The approach therefore is to model the time driven channel and scale-up the target QoS metric to compensate for mobility within the CRN. Proposed therefore is a mobility driven power control algorithm with incumbent protection via spectrum sensing (MDPC-IPSS). Results show that MDPC-IPSS algorithms give performances in proxy of quasi-static cognitive radio networks. Finally, the user selection problem is further considered, however, it is known that the optimum user removal problem is NP-hard; therefore most approaches in literature employ sub-optimal centralised or semi-distributed mechanisms to solve the user removal problem. This thesis formulates an efficient outage based fully distributed user removal algorithm which is jointly executed with the DCP-IPSS scheme called (ODPC-IPSS). In ODPC-IPSS schemes, users autonomously remove themselves from the network based on a self accessed outage criterion. This results in reduced interference at the incumbents and within the CRN with a resulting qualitative increase in the cognitive subscriber base