Research On The Resource Management In Cognitive Radio Networks

The concept of cognitive radio is proposed to improve spectrum efficiency since itequips wireless users with the capability to optimally adapt their operating parametersaccording to the interactions with the surrounding radio environment. The resourcemanagement in cognitive radio networks has been receiving an increasing attention in recentyears.Based on the survey on the fundamentals of cognitive radio networking technology andrecent advances in related research, the resource management in cognitive radio networks isinvestigated and the contributions and innovations are achieved as follows:1) To address the wireless channel heterogeneity, which is a unique feature in cognitiveradio networks, a cross-layer optimization framework is presented, which jointly considersthe spectrum sharing, scheduling and routing with the objective of minimizing the systemtotal activation time, i.e., maximizing the system capacity. A novel algorithm is proposed togenerate all of the system concurrent transmission link-channel sets. Based on that, the MILP(Mixed Integer Linear Programming) optimization framework is simplified to a LP (LinearProgramming) problem, so the global optimal solution could be easily found. Thecomputational complexity to construct the schedule matrix is simplied from O(2M) to O(M2)by the traversal algorithm. Simulation results verify the significant benefit for the cross-layerspectrum sharing in cognitive networks.2) To address the dynamic nature of the wireless channel and the network environment, aSINR-interference-model-based algorithm is proposed for resource management. Firstly, thenetwork utility maximization problem is decoupled into the congestion-control problem andthe scheduling problem. And then a distributed algorithm is presented to find the maximalthroughput of the network. The innovated idea is that not only the differential backlogbetween neighborhood nodes but also the capacity of the wireless channel are used fordetermining the priority of sending packets, and a conflict-free MAC is proposed to improvethe network performance. Simulation results show the throughput is improved when thenetwork is heavily loaded.3) In order to minimize the end-to-end packet delay for delay-sensitive applications, aSINR-interference-model-based algorithm is proposed to manage network resources. It allowsnodes to exchange local information with the neighbors, and find their own feasible actionsincluding the relay node and the transmission frequency channel, and then autonomously select the optimal one with minimal expect delay to destination. The distributed algorithm isshown through numerical simulations to outperform other solutions leading to a lowerend-to-end packet delay and packet loss rate.