A Research on Resource Allocation in Cognitive OFDM Networks
|School||Nanjing University of Posts and Telecommunications|
|Course||Signal and Information Processing|
|Keywords||Cognitive Wireless Network Orthogonal Frequency Division Multiplexing Resource Allocation Quality of Service Convex Optimization Fairness|
Orthogonal Frequency Division Multiplexing (OFDM) and Cognitive Radio (CR) technologies were proposed for reconciling the conflict between limited spectrum resource and increasing QoS requirements. OFDM is an efficient broadband capacity approaching technique and CR is a dynamic spectrum access technique under certain interference-limited criterion. This dissertation integrated these techniques and focused on the topics as follows.Firstly, CR and OFDM were combined. The type of resource allocation problem and their numerical solution methods were exhaustively summarized. On the base of two classic interference models (0-1 model and interference temperature model), an improved cognitive OFDM sub-carrier interference-limited model was proposed.Meanwhile, the optimal power allocation for single cognitive user was investigated. The necessary and sufficient condition of the optimal allocation was formulated using convex optimization. Through the definitions of equivalent noise and interference limit, an optimal algorithm was obtained under the compare between cognitive OFDM power allocation and classic water-fill method. With a further investigation of our algorithm’s efficiency, a sub-optimal high-efficient algorithm was proposed based on the introduction of variable length searching method using optimization theory. The three-stage relationship between interference limits and power restrictions was also investigated in numerical simulations.Finally, the multi-user resource allocation problem was intensively researched. For the network capacity-maximization problem, multi-users’ water-fill and sub-carriers selection strategy was proposed through a joint sub-carriers power variable, which transformed the non-convex problem into a convex one. For the QoS guarantee among cognitive users, weighted factors were specified on each link to improve the fairness of different cognitive links through different utility influences on the objective function. For real-time services, the allocation model was established as a heterogeneous network composed of fixed data rate (FDR) users and variable data rate (VDR) users, which was divided into two sub-issues based on a packing problem, and put forward to an allocation algorithm both meet the requirements of FDR users and the weighted capacity maximization of VDR users. The availability and efficiency of these algorithms were fully proved by simulation results.