Optimization Of Cooperative Spectrum Sensing In Cognitive Radio Networks

In cognitive radio networks, secondary users (SUs) conduct spectrum sensing to detect the presence of primary users (PUs). In fading environments, independent spectrum sensing of individual SUs often cannot obtain satisfactory sensing performance. Therefore, SUs can perform cooperative spectrum sensing (CSS) to improve the sensing performance, which is called cooperative spectrum sensing (CSS). Recently, how to optimally design the CSS method to maximize the interest of SUs under the requirement of PU protection has attracted increasing attentions and hence is studied in this thesis.We mainly consider two optimization problems for CSS. Firstly, we focus on the case of decision fusion and aim to obtain the sensing time that maximizes the throughput of SUs subject to the requirements of PU protection and QoS. To this end, we propose a constrained nonlinear optimization formulation and develop a penalty function based algorithm to derive the optimal solution. Secondly, we consider the case of data fusion. In order to detect the PUs accurately at a low computation cost, we use a linear CSS method, in which the sensing results are based on the linear combination of local statistics from individual SUs. To obtain the optimal weights allocated to the SUs, we propose a constrained nonlinear optimization problem. The requirements of fairness, QoS and PU protection are adopted in the constraints. We then design a penalty function based algorithm to solve this problem. Finally, computer simulations are conducted to validate the effectiveness of the proposed algorithms.