相关盒维数的分步合作频谱感知方法

Step Cooperative Spectrum Sensing Based on Related Box Dimension

提出一种基于分形相关盒维数的分步合作频谱感知方法,通过对信号进行自相关运算,求取其盒维数值,利用恒虚警检测得到判决门限进行判决。在所有认知用户中随机选择参与运算的认知用户,并反馈有益信息。仿真结果表明,该方法在保证检测准确率的同时,降低了认知系统合作频谱感知所带来的额外开销和频谱检测时长,减轻合作控制中心的负荷,整体上提高了认知系统的检测性能。

A CFAR（ constant false alarm） step cooperation spectrum sensing method based on fractal related box dimension is presented. Through the autocorrelation operation on signals,the box dimension can be gotten. By the threshold resulting from the constant false alarm detection,whether the signal is existing can be gotten. This method chooses cognitive numbers and its users from all users randomly,and then produces beneficial feedback to central controller. Simulation results show that the method can not only ensure the correctness,but also reduce the extra data overhead resulting from the cooperative sensing and shorten the sensing search time. Meanwhile,it is able to ease the burden of central controller and improve the performance of system as a whole.