摘要
协作频谱感知(cooperative spectrum sensing,CSS)可以有效消除频谱感知过程中路径阴影和衰落对单节点频谱感知的影响,但是此感知方式中却有很多潜在的安全问题.针对CSS中较为常见的频谱感知数据篡改(spectrum sensing data falsification,SSDF)攻击,利用感知用户的本地能量检测值和频谱分割技术得到的信道状态值作为数据集进行训练,采用遗传算法优化加权协作中的加权系数,通过改变SSDF攻击者加权系数的方式降低SSDF攻击对系统检测性能的影响.仿真结果表明,加权系数经过遗传算法进化15代左右会趋于稳定值,检测概率明显提高,克服了SSDF攻击对系统造成的危害,更有利于实现认知无线电网络快速及高效的检测要求.
Cooperative Spectrum Sensing can effectively eliminate the effects of path shadows and fading on single-node spectrum sensing during spectrum sensing, but there are many potential security issues in this sensing method. For the more common Spectrum Sensing Data Falsification attacks in CSS, utilize the user’s local energy detection value and the channel state value obtained from spectrum segmentation technology as the data set for training. In view of the more common SSDF in CSS, the local energy detection value of the sensing user and the channel state value obtained by the spectrum segmentation technology are used as the data set for training, and the genetic algorithm is used to optimize the weight. The weighting coefficients in operation can reduce the impact of SSDF on system detection performance by changing the weighting coefficients of SSDF attackers. The simulation results show that the weighting coefficient will be stabilized after about 15 generations of genetic algorithm evolution, and the detection probability will be significantly improved. It overcomes the harm caused by SSDF attacks to the system and is more conducive to achieving the rapid and efficient detection requirements of cognitive radio networks.
作者
李敏
陈跃斌
吴孟礼
LI Min;CHEN Yue-bin;WU Meng-li(School of Electrical and Information Technology,Yunnan Minzu University,Kunming 650500,China)
出处
《云南民族大学学报（自然科学版）》
CAS
2020年第3期279-284,共6页
Journal of Yunnan Minzu University:Natural Sciences Edition
