基于指导型盲源分离和非高斯准则的全双工认知

Guided Blind Source Separation and Non-Gaussian Criterion Based Full-Duplex Cognitive Radio

为了克服半双工认知中存在的静态感知问题和常规全双工认知中残余自干扰限制的影响,提出了一种基于指导型盲源分离和非高斯准则的新型全双工认知方法。该方法将频谱感知和数据传输设计于同位置执行,避免在感知信息中的任何不匹配和资源损失,利用同位置配置中已知的次用户信号作为指导信号辅助执行盲源分离工作。在分离处理后,利用相关性识别出次用户自发信号,另一个信号通过非高斯准则判定,进而判决主用户的活动状态。仿真实验分析和讨论验证了所提方法的有效性,与基于自干扰消除的全双工频谱感知方案相比在计算复杂度和感知性能上具有明显的优越性。

Abstract: To overcome the static sensing problem in half-duplex cognitive radio(HD-CR),as well as the harmful restriction of residual self-interference in conventional full-duplex cognitive radio(FD-CR),this paper proposes a novel full-duplex cognitive radio method based on guided blind source separation(BSS) and non-Gaussian criterion.The functions of spectrum sensing and data transmission are designed in the same position avoiding any mismatch and resource loss in sensing information.The known secondary user in same position configuration is utilized as a guided signal to assist performing blind separation assignment.After separation,the secondary user signal is identified by correlation processing,and the other signal is recognized through non-Gaussian criterion to further decide the status of primary user.Simulation experiments and discussions corroborate the effectiveness of the proposed method.Moreover,the comparative experiment results demonstrate that the proposed method has the superior performance in computation complexity and sensing performance compared with the self-interference cancellation based spectrum sensing scheme.