中文摘要：

提出了基于指数嵌入族（EEF）准则的认知无线电宽带频谱感知算法.在EEF准则宽带感知算法的基础上,充分考虑信号子空间和噪声子空间的性质,利用Gerschgorin酉变换改进EEF算法,提出了基于Gerschgorin EEF（GEEF）的宽带感知算法.该方案主要是利用EEF准则或GEEF准则估算出授权频带中被主用户占用子带集合的势,从而完成宽带授权频段上空闲频带的感知,以便于被次级用户动态接入.理论推导和仿真结果表明,所提出的方案无须知晓任何有关噪声功率和主用户信号的先验信息,对噪声功率的不确定性具有鲁棒性.

英文摘要：

Exponentially embedded families（ EEF） criterion based wideband spectrum sensing schemes were proposed for cognitive radios. Based on the EEF criterion,the signal and noise subspaces were considered in spectrum observation data and the EEF wideband sensing was improved by employing Gerschgorin unitary transformation in data processing. Correspondingly,the developed wideband sensing algorithm was operated under Gerschgorin EEF（ GEEF） criterion. The proposed two algorithms estimate the cardinality of the primary user signal occupied subband set via EEF or GEEF criterion to identify idle subbands for secondary users dynamically accessing the licensed frequency band. Analysis and simulation verify that the proposed algorithms do not need any a prior knowledge of the noise power and the primary user signal. They are robust against the noise power uncertainty problem.