Cognitive radio, a solution for resolving spectrum scarcity problem encountered in many countries, has been regarded as one of the most promising technologies for future wireless communications. The critical requirement in cognitive radio design is to ensure that the primary users are well-protected. One way to do so is called spectrum sensing. Due to various practical constraints such as noise power uncertainty, conventional sensing methods are difficult to achieve satisfied detection performance in low signal-to-noise ratio (SNR) environment. This paper discovers how the spatial- and time-domain correlations of the received signals can be utilized for spectrum sensing design. Observing that the correlation matrix of the received signals changes from an identity matrix to a non-identity matrix when the primary signal becomes active, the authors make use of the eigenvalues of the correlation matrix to perceive the existence of the primary signals. The proposed sensing schemes are able to combat the noise power uncertainty issue, and can achieve good performance in low SNR environment. Furthermore, it builds up an interesting bridge between spectrum sensing and the random matrix theory, from which the sensing performance can be quantified with accurate closed-form expressions.
The methods proposed in this paper have been adopted in IEEE 802.22, the first international standard based on cognitive radio technology. The paper is the winner of the first IEEE Communications Society Asia Pacific Board best paper award.