An Innovative Signal Detection Algorithm in Facilitating the Cognitive Radio Functionality for Wireless Regional Area Network Using Singular Value Decomposition

This thesis introduces an innovative signal detector algorithm in facilitating the cognitive radio functionality for the new IEEE 802.22 Wireless Regional Area Networks (WRAN) standard. It is a signal detector based on a Singular Value Decomposition (SVD) technique that utilizes the eigenvalue of a received signal. The research started with a review of the current spectrum sensing methods which the research classifies as the specific, semiblind or blind signal detector. A blind signal detector, which is known as eigenvalue based detection, was found to be the most desired detector for its detection capabilities, time of execution, and zero a-priori knowledge. The detection algorithm was developed analytically by applying the Signal Detection Theory (SDT) and the Random Matrix Theory (RMT). It was then simulated using Matlab® to test its performance and compared with similar eigenvalue based signal detector. There are several techniques in finding eigenvalues. However, this research considered two techniques known as eigenvalue decomposition (EVD) and SVD. The research tested the algorithm with a randomly generated signal, simulated Digital Video Broadcasting-Terrestrial (DVB-T) standard and real captured digital television signals based on the Advanced Television Systems Committee (ATSC) standard. The SVD based signal detector was found to be more efficient in detecting signals without knowing the properties of the transmitted signal. The algorithm is suitable for the blind spectrum sensing where the properties of the signal to be detected are unknown. This is also the advantage of the algorithm since any signal would interfere and subsequently affect the quality of service (QoS) of the IEEE 802.22 connection. Furthermore, the algorithm performed better in the low signal-to-noise ratio (SNR) environment. In order to use the algorithm effectively, users need to balance between detection accuracy and execution time. It was found that a higher number of samples would lead to more accurate detection, but will take longer time. In contrary, fewer numbers of samples used would result in less accuracy, but faster execution time. The contributions of this thesis are expected to assist the IEEE 802.22 Standard Working Group, regulatory bodies, network operators and end-users in bringing broadband access to the rural areas.