Application of SVD to sense wireless microphone signals in a wideband cognitive radio network

Abstract

In a wireless regional area network (WRAN), the presence of wireless microphone (WM) signals must be detected as primary users. However, very narrow bandwidth and low power makes it difficult to sense WM signals especially when these WM signals are distributed in a wideband spectrum. In this paper, a singular value decomposition (SVD)-based approach is presented to sense and estimate multiple WM signals in a wideband spectrum. After performing SVD on the received data matrix, the presence and the number of WM signals can be detected and then the center frequencies of these WM signals can be estimated. By doing so, it can be determined that which channels are occupied and which are still vacant. Such that those unoccupied spectra are still available for the secondary users and the spectrum efficiency can be improved. Simulation results prove the better detection performance by comparing the proposed method and the traditional energy detection. Simulations also show a high frequency estimation precision by using the proposed SVD-based algorithm as well.