Correntropy-based spectrum sensing for wireless microphones in man-made noise environments

Correntropy is a generalized similarity measure that is here applied to estimate the spectral density which we called the correntropy spectral density. When the Gaussian kernel is utilized in correntropy, there is a free parameter. In this paper, we propose a weighted method to reduce the concerns about the free parameter selection. The proposed method is applied for detecting the presence of narrow-band wireless microphone signals in TV white space. The simulation results show that the proposed method has higher detection ability than the conventional power spectral density method in man-made noise environments and very close to the conventional in Gaussian noise, while maintaining low false alarms in both.