Estimation of the Number of Signals in the Presence of Color Noise

Lu Gan,Xin Qi,Liping Li,Huaguo Zhang
DOI: https://doi.org/10.1109/icccas.2009.5250500
2009-01-01
Abstract:In this paper, a new method to detect the number of signals in the presence of color noise is proposed. Firstly, pseudocovariance matrix is constructed using the covariance of sensors. Then, the Information Theoretic criteria and Gerschgorin Disks are used to estimate the number of signals. Simulation results show that the proposed AIC and MDL can work well in the presence of color noise, while the classical ones fail. What's more, the proposed GDE method has superior performance to the original GDE.
What problem does this paper attempt to address?