Detection of the Number of Cyclostationary Signals in Unknown Interference and Noise
Stephan V. Schell, William A. Gardner · 2005
A criterion for detecting the number of cyclostationary signals having a specified cycle frequency and impinging on an array of sensors along with unknown interference and noise is presented. Such a criterion is required for practical implementation of signal-selective direction-finding algorithms such as Cyclic MUSIC. The new detection criterion is based on a combination of existing results in multivariate analysis regarding the significance of common factors, which are signal components that can be estimated from either of two data sets by means of a linear transformation, and information-theoretic detection, in which no subjectively chosen threshold is required. The new detection criterion is shown to operate properly in the presence of more signals than there are sensors and in the presence of unknown, spatially colored noise.