System and Method of Use for Non-parametric Circular Autocorrelation for Signal Processing

A computer-implemented method to characterize a signal structure is provided. The method may receive a first set of timedependent data vectors; select a sample size of data vectors; determine a minimum number of lags to process to characterize the first set of data vectors; calculate a cross-product term for each data vector in the sample size for each lag, a mean value and a variance value based on the sum of the cross-product term for each lag; calculate a circular correlation function for each lag; determine an upper and a lower correlation bound for the circular correlation based on the mean value, the variance value, and a desired false alarm rate; and characterize the signal structure based on the number of calculated circular correlation functions that exceed the upper or lower correlation bounds so as to generate a signal characterization data signal. The original document contains color images.