Title :
Efficient autocorrelation estimation using relative magnitudes
Author :
Sullivan, Mark C.
Author_Institution :
ST Res. Corp., Newington, VA, USA
fDate :
3/1/1989 12:00:00 AM
Abstract :
The relative magnitude estimator computes the normalized autocorrelation of a stationary Gaussian process using simple nonlinear functions to eliminate most multiplications. Approximate expressions for bias and variance of the estimate are presented along with the results of a computer simulation. The estimator performs well when compared to similar techniques
Keywords :
correlation theory; autocorrelation estimation; bias; computer simulation; nonlinear functions; relative magnitudes; stationary Gaussian process; variance; Acoustic signal processing; Artificial intelligence; Autocorrelation; Computer simulation; Costs; Gaussian processes; Signal processing algorithms; Speech processing; Taylor series;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
