“Estimation of the Parameters of Sinusoidal Signals from Noisy Observations”

Abstract

Estimation of the amplitude and frequency parameters of sinusoidal signals is an important problem  in many applications such as radar, sonar, and telecommunications. Statistical efficiency  and computational simplicity are among the desired properties of estimation methods.  The Gaussian maximum likelihood (GML) method has the first property but suffers from  serious computational problems. This talk reviews some alternative methods and discusses  in detail an iterative filtering technique that overcomes the computational problems and  retains the statistical efficiency of the GML method. It also discusses a nonlinear method capable of outperforming GML in statistical efficiency under heavy-tailed non-Gaussian conditions.

Bio

Ta-Hsin Li is a Research Staff Member at the IBM  T. J.  Watson Research Center. He received the Ph.D. degree in applied mathematics from the University of Maryland at College Park in 1992. From 1992 to 1998, he was a faculty member of the Statistics Department of Texas A&M University at College Station and the Statistics and Applied Probability Department of the University of California at Santa Barbara. He has been with the IBM  T. J. Watson Research Center since 1999. His main research interests include statistical theory and methods for time series analysis, signal and image processing, and spatial data analysis and modeling. He served as Associate Editor for the IEEE Transactions on Signal Processing (2000-2006) and the EURASIP Journal on Advances in Signal Processing (2006-present). He is also Adjunct Professor at the Electrical Engineering Department of Columbia University.

Dr. Li is a Fellow of the American Statistical Association (ASA) and an elected senior member of the Institute of Electrical and Electronic Engineers (IEEE).