Development and Analysis of ARMA Parameter Estimation Schemes in the Presence of Noise
Abstract
This research includes the development and analysis of signal processing estimation algorithms. The main areas of application are sensor array processing for source localization, adaptive signal processing, system identification, and estimation. In the sensor array processing area we developed algorithms for source localization with decentralized array processing, sensor localization, passive range and bearing estimation, and source localization in multipath and short data applications. We introduced statistical tools that were used to provide compact expressions of the asymptotic variances of source localization algorithms and to obtain, for the first time, concrete analytical performance comparisons. Performance bounds of the Cramer-Rao type were found and used to analyze statistical efficiency. We introduced the use of electromagnetic and acoustic vector sensors for passive source localization and for active target localization and identification. We derived performance bounds and investigated the potential advantages of methods using vector sensors. Simple algorithms for estimating source direction with a vector sensor were proposed along with their statistical performance analysis. In the adaptive signal processing and system identification area we developed algorithms for root factorization and nonlinear filtering. We analyzed the tracking properties of our previously introduced adaptive notch filter for nonstationary signals.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1993
- Accession Number
- ADA267184
Entities
People
- Arye Nehorai
Organizations
- Yale University