Application of Statistical Estimation Procedures to the Identification Problem
Abstract
The method of maximum likelihood parameter estimation is applied to the problem of measuring parameters of an unknown linear filter or control system from input-output data when it is assumed that the output signal is corrupted with an additive Gaussian noise signal. The physical realization suggested by the integral formulation of the estimation technique is discussed and illustrated. Approximate expressions for the parameter estimates and the covariance matrix of the errors in the parameter estimates are obtained in the strong signal case. This analysis also has applications to the adaptive radar problem.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 03, 1961
- Accession Number
- AD0266877
Entities
People
- J. C. Lindenlaub
- R. P. Wishner
Organizations
- Massachusetts Institute of Technology