A NON-PARAMETRIC RECURSIVE IDENTIFICATION ALGORITHM FOR FINITE-MEMORY LINEAR SYSTEMS,
Abstract
The problem of identifying a linear, finite memory, discrete-time system with noisy input and output is considered. A recursive algorithm is proposed which gives asymptotically unbiased estimates of the system parameters. The procedure does not require any specific knowledge of the noise statistics, in contrast to other proposed algorithms which generally need some knowledge of the statistics in order to obtain unbiased parameter estimates. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1969
- Accession Number
- AD0698770
Entities
People
- S. C. Schwartz
- Y. Bar-shalom
Organizations
- Princeton University