Estimation of Parameters in a Partially Specified Stochastic System,
Abstract
A partially specified stochastic linear system is one in which the input disturbance is uncorrelated with only a finite number of the previous values of the noise, output and the control input. The remaining higher order correlation coefficients of the noise are unspecified. In particular, the input disturbance could be a discrete white noise. The author develops a recursive algorithm for the estimation of the coefficients in a partially specified linear system, proves the consistency of the estimates, and illustrates the technique by numerical examples. Simulation studies indicate that the variance of the estimate is of the same order as the Cramer-Rao lower bound, even for sample sizes as small as 50. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1972
- Accession Number
- AD0746065
Entities
People
- R. L. Kashyap
Organizations
- Purdue University