IDENTIFICATION FROM APERIODIC DISCRETE-TIME DATA WITH APPLICATIONS TO THE ESTIMATION OF EXPONENTIAL PARAMETERS,
Abstract
A method for the identification of a single-input single-output time-invariant linear system from an arbitrary set of discrete-time observations is described. Interpolatory procedures are used to express the unknown parameters in terms of the data. Applications of the procedure to an overdamped second-order system are included and the identification method is shown related to the resolution of a set of discrete-data into a weighted sum of exponentials, and hence to Prony's method for equally-spaced discrete-data. The results compare well with those of other methods when linear interpolation is used to obtain the equally-spaced or other special data required by alternate procedures. Computer results indicate that the procedure can be used to identify systems when only limited quantities of aperiodic data are available, and that it could be used in a comprehensive digital computer program for identification for an adaptive control system. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1966
- Accession Number
- AD0630300
Entities
People
- Allen Klinger
Organizations
- RAND Corporation