Simulation Pratique de Fonctions Aleatoires (Practical Simulation of Random Functions),
Abstract
Identification of single input-single output systems is achieved by minimizing the integral of the square of the deviation between model and system output, subject to same pseudo random input. The method is known to be effective if no unknown disturbance affects the system. If an uncorrelated additive noise corrupts the output, the identification is less accurate. The acheivable accuracy is a function of the signal to noise ratio and of the accuracy of the measurement of the criterion to be minimized. For small deviations, the sensitivity in parameter space is described accurately enough by a quadratic form. The effect of the errors in parameter identification on the response of the model is small, except in some frequency bands. In both deterministic and stochastic problems, the accuracy of the method could be significantly increased by the use of appropriate high pass or band pass filters. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1970
- Accession Number
- AD0721496
Entities
People
- J. Kryze
Organizations
- Mines ParisTech