CONFIDENCE REGION OF IDENTIFIED PARAMETERS DUE TO EFFECTS OF MEASUREMENT NOISE

Abstract

This paper presents a method to estimate the effect of measurement noise on identified parameters in structural parameter identification problems. The effects of noise is given as a confidence region of identified parameters using the propagation law of errors and sensitivity of parameters with respect to errors. Furthermore, regarding an identification method based on the concept of prior information, a similar formula for the confidence region is derived. The validity of the formula is confirmed by numerical simulation.