COMPARATIVE SIUDY OF SEVERAL PARAMETER ESTIMATION TECHNIQUES FOR A CERTAIN CLASS OF DYNAMIC SYSTEMS

Document Type : Original Article

Author

DR., ( PH.D )., Deputy chief of optoelectronic and automatic control department, M.T.C, Cairo.

Abstract

The desire of the practical engineers has been to obtain better knowledge of specific plants, thus facilitates the design of improved control systems and consequently lower operating costs. In general the identification problem is characterized by three elements :
— Mathematical models of the systems
— Input signals
— Criteria for best estimation
Regarding to these elements, the auther has chosen the most common mathematical models (and the related algorithms) suited for linear time invariant system with finite memory. At the same time, the used algorithms do not make strict limitations on the input signals. With respect to the criteria, it was chosen to be a functional of the error of the output response. A comparative measures of parameter estimation for three algorithms have been introdces. These algorithms are :
— The algorithms derived by the auther
— The algorithms derived by. R. Subbayyan; R.Nagarajan
— The algorithms derived by. L.C. Westphal.