Accession Number : AD0259906

Title :   ON THE IDENTIFICATION AND ADAPTIVE CONTROL OF SAMPLED-DATA SYSTEMS

Corporate Author : JOHNS HOPKINS UNIV BALTIMORE MD DIELECTRICS LAB

Personal Author(s) : LENDARIS,G.G.

Report Date : 21 MAR 1961

Pagination or Media Count : 1

Abstract : A procedure is presented for identifying a linear, lumped-parameter, time-invariant (at least during the measurement period), single-input system from its response to a step excitation. Then, an adaptive sampled-data control system is designed using a digital computer to perform this identification procedure and to act as the controller for the system. The system was actually built, and some experimental results are cited. This identification technique requires 2n + 1 measurements of the magnitude of the output, where n is the order of the system and the measurements are taken at equally spaced intervals of time. These measurements are used as the coefficients in n scalar equation whose n unknowns are the coefficients of the z-transformed characteristic equation of the system. This set of n scalar equati ns is solved for these coefficients (on a digital computer). For some applications, this characterization of the system is sufficient; when it is not, then this equation is factored to yield the system's z-plane poles, and from this, the corresponding s-plane poles are found. This procedure is independent of initial conditions and will work for any such system (even for the multiple eigenvalue case) the only restriction being that all the poles of the system be excited when the measurements are made. (Author)

Descriptors :   *CONTROL SYSTEMS, DIGITAL COMPUTERS, LINEAR SYSTEMS, MATHEMATICAL ANALYSIS, SAMPLING

Distribution Statement : APPROVED FOR PUBLIC RELEASE