Identification of Autonomous Systems based on The Standard Deviation

Abstract

This study describes an autonomous system identification technique using a digital computer. This method constructs the system model in state space in a digital computer beforehand and identifies repeatedly until the model state transient follows the system state transient to a prescribed accuracy. Simulative experiments of second order systems based on this principle were tried and the relation between sampling period and error (or standard deviation) of the estimated system's parameter was made clear. Further an algorithm of identification based on the standard deviation is proposed and this one shows to be succesful to identify unknown parametric systems.