State Estimation in Linear Dynamic Systems via Initial State Identification

This paper deals with the estimation of the unmeasurable states of a linear, time-invariant dynamic system by first identifying the initial state of the system. This new approach makes use of the state transition concept to predict the future behavior of the system. A state estimator which operates in series with the initial state identifier is developed. Satisfactory state estimates have been obtained even in the presence of significant noise levels. A simple example is considered to illustrate the theory discussed in the paper.