Abstract
The elements of the observer matrix are determined to meet the specified eigenvalues, distinct and multiple, and at the same time, the measure of quadratic estimate error expressed in block pulse functions is minimized. It is relatively simple as compared to other design methods for optimal observers.