Using the bicausality concept to build reduced order observers in linear time invariant systems modelled by bond graph

In this paper a bond graph approach to build reduced order observers for linear time invariant systems is shown. This approach uses the bicausality concept to simplify the construction and the calculation of the observer. The method is based on the Luenberger's algebraic method for the design of reduced order observers; some classical matrix calculations are simplified and the calculation of the inverse of matrices is not needed, which is an important improvement mainly for large-scale systems. The calculation of the observer gain is based on the pole placement techniques for linear systems modelled by bond graph. As application, one example with multiple outputs is developed.