ABSTRACT
In the current work a new, signal based observability measure is proposed and applied. This measure surveys the observability from the signal analysis point of view. In contrast to existing measures it can be computed for nonlinear systems and makes no stringent requirements on the signal properties. As an example for a highly nonlinear system a railway bogie model is implemented that comprises a nonlinear wheel-rail contact formulation. The newly developed index is evaluated with regard to the longitudinal dynamics of the bogie and in this way the index allows for a reasonable choice of the most meaningful sensor signals. Finally, a comparison with already existing indices illustrates the benefits and weaknesses of the new observability measure.
