Observability Checking to Enhance Diagnosis of Real Time Electronic Systems

This paper describes a new property checking approach in order to enhance the diagnosis ability of an electronic embedded system, included in an automotive application. We consider functional diagnosis that is not necessarily oriented towards electronic components, and may result from flaws in the design process. The idea is, at the highest levels of the design procedure, to be able to assess the observability degree of the aimed system. Our approach is based on the analysis of the system simulation results. It allows checking the observability property in a real time electronic system in order to improve its diagnosis capacity.To reach this objective, we have set up an iterative modeling process. The model of the real time electronic system represents the input for the simulation step, which precedes the property checking. It is possible to model the system in two different ways: either with a combined architectural and functional modeling using SystemC, or solely functional modeling using Matlab/Simulink. Then, we check the observability property of the system, if it is not verified, our contribution consists in adding a feedback to the modeling step to improve the model. Otherwise, we validate the result through a test in an electronic automotive platform using as input the Simulink model used for the simulation step.