An AADL Model-Based Safety Analysis Method for Flight Control Software

Abstract

This paper proposes a model-based software safety analysis method for flight control software. Firstly, the AADL architecture model and error model of flight control software are constructed separately. Then we convert the AADL model of flight control software into extended Markov chain model. We can evaluate safety of software component by calculate the probability of component state in different hazardous levels with the extended Markov chain model. Finally, safety of entire flight control software system can be calculated according to AADL error states composition relation, which can avoid states explosion problem of Markov chain model.