Research On Safety Testing For Airborne Software

Safety is one of the important properties of airborne software, airborne software still exists the following problems :(1) When evaluate the safety of the software, itusually tends to be isolated software from system, so software‘s risk contribution to the system is hard to evaluate;(2)Safety testing is independent of the safety analysis, so it is difficult to ensure the completeness and correctness of safety testing requirements;(3) Airborne software safety testing framework is not yet into the system;(4)The study of safety testing resource distribution is relatively rear, and less consider safety hazard degree of components for software safety,when refer to testing resource allocation. Based on studies of airborne software safety related technologies, an airborn software safety testing framework, of some commonality, is proposed.Then new methods of safety analysis, safety assessment and test strategy distribution in the frame are given, which can solve above problems in some extent. This article’s main research content is as follows:(1) Based on existing research of safety testing frameworks, this paper puts forward an airborne software safety testing framework of some commonality, it conbines safety testing and safety analysis to ensure the correctness and completeness of the safety testing requirements and safety requirements of software.(2) Airborne software’s failure cause is analyzed, and put forward SISM to describe how the system entities interact with software.Combined DFM with FTA to conduct safety analysis, and extract safety information to generate software safety SISM,which focus on safety description anddepicts relationships between safety standard scenario, basic scenario and software behevior.Then based on normalized scenarios’ structure and frequencies of entitys’ states, structural importance for normalized scenario’s is defined,which will be a compute element for normalized scenario’s com-importance.(3)Considering airborne software’s behaviors’ trigger is associated with its interactions with all entitys’ states, Single Entity Condition Dependency Graph(SECDG), which posses Markov characteristic, is put forward, then the method to compute occurrence rate for related scenarios is given. On this basis, through defining start propability importance, running propability importance, start failure weight and running failure weights, combind with the above structural importance, com-importance, which have harzard importance, is defined. It is one of the parameters for the follow-up test strategy distribution.(4) Considering failure consequences of system failure is different when it starts and when in the runtime, a new safety assessment method is put forward, which evaluate the safety of software from the system level.(5) Comprehensive Considering normalized scenario’s structure importance, start occurrencerate, running occurrence rate and corresponding consequences, a parameter to evaluateharzard severity for normalized scenario: is defined,which is called Com-importance. As the same, Com-importance for basic scenario is defined, too. Based on both Com-importance, the test risk allocation strategy, which take risk target as test stop indicator, is proposed.Safety reliability is set according to the risk target, then specifictesting time for every normalized scenario is distributed through the setted safety reliability. Safety reliability, in the test-time allocation method, changes dynamically within testing, when the system’s safety is improving, the total time required to test is decreasing.Furthermore this method can ensure, higner the important degree and lower the safety reliability of a scenario, the scenario will be allocated more test time.Finally, this paper uses an example and illustrates the feasibility of the above safety testing strategy and method.