Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Issue N5 2019 year
A new method to enhance spacecraft onboard software testing process based on decision rules application is proposed in the paper. Consider spacecraft onboard software consist of several interconnected subsystems, each of them must be thoroughly tested to meet requirements as well as the whole software complex. Existing onboard software testing process and its significant milestones are analyzed. As a result, weak points of the whole process are determined. Especially, subsystems test result analysis stage takes relatively a long time and a lot of effort because of its complexity and current low automation level. Also results arent saved systematically which makes it difficult for future analysis. The proposed method uses knowledge bases to produce conclusions such as "Ok" or "Failed" as a result of software testing both for its subsystems separately and for the whole software. Knowledge bases software is integrated into The Tests Automation System (TAS) software which is already developed. Every knowledge base comprises a software module that corresponds to particular software subsystem. Modules interact with TAS according to specially developed protocol based on xml standard. TAS executes the modules automatically after completion of software test to combine the results and save them into the database. The proposed method is evaluated within "Mars" design bureau during the development of onboard software for several spacecraft (Spektr-R, Elektro-L, Arktika-M, Spektr-RG) and has proven its effectiveness.