As the fast development of microprocessors, more and more safety-critical system (SCS) are applied and adapted to many safety-critical industries. The dependability of SCS becomes imperative. Generally, the system dependability includes usability, reliability, maintainability and safety. But the SCS mainly focuses on reliability and safety. Dependability testing requires plenty of resource to ensure that the system under test (SUT) satisfies the dependability requirement.In the thesis, firstly, the test features of reliability and safety are mainly analyzed as the core parts in dependability testing. Secondly, the relative theoretical backgrounds on pattern-based testing and Bayesian evaluate algorithm are provided. And then a new SCS rapid testing method is proposed to improve SCS adaptive dependability testing. Verification pattern(VP) technology is applied to generate test script, so that the mass of test cases can be easier to generate automatically. Besides, the results of each test execution are evaluated and saved in calculation memory unit. The least quantity of scenario test case for next test execution will be calculated according to the promised SUT's confidence level. The feedback data are generated to weight controller as the guideline for the further testing. Finally, a comprehensive experiment study demonstrates that this adaptive testing method can really work in practice. This rapid testing method, verification pattern-based adaptive control, makes the SCS dependability testing much more effective. |