Date of Award


Document Type


Degree Name



Computer Science

First Advisor

Anneliese Andrews

Second Advisor

Gareth Eaton

Third Advisor

Scott Leutenegger

Fourth Advisor

Chris GauthierDickey


Safety-critical systems are those systems that when they fail they could cause loss of life or significant physical damages. Since software now is an essential component of these types of systems, failures caused by software faults could come from flaws in the software development life-cycle. As a result, challenges unfold in two directions. First, in verifying that the software will not put the system in an unsafe state, and identifying external failures and mitigate them properly. Second, in providing sufficient evidence for an efficient safety certification process. In this study, we propose an approach for testing safety-critical systems called Model-Combinatorial Based Testing framework (MCbt). MCbt is designed by integrating combinatorial testing and fault modeling into model-based testing to generate tests for normal behavior, and robustness. MCbt is driven by safety certification and standards compliance. In MCbt, we model each component of the System Under Test (SUT) separately using Extended Finite State Machines (EFSM) to include unit level and integration level tests. MCbt also uses Communicating Extended Finite State Machines(CEFSM) to model the SUT interactions and generates tests at the system level. Combinatorial testing is used to efficiently combine tests from different components of the SUT. Fault modeling using fault trees is used to identify which of the combinations could cause failures to test for failure mitigation. We apply MCbt to various domains through case studies. The results show that MCbt is applicable, efficient and provides a variety of evidence to use in safety certification.

Publication Statement

Copyright is held by the author. User is responsible for all copyright compliance.


Received from ProQuest

Rights holder

Aiman S. Gannous

File size

392 p.

File format





Computer science

Available for download on Saturday, October 02, 2021