AbstractTo deploy a safety-critical system it is imperative to have confidence in the system's under-pinning software. This is gained by performing software safety assurance. If there is not a sufficient level of confidence in the software then there is not a sufficient level of confidence in the system. Therefore, the system would not be able to be deployed in applications where safety is paramount. A traditional method to gain confidence in software is to develop it to a process centred on the life-cycle. This is subsequently judged against a set of predefined objectives and the judgement on the level of compliance to the objectives is taken to warrant a degree of confidence in the software. However, if only certain types of evidence are accepted to demonstrate compliance, e.g. process-based evidence, then the solution space is reduced and some technical solutions potentially excluded.
The aim of the thesis is to provide additional methods and success factors to potentially expand the scope of the current safety assurance processes.
This research has demonstrated how the use of diverse evidence can achieve an equivalent level of compliance to a full process-based approach and therefore that it can form part of a software safety assurance strategy. The research outputs have not previously been implemented within the software safety assurance domain prior to this research.
|Date of Award||23 Jan 2020|
|Supervisor||John H R May (Supervisor), Theo Tryfonas (Supervisor) & M.J. Hadley (Supervisor)|