Human error probability estimation for safety and diagnostic systems in railway engineering

Abstract

Human reliability is a very important multidisciplinary topic. The reliability assessment of safety related system such as railway signaling system should also consider the impact of the human behaviour. RARA (Railway Action Reliability Assessment) is the only human reliability analysis method used in railway field. It is a powerful and effective technique; however, it strongly depends on the expert judgment, suffering of subjectivity. Generally, there is a lack of information for the human aspect of a certain task, fuzzy theory could handle this type of uncertainty and provides a trustworthy result. This paper proposes a new technique customized for the railway field which involves the fuzzy theory to evaluate the Human Error Probability. The approach is validated on a railway case study and the obtained probability is compared to the value provided by RARA method, to discuss advantages and disadvantages of the techniques.

Keywords

• Artificial intelligence
• Diagnostic
• Fuzzy logic
• Human reliability
• Railway engineering