Fuzzy risk matrix as a tool for the analysis of the air traffic safety

Air transport has been since years the safest way of traveling. It is a result of wide usage of risk analysis methods, affecting the majority of actions performed in the field of the air traffic management. Risk assessment is usually based on fusing probability of an adverse event with severity of its consequences. They are usually classified into one of the five exclusive categories. That practice seems to be inadequate. Assessments of probabilities and consequences are often imprecise, but are considered to belong to one particular category, that determines further risk assessment. However, there is a whole range of intermediate situations. In classic risk matrix it is difficult to take those situations into account. The aim of this study is to present the risk assessment method that allows expressing the risk using a continuous numeric scale. The fuzzy risk matrix was proposed that is in fact a fuzzy inference system having two inputs and one output. Thanks to applying that mechanism, we are able to compare the risk for different situations, even when they belong to one particular category e.g. tolerable. Such comparisons allow rationalization of actions as a part of risk management systems. In this paper the concept of a fuzzy risk matrix is applied to air traffic accident analysis. For the estimation of the accident probability a model based on Petri nets was used. To estimate consequences of a potential accident, an expert knowledge was used. In case of the analyzed incident the system that was created has a hierarchical structure. Experiments conducted by using this tool allow one to conclude that a fuzzy risk matrix is a powerful tool to assess the risk in air transport.