Takeoff runway overrun risk assessment in aviation safety based on human pilot behavioral characteristics from real flight data

Aviation safety is an important part of safety science and pilot behavior is a major factor in flight safety analysis. However, when analyzing unsafe events, such as runway overrun, majority of studies do not quantify the impact of pilot behavior on flight safety. The aim of this paper is to discover the pilot behavioral characteristics from actual flight data and apply them to construct a takeoff runway overrun risk assessment model. Concretely, the time series clustering is used to mine the pilot behavioral characteristics based on pilot operating data. Then, the similarity theory is introduced to construct a data-based man–machine-environment system expression including pilot behavioral characteristics to get the takeoff distance for the runway overrun risk assessment. The results of case study showed that 3 types of pilot behavioral characteristics were found in selected fleet, which mainly differ in the pilot reaction time after V_R and operation input speed of the control column, and RO is more likely to happen under the combined influence of longer reaction time and slower operation input speed. The proposed method makes full use of real flight data from airlines under different conditions incorporating pilot behavior characteristics and can inform risk analysis of other unsafe events such as hard landing, tail strike, etc.