Assessment of piloting behavior impact on landing risk of transport aircraft

The human factors and their interaction with other factors play an important role in the flight safety of transport aircraft. In this paper, a paradigm of risk assessment for transport aircraft interacting with piloting behaviors is proposed, with focus on landing which is the most accident-prone flight stage in aviation safety statistics. Model-based flight simulation serves as our data source for landing risk analysis under uncertainties. A digital pilot in the loop that reflects the human piloting behaviors is employed to facilitate simulation efficiency. Eight types of unsafe events in landing are identified from statistics. On this basis, the landing safety boundary is extracted via stochastic simulation to divide safety and hazardous flight status domains, which contributes to flight status management and risk warning. The simulation results indicate that appropriate piloting behavior, which is active response and fast target acquisition with minimum overshoot and fluctuation, shows benefit to landing safety. The subset simulation technique is employed to further refine the boundary with less computational workload. Furthermore, the effect of airspeed, windspeed, and other factors on landing risk is also discussed. The proposed risk assessment method would help optimize operation procedure and develop targeted pilot training program.