Improving Pilot Safety Performance Assessment: a Focus on Human Errors with Fuzzy Bayesian

EasyChair Preprint 15662

Abstract

Safety is one of the objectives of ergonomics in human-machine interaction. Accurately assessing human errors is of great significance for improving ergonomic design of human-machine interfaces and enhancing the safety of human-machine systems. Human errors in aviation are key factors affecting flight safety. Current flight data records a large amount of interaction data between pilots and aircraft, which can provide reference for the evaluation of human errors in aviation. This study extracts typical human error interpretation rules based on Flight Crew Standard Operating Procedures (SOP), including four evaluation dimensions: speed control, attitude control, configuration control, and trajectory control, to form indicators for judging pilot human errors based on flight data. Furthermore, an evaluation model for human errors in aviation is constructed based on fuzzy Bayesian networks using expert experience, identifying key human errors that affect flight safety, and guiding pilots’ daily flight training. This research aims to enhance human-machine interaction safety from the perspective of human factors, providing references for further reducing aviation accident risks and enhancing ergonomic design of human-machine interfaces.

Keyphrases: Fuzzy Bayesian, human errors, human pilot, safety performance

@booklet{EasyChair:15662,
  author    = {Chongfeng Li and Ziyao Li and Xing Pan},
  title     = {Improving Pilot Safety Performance Assessment: a Focus on Human Errors with Fuzzy Bayesian},
  howpublished = {EasyChair Preprint 15662},
  year      = {EasyChair, 2025}}