Abstract: For the purpose of effective assessment of human reliability of air traffic controllers, a HRA technique tailored for air traffic controllers operation was developed based on HEART. GTTs were determined by going through the controller's daily tasks. Using the geometric mean method and sampling method, the nominal human error probabilities were calculated by processing the data from multi-sauces. EPCs were structured by following the idea of "4M" (Human-Machine-Media-Management), and EPCs were also referred to existing HRA methods EPC or performance shaping factors(PSF). The maximum impact values of EPCs were obtained by the Delphi method and the BN. The controllers' human error probability(HEP) in the scenario of air traffic de-conflict was analyzed by applying the method proposed in the paper. The result is between that by the traditional HEART and that by the traditional cognitive reliability and error analysis method (CREAM), which demonstrates the validity of the method.

Key words: air traffic controller, human reliability analysis(HRA), error producing conditions (EPC), human error assessment and reduction technique(HEART), generic task type (GTT), Bayesian networks (BN)