부산대학교 도서관

In order to reduce the probability of human error in the cockpit of a civil aircraft, based on the SHEL model, this paper classifies performance shaping factors (PSF) that affect the human factor reliability of the human-machine interface. Then it constructs the relevant PSF system and uses the interpretation structure model to establish the corresponding adjacency matrix. When a multi-level recursive structure is obtained, the Noisy-OR model is introduced to construct the Bayesian network, thus establishing the new human-machine interface evaluation method. Real cases are used to verify the effectiveness of the method thus established through causal inference with the Bayesian network. The evaluation method for the reliability of a set of human factors this paper proposes provides a new approach to evaluating the human-machine interface in the cockpit of a civil aircraft.