An integrated graphic-taxonomic-associative approach to analyze human factors in aviation accidents

Human factors are critical causes of modern aviation accidents. However, existing accident analysis methods encounter limitations in addressing aviation human factors, especially in complex accident scenarios. The existing graphic approaches are effective for describing accident mechanisms within various categories of human factors, but cannot simultaneously describe inadequate human-aircraft-environment interactions and organizational deficiencies effectively, and highly depend on analysts' skills and experiences. Moreover, the existing methods do not emphasize latent unsafe factors outside accidents. This paper focuses on the above three limitations and proposes an integrated graphic-taxonomic-associative approach. A new graphic model named accident tree (AcciTree), with a two-mode structure and a reaction-based concept, is developed for accident modeling and safety defense identification. The AcciTree model is then integrated with the well-established human factors analysis and classification system (HFACS) to enhance both reliability of the graphic part and logicality of the taxonomic part for improving completeness of analysis. An associative hazard analysis technique is further put forward to extend analysis to factors outside accidents, to form extended safety requirements for proactive accident prevention. Two crash examples, a research flight demonstrator by our team and an industrial unmanned aircraft, illustrate that the integrated approach is effective for identifying more unsafe factors and safety requirements.