Human performance modelling for image analyst decision support design

Abstract

The critical element in human in the loop military systems, image analysts must make decisions quickly. Improving analyst performance remains a priority of military leadership. This model, and previous findings of synthetic aperture radar (SAR) human in the loop experiments were used to design and test a decision support system (DSS) for analysing electro-optical (EO) and SAR images. Cognitive task analysis was used to develop an operator function model of the analyst classification task. Findings indicate that a DSS increased decision confidence and time. Interestingly, analyst trust is higher when employed in the analysis of EO images, yet aided performance was more accurate on SAR images. As uncovered through human performance modelling, analyst accuracy, confidence and time aided by a trusted DSS is crucial to analyst performance. These results suggest while the developed DSS improved analyst confidence and accuracy, it did not do so while concurrently decreasing task time.