Explanatory tool for decision support on data-fused panoramic cockpitdisplays,

摘要

Abstract: Data-fused panoramic displays are currently being developed to provide decision support to military pilots in air-to-air combat environments. This paper reports the findings from a series of studies conducted using simulated data-fused explanatory displays. The displays provided explanations, rather than advice, in an endeavor to keep the pilot in the decision loop while improving the accuracy and/or speed of his mission-critical decisions. In addition to the basic display of the threat type, speed and direction (the control condition), three forms of explanatory displays were presented: text only, graphics only, and text and graphics (redundant). These displays provided information regarding the hostile aircraft's missile engagement zones and success envelopes. For effective human-system interaction, an appropriate level of operator trust is required. A critical determinant of trust is the transparency of the system interface, which should allow the operator to assess the system's accuracy. Two experiments were conducted to investigate the issues of trust and transparency of the different display formats used in the decision support system. Participants were asked to identify the highest threat posed by three hostile aircraft in an air combat scenario in the first experiment, and to assess whether explanations generated for a single hostile aircraft were correct or incorrect in the second. The scenarios, which were designed to present real-world decision tasks, were developed through discussions with RAF aircrew. Low trust was found to have a negative impact on decision-making and situational awareness, with subjects in the low trust condition making fewer correct decisions and reporting lower levels of subjective situational awareness than subjects in the high trust condition. Decision-making with low trust was particularly impaired in the test only condition. Poorer memory for hostile aircraft was observed where explanations were provided, compared to the control condition, in which no explanations were provided. The inability of subjects to identify erroneous textural explanations suggested a lack of transparency in the textual explanations. These results are discussed with regard to the implications of the different display formats for decision-making and situational awareness with data- fused cockpit displays.!7