Developing Human-Machine Interfaces to Support Appropriate Trust and Reliance on Automated Combat Identification Systems (Developpement d'Interfaces Homme-Machine Pour Appuyer la Confiance dans les Systemes Automatises d'Identification au Combat); Contra

摘要

A series of laboratory studies examined the effects of system reliability information and interface features on human trust in, and reliance on, individual combat identification systems. The first experiment showed that participants had difficulty estimating the reliability of the 'unknown' feedback from these systems. However, providing reliability information through instruction led to more appropriate reliance on that feedback. The second experiment showed that both the system's activation mode and the feedback form influenced participants trust in the 'unknown' feedback. However, their reliance on 'unknown' feedback was not affected. Drawing from the results of these two experiments, information requirements for effective use of feedback from combat identification aids were derived. Several display prototypes were created from these requirements. A third experiment showed (1) that the method of displaying reliability information affected the participants sensitivity in discriminating the target from noise, and (2) that the display format (integrated vs. separated) affect the participants reliance on the system. Taken together, the experimental findings yield implications for the design of interfaces for individual combat ID systems and the training of infantry soldiers. Finally, a new method of measuring reliance on automation was developed and employed across all three experiments, demonstrating several advantages over previous methods. This methodological innovation represents a substantial contribution to the analysis of reliance behaviour in joint human- automation systems.