Trust-Based Analysis of an Air Force Collision Avoidance System

Abstract

<jats:p> This case study analyzes the factors that influence trust and acceptance among users (in this case, test pilots) of the Air Force’s Automatic Ground Collision Avoidance System. Our analyses revealed that test pilots’ trust depended on a number of factors, including the development of a nuisance-free algorithm, designing fly-up evasive maneuvers consistent with a pilot’s preferred behavior, and using training to assess, demonstrate, and verify the system’s reliability. These factors are consistent with the literature on trust in automation and could lead to best practices for automation design, testing, and acceptance. </jats:p>

Would You Fix This Code for Me? Effects of Repair Source and Commenting on Trust in Code Repair

“…Importantly, the trust in automation literature has focused on transparency as a key antecedent of trust in automation. For example, Lyons et al [29] found trust in an automatic ground collision avoidance system in fighter jets increased when information about the functioning of the system was displayed on the screen. Similarly, Alarcon et al [16] found transparency was a key factor in whether programmers would reuse code that was generated by other programmers, and research has consistently demonstrated transparency as a key factor in perceptions of software [14,20].…”

“…One alternative may also be that programmers do not understand how GenProg operates "under the hood." Specifically, the process by which GenProg uses extant information and creates patches that pass test cases may not be understood by programmers, and this lack of transparency leads to a lack of trust [29]. Future research should investigate whether transparency is indeed the most important factor leading to differences between trust towards human and automated repair tools in software evaluation contexts.…”

Innovation Potentials and Pathways Merging AI, CPS, and IoT

“…It is interesting to compare the development of autonomous (road) driving with existing technologies and concepts in other areas of transportation: Within the airline as well as rail and shipping industries, partly automated and autonomous piloting is a long-standing and accepted situation (see e.g., [67,68]). Therefore, it is astounding that current expectations regarding autonomous driving within road transportation and trucking are on the one hand met with many reservations and on the other hand extreme job losses are feared-in spite of the fact that in the airline, railway or shipping business such job losses have never been happened.…”

Drivers Information Gathering Pattern During Transitions to Manual Control: A Study About Hmi Design for Autonomous Vehicles

“…This is caused by effects of over trust and overreliance on the automation (Parasuraman & Riley, 1997), where even when the human is monitoring the system, he/she assumes it will not fail. One example that can be seen of previous studies about this phenomenon was reported by Lyons et al (2016) on the field of military aviation, where the researchers found that pilots tend to push over the safety margins of the aircraft, the more reliable its automated control system is.…”

“…As already observed on the previous chapter in the OODA LOOP model, this is an interactive process, deeply related to the operator's past experience with the system, which reaffirms the dynamic aspect of the automation bias and complacency (Manzey & Bahner, 2005). Other factors such as the system reliability and interface clarity also seem to interfere directly with the phenomena, as it deals directly with the operator's trust (Dixon & Wickens, 2006;Lyons et al 2016). Some authors, such as Jones (1992) believe that this is a fatalistic process, where every individual that is exposed to automation will eventually become complacent due to a lack of stimuli for their attention.…”