Key points are not available for this paper at this time.
Automation failures in Unmanned Aircraft Systems (UASs) significantly lead to a decrease in overall system performance, an increase in operator workload, and a deterioration in automation trust. This study investigates how the frequency and intensity of automation failures differ in multi-subsystem environments. An improved automated MATB (Multi-Attribute Task Battery) paradigm was used to quantify the frequency and intensity of automation failure at four levels. Through operational experiments incorporating eye-tracking technology, we examined the effects of different failure levels on dependent variables. Data were analyzed using descriptive statistics, ANOVA, and nonparametric tests, revealing that while failure frequency and intensity significantly deteriorated workload, automation trust, and task performance, not all dependent variables showed consistent changes across failure levels, indicating the presence of a plateau effect in certain cases. Trust in automation negatively mediated participants’ perceptions of workload in the context of failure. These results suggest that different failure frequency and intensity contexts can have differing effects on operators, especially in complex socio-technical systems involving multiple subsystems, which should not be generalized regarding whether they fail or not. In practical applications, designers could consider how to increase operator trust in automation (through personnel training, system design, etc.) and reduce the negative impact of automation failures on performance and workload.
Wang et al. (Sun,) studied this question.