Effects of unreliable automation and individual differences on supervisory control of multiple ground robots

Abstract

A military multitasking environment was simulated to examine the effects of unreliable automation on the performance of robotics operators. The main task was to manage a team of four ground robots with the assistance of RoboLeader, an intelligent agent capable of coordinating the robots and changing their routes based upon developments in the mission environment. RoboLeader's recommendations were manipulated to be either false-alarm prone or miss prone, with a reliability level of either 60% or 90%. The visual density of the targeting environment was manipulated by the presence or absence of friendly soldiers. Results showed that the type of RoboLeader unreliability (falsealarm vs. miss prone) affected operator's performance of tasks involving visual scanning (target detection, route editing, and situation awareness). There was a consistent effect of visual density for multiple performance measures. Participants with higher spatial ability performed better on the two tasks that required the most visual scanning (target detection and route editing). Participants' attentional control impacted their overall multitasking performance, especially during their execution of the secondary tasks (communication and gauge monitoring).

Imperfect Automation in Scheduling Operator Attention on Control of Multi-Robots

“…Participants in the 6 robot condition found no more victims than those controlling only 3 while missing significantly more that passed through their view, indicating less efficiency when more resources were available. The finding is consistent with the studies (Chen, Barnes, & Kenny, 2011;Squire & Parasuraman, 2010) which suggest that participants can perform less well with access to more resources. The second hypothesis that decreases in reliability would increase perceived workload was not borne out.…”

Designing Effective Soldier-Robot Teams in Complex Environments: Training, Interfaces, and Individual Differences

“…The PAC interactions were even more pronounced in the AiTR reliability study [5]; high PACs tended to perform poorly when using falsealarm-prone aids whereas low PACs tended to perform more poorly on the missprone aids showing a classical type X interaction indicating possible mistrust in the former case and over trust in the latter case. The RoboLeader studies investigated the usefulness of an intelligent agent to help supervise multiple robots [6] [7]. RoboLeader supervision vs. operator supervision did not result in improved target detection for both the 4 and 8 robot conditions but RoboLeader supervision did result in more rapid transit for the robots.…”

Neural Correlates of Trust in Automation: Considerations and Generalizability Between Technology Domains

“…When receiving feedback, Increased activation for human agents who provided good feedback in the left dorsomedial PFC and medial frontal gyrus. decrease in trust (Chen et al, 2011;Hu et al, 2016). As the PFC, responsible for complex cognition and working/short-term memory, was identified by a subset of the reviewed articles, investigations into the neural correlates of trust should control for the co-varying influence of workload on trust perceptions in the accompanying neural signatures.…”