Abstract
This study examined performance assessment methods used to initiate mode transfers between manual control and automation for adaptive task reallocation. Participants monitored two secondary displays for critical events while actively controlling a process in a fictional system. One of the secondary tasks could be automated whenever participant performance surpassed absolute threshold values or after evidence of a continued change in performance beyond the threshold criteria. Transfers were either machine-initiated or human-initiated after the computer signaled the participant to change modes. First, the results support previous findings (Montgomery, 2001). In addition, it appears that by including the human operator in machine-initiated transfers the proportion of mode errors (i.e., accidental responses while a task is automated) may be reduced relative to transfers that are completely machine-initiated with the operator being simply informed of the current mode. Secondly, the results indicated that the assessment method that required a change in performance produced performance advantages, including evidence of a decrease in the proportion of mode errors, relative to the absolute threshold criteria without a heavy reliance on automation.
Get full access to this article
View all access options for this article.