Assistive machine design based on human physiological measurements offers variety of opportunities in both civilian and military applications. In such settings the machine would need to compute certain metrics associated with these measurements to decide how to assist the humans. In this paper, we study a set of metrics pertaining to human behavioral patterns in easy and difficult tasks taken on a touch screen computer game. Analysis of these metrics reveal that some of them can be utilized as indicators of task difficulty showing promise for their use by machines. The metric results are also compared with task performance metrics, as well as a reliable and established metric called pNN50 calculated from subjects’ heart rate measurements and serving as the baseline.

This content is only available via PDF.
You do not currently have access to this content.