Quantifying Cognitive Efficiency of Display in Human-Machine Systems

Abstract

As a side effect of fast growing informational technology, information overload becomes prevalent in the operation of many human-machine systems. Overwhelming information can degrade operational performance because it imposes large mental workload on human operators. One way to address this issue is to improve the cognitive efficiency of display. A cognitively efficient display should be more informative while demanding less mental resources so that an operator can process larger displayed information using their limited working memory and achieve better performance. In order to quantitatively evaluate this display property, a Cognitive Efficiency (CE) metric is formulated as the ratio of the measures of two dimensions: display informativeness and required mental resources (each dimension can be affected by display, human, and contextual factors).

The first segment of the dissertation discusses the available measurement techniques to construct the CE metric and initially validates the CE metric with basic discrete displays. The second segment demonstrates that displays showing higher cognitive efficiency improve multitask performance. This part also identifies the version of the CE metric that is the most predictive of multitask performance. The last segment of the dissertation applies the CE metric in driving scenarios to evaluate novel speedometer displays; however, it finds that the most efficient display may not better enhance concurrent tracking performance in driving. Although the findings of dissertation show several limitations, they provide valuable insight into the complicated relationship among display, human cognition, and multitask performance in human-machine systems.