Workload and task performance in human-robot peer-based teams

Abstract

Robotic technology is developing rapidly, and humans are beginning to be partnered with robots. Human-robot interaction is an interdisciplinary field encompassing research questions regarding the extent of the impact of robotic design, presence, task assignments, and behaviors on a human. This dissertation focuses on modeling and assessing human task performance and workload, while humans directly interact with robots to achieve tasks. The goal of this dissertation was building a usable database of workload and task performance metrics that are appropriate for investigating the differences between human-human and human-robot interaction, because these workload and task performance differences influence the design of robotic peers. Specific differences in workload and task performance were assessed via human performance modeling techniques and the analysis of appropriate metrics in three evaluations of peer-based human-robot teams in the first response domain. The first response domain has constraints that impact the measurement of task performance and workload, such as an unpredictable environment, mobility of tasks, specialized protective gear, and time-critical task goals; thus, potential metrics were investigated to address these constraints and included a variety of appropriate and collectable methods (e.g., commercially available physiological sensors, observational data, subjective responses).