Human Performance in a Mixed Human-Robot Team: Design of a Collaborative Framework

In a mixed human-robot team, adaptive automation methods can be used based on mental and cognitive states of human operators. Such adaptive behaviors can be designed such that lead to mitigation of human errors and consequently improvement of the task performance. However, real-time estimation of human internal states and their effects on the task performance remained a challenging issue and it has been the focus of many research in the recent years. Several studies have shown the capabilities of physiological feedbacks to assess human states in multi-tasking environments. In this paper, we present the early development of an experimental setup to investigate human physiological data during interaction with a small group of robotic agents. A simulated tele-exploration task is accomplished by participants and their brain activity and eye movements are recorded across the experiment. Statistical analysis are applied on the quantitative metrics to investigate the main effects and correlations between task performance and physiological features.