EPIIC

In order to enhance human-machine interaction, usual metrics include questionnaires and performance measures which do not allow for a continuous and online assessment of the quality of interaction, nor for a direct cognitive state monitoring of the human operator. In recent years, the development of physiological computing methods including that of brain-computer interfaces has enabled the rise of symbiotic systems that adapt the interaction using involuntary user inputs. Yet, to our knowledge, this technology has never been applied to human-robot interaction (HRI) in the context of mobile and collaborative robotics. This might be due to several challenges that need to be overcome, including the impact of user physical activity on the acquired metrics. This project will provide the first evaluation of the usability of electrophysiological metrics from wearable sensors for a rich, out-of-the-lab and online quality of interaction (QoI) assessment for collaborative robotics. The main objectives will be to: i) characterize the users’ cognitive state -i.e. cognitive effort and automation surprise- during collaborative and mobile HRI using involuntary electrophysiological features elicited by two standard collaborative robotic tasks -i.e. a joint navigation task, and an interactive manipulation task; ii) create an enriched QoI index that takes as inputs these features processed through a mental state estimation pipeline; iii) adapt the HRI thanks to this new QoI index; as well as iv) promote good scientific practices including data sharing.