Psychophysiological feedback control in physical human-robot interaction

Resumen

Human state aware" systems may be able to improve physical human-robot interaction (pHRI) by adapting their behavior and cooperation level accordingly to human psychophysiological feedback. This research work reviews the impact of the emotions on human motor control performance, and studies the possibility of im- proving the perceived skill/challenge relation on a multimodal pHRI platform. For this purpose, a biocybernetic controlled physically interactive multimodal platform that modulates the difficulty and assistance given to a subject in reaction to unde- sirable physical and emotional states was designed and implemented. This research work aims to a novel biocooperative system which is able to adjust haptic and dif- ficulty parameters in task time, according to the instant needs of the subject. The controller that drives the haptic interaction between subjects and the robotic device was designed using a human centered approach that includes the human into the control loop. The goal of the controller is to enhance the user experience by using Psychophysiological feedback as a guide of the overall state of the patient regarding physical and emotional loads experienced at task time. A Takagi-Sugeno type fuzzy inference system (FIS), was implemented, synthesizing several rules created from psychophysiology theory, and previous observations made by this and other groups. We have designed, developed and tested the necessary technology in robotics, digital signal processing, artificial intelligence and other engineering and medical areas, to create a multimodal physically interactive platform in order to perform open and closed loop experiments. Results of those experiments have shown that psychophisiological feedback estimates by themselves and in conjunction with performance calculations, can help considerably on improving interaction in a human-robot cooperative scenario. The results also suggests that as a side effect of modulating robotic assistance with the help of Biocybernetic control, leads to better overall user experience based on the reported levels of emotional valence, arousal and dominance assessed after each experiment.