Enabling autonomous systems to perceptually detect human performance improvements and their applications

Abstract

The research on human-robot interaction (HRI) has been an emerging topic of interest for both basic research and customer application. The studies specially focus on behavioral and cognitive aspects of the interaction and the social contexts surrounding it. HRI issues have long been a part of robotics research because the goal of fully autonomous capability has not been met yet. One of the most challenging problems is giving the robots an understanding of how to interact with human beings at the same logical level so that they may function not as passive tools, but rather as active agents that can drive the human interaction, instead of merely reproducing a sequence of movements. Hence, these robots must have higher level cognitive functions that include knowing how to reason, when to perceive and what to look for, how to integrate perception and action under changing conditions, etc. These functions will enable robots to perform more complex tasks which require tight human interaction; consequently, the robots can perform high level interactions such as teaching motor skills to unskilled people. Such functions include the extraction of symbolic descriptions based on the integration of the information coming through the robot’s sensors and the analysis of these descriptions to decide the action to be carried out in order to improve the humans’ performances. In this paper; in particular, some approaches and applications towards enhancing the understanding of the human performance to implement the evaluation module of the proposed General Transfer Skill System (from robot to human) are described.