Development of a human-friendly robot for socially aware human-robot interaction

Abstract

One of the fundamental issues for service robots is human-robot interaction. In order to provide desired services, these robots need audiovisual perception of humans and to make appropriate feedback. In this paper, we present our system for spontaneous speech recognition, localization and identification of the user, recognition of the user's gestures, so as to perform the corresponding tasks like mapping, following the user and heading to designated location meanwhile avoiding the obstacles on the way. The system employs a new method of multi-feature detection for robot self-localization based on the recognition of artificial and natural features extracted from laser scans and RGB-D images, and they are quite discriminative in cluttered environments. An angle potential field (APF) method was used for obstacle avoidance. Experiments demonstrate the feasibility and effectiveness of the strategies in the paper.