A Novel Architectural Pattern to Support the Development of Human-Robot Interaction (HRI) Systems Integrating Haptic Interfaces and Gesture Recognition Algorithms

Hap tic and robotic interfaces are recently gaining momentum to be pervasively integrated in modern everyday life. In fact, they can be employed in several different fields, ranging from manipulation of small and dangerous objects to rehabilitation, assistive and service technologies, and are also integrated in mission critical systems. Modern research is rapidly shifting to investigate novel and more intuitive ways of controlling these interfaces. In particular, gesture-based control is one of the most interesting scenario for Human-Robot Interaction (HRI), since we human perceive gestures as a natural way of interaction with the external world. In this work we present a novel architectural pattern, entirely based on the Robotic Operating System (ROS), to support the development of applications and systems where computer vision techniques are applied to control robotic interfaces. As case study, the presented pattern is used to develop and assess the overall PARLOMA system. PARLOMA project aims at developing a system to enable remote communication between deaf-blind subjects. The system is designed to send, remotely and in real-time, messages in tactile Sign Language from a sender to a deaf-blind recipient (or many recipients) by integrating hand tracking and gesture recognition algorithms coupled with bio-mimetic hap tic interfaces.