Developing a Home Service Robot Platform for Smart Homes

Abstract

The purpose of this work is to develop a testbed for a smart home environment integrated with a home service robot (ASH Testbed) as well as to build home service robot platforms. The architecture of ASH Testbed was proposed and implemented based on ROS (Robot Operating System). In addition, two robot platforms, ASCCHomeBots, were developed using an iRobot Create base and a Pioneer base. They are equipped with capabilities such as mapping, autonomous navigation. They are also equipped with the natural human interfaces including hand-gesture recognition using a RGB-D camera, online speech recognition through cloud computing services provided by Google, and local speech recognition based on PocketSphinx. Furthermore, the Pioneer based ASCCHomeBot was developed along with an open audition system. This allows the robot to serve the elderly living alone at home. We successfully implemented the software for this system that realizes robot services and audition services for high level applications such as telepresence video conference, sound source position estimation, multiple source speech recognition, and human assisted sound classification. Our experimental results validated the proposed framework and the effectiveness of the developed robots as well as the proposed testbed.