Real-Time Interaction System of Human-Robot with Hand Gestures

Abstract

With the declining birthrate and the aging population, the shortage of labor has become more obvious. Therefore, the teleoperation of the robot as the substitute of human labour have become popular. In this study, an application of real-time gesture-based human-robot interaction is proposed using a Kinect sensor for elderly care. By employing gesture-based teleoperation of robots, demands, and conditions of the elderly such as sudden falling and the abnormality can be quickly recognized. The robot used in this study, iRobot Create, is a popular mobile robot for indoor usage which will recognize the following gestures: right-hand raising, left-hand raising, right-hand waving, and left hand waving, in corresponding to the control the robot's moving forward, backward, right turn, and left turn, respectively. Experimental results shown that the four types of gestures can be recognized by the iRobot Create with human gestures captured by the Kinect sensor. The corresponding actions in real-time were realized through the designed gesture-based remote human-robot interaction system. The experimental results are also can be seen from the Youtube channel.