Gesture Control for Human-Robot Interaction Based on Three-Way Decision Model

Abstract

The Gesture interaction is a new Human-Computer Interaction method which has gradually become the heated topic and attracted much attention. As the gesture is intuitive and natural, it has been widely used in the area of human-robot interaction. However, most existing gesture control processes focus on the effectiveness of successful operation. The consideration of the risk of misoperation is usually ignored. It may cause a great risks and much trouble to the robot control system. Therefore, this paper introduces three-way decision model into human-robot interaction process. In the three-way decisions, it is important to acquire the thresholds and decision costs. In this paper, the thresholds of the three-way decision model are the combination of decision risk and classification accuracy. The decision costs come from the aid of group decision making process. Moreover, as the gesture control process is a multi-category problem, it is transformed into several two-category problems. Because the transformed problems may cause conflicts in the decision actions, a divided threshold is considered to deal with the problem. To demonstrate the effectiveness of our proposed method, an illustration is presented. Finally, our experimental results show that the performance of the proposed method is better that that of the method based on two-way decision model.