Abstract
Because of the rich potential of the gesture recognition technology, the research about gesture recognition is always a popular issue. Traditionally, the gesture recognition system was implemented by the method of image recognition, which is a classic and mature technique, but the image recognition system not only spends a lot of computing time but also needs expensive devices. So we design and implement a system to recognize the human gesture and use a humanoid robot to demonstrate this system. This system consists of wireless sensor networks, triaxial acceleration sensors, and a humanoid robot. The whole system can be divided into three subsystems: 1) Wearable Sensors, 2) Host, and 3) Humanoid Robot. In the Wearable Sensors, the sensor nodes with triaxial acceleration sensor are installed in the human limbs and transmit acceleration data to Host. In the Host, a PC or laptop receives triaxial data and then recognizes the human gesture. After that, Host transmits a command message to the Humanoid Robot. In the Humanoid Robot, the robot receives command message from Host, and does the same gestures as user does.