GESTURES RECOGNITION BY INERTIAL SENSORS USING KOHONEN NETWORK

Abstract

In this paper, a new wearable system to measure a gesture is described. It consists of a set of sensors (accelerometer and gyroscope) attached to the index finger, and of a microcontroller for index finger acceleration and angular velocity acquisition and conditioning. The paper focuses on the gesture recognition using the Kohonen neural network. The results show that the proposed system allows recognizing the basic gestures in aims to further implementation of the control device for man-machine gesture interaction.