Design and Implementation of Real Time Face Tracking Humanoid Robot

Abstract

Humanoid robots are not solely a part of the science fiction; they are at the focus of many study areas in real life. After the developments in robotics and automation in the last century, robots are made not just to perform certain tasks; some are designed to bear a resemblance to humans. The main thrust of this project is to design and implement a humanoid robot that is capable of seeing, detecting and tracking a human face in a cost efficient and robust manner. In this study Viola-Jones face detection framework and Kanade-Lucas-Tomasi (KLT) feature tracker are utilized to build the desired humanoid robot. Certain constraints such as limited joint range, manipulability measures and singularity avoidance are all regarded on the end-effector for the implementation.