Intelligent robotic system for home services - computer vision system

Abstract

With the increasingly reliance on robotic technology in human‟s life, research in servicing robot becomes more and more popular in the current era. This project mainly focuses on the development of a basic computer vision system for a home service robot to participate in RoboCup competition which is held in Singapore, 2010. The system is supposed to act as robot‟s eye in recognizing human faces and memorized new person in a dynamic environment. This function is highly important in performing human-robot interaction. The face recognition system has been designed based on Principal Component Analysis (PCA) together with k-nearest neighbor classifier. PCA is found to be an efficient method for real time implementation. However, a lot of factors may affect the accuracy of the system in a dynamic environment. This is always the greatest challenge in most of the robotic project. To solve this problem, some basic idea of image processing technique is introduced in the system. Real time testing has been performed afterward in different environments such as laboratory, bedroom, and exhibition hall in order to obtain a consistent performance and increase the reliability of the system. Lastly, this system is also combined with other parts of the robot such as speech recognition system and movement control system. The robot is now able to move by following the designated person according to his face after receiving a voice command “follow me”.