Design and Implementation of a Human Thinking Based Cognition Learning Algorithm for Humanoid Robots at a Basketball Competition of the FIRA HuroCup

Abstract

The main concept of this thesis is to design an algorithm that allows a humanoid robot to imitate human thinking behavior so as to learn the shooting pose for the FIRA HuroCup basketball competition The systems proposed in this thesis include image processing algorithms a learning algorithm and hardware architecture of adult-sized humanoid robots Before instructing the robot to do some motions a vision feedback control system is processed first and then executes the learning on the computer In an image processing system a CMOS webcam sensor is used on the robot as the eye and two internet protocol cameras are installed on the side and above the basket separately To recognize and segment the objects a recursive searching algorithm is developed for the issue Then a novel learning algorithm is designed by a human thinking conception proposed in “Thinking Fast and Slow” by Daniel Kahneman the Nobel Memorial Prize winner in 2002 Economic Science This algorithm is based on cognitive psychology which divides human thinking into two modes fast and slow The fast mode favors intuitive thinking while the slow mode favors rational thinking Furthermore the cognitive bias of the anchoring effect and the peak-end rule are integrated This algorithm is implemented and applied on the robot for it to learn the experience curves After the learning process the robot also accomplishes the basketball game in the FIRA HuroCup Eventually the experimental results show that the performance of its bionic learning method is very efficient In other words the learning method also verifies that the thinking mode of the human being is reasonable and can be applied on the robot

Date of Award	2014 Aug 6
Original language	English
Supervisor	Tzuu-Hseng S. Li (Supervisor)