Document Details

Document Type : Thesis 
Document Title :
Gesture-Based Assistive Robotics Children Education through Enhanced Interaction
التعليم التفاعلي للاطفال باستخدام الايماءات مع الانسان الآلي
 
Subject : Faculty of Computing and Information Technology 
Document Language : Arabic 
Abstract : Due to the huge revolution in research of using a robot in the surrounding environments, the need for enhancing the way of interaction with a robot revolution. Usually, a user uses voice commands to control the robot, which uses speech recognition to act accordingly. The hand gesturing is one of universal expression way that is used by all humans even those people with special needs regardless of their age. In addition to being an essential tool for human communication, nowadays, we can use them to control and interact with robots. Hence, the need to study the best devices to capture hands and employ them effectively to improve the interactive environment between man and robot. The evolution in the field of human interaction with the computer resulted in the use of robots as in most aspects of life and educational entertainment for all age groups. The employment of the machine in some respects requires a greater accuracy in the results, and this requires the use of additional tools and software to overcome the weaknesses of the robot and improve its capabilities even further. It is also known that the camera attached to the NAO robot head (color camera) is usually a normal and that the proportion of accuracy in distinguishing human hand gestures is much lower than the validity of the results using depth cameras on or special hand gesture sensors, as studies have shown. The research proposed hand gesture classification models for signs prediction that enrich the assistive robotics systems that use hand gestures as a communication means between human and robots. We proposed an architecture of full robotics system using hand gesture while we concern with enhancing the part of hand gesture recognition on that system. The architecture includes many services like hand gesture collection, generation classification models and prediction of new hand sign gestures. Two classifiers were used in models’ generation, SVM and KNN. We prepared the environment where the test experiments conducted. Participants are grouped according to their age, then hand signs of numbers (from 0 to 9) collected. Two samples of hand signs were arranged, each one is dedicated to achieving a specific goal. The validation sample were tested over our six prediction models we have created in implementation chapter. A comparative study for six different prediction models is conducted on validation dataset to compare the accuracy of prediction of hand sign gestures and test the processing speed among models. It shows that the prediction accuracy of both of the proposed classification models for hand signs increased as follows (from 69.25% to 98.29% for SVM models) and (from 55.5% to 83% for KNN models). 
Supervisor : Dr. Mohammed I Buhari 
Thesis Type : Master Thesis 
Publishing Year : 1438 AH
2017 AD 
Added Date : Thursday, August 3, 2017 

Researchers

Researcher Name (Arabic)Researcher Name (English)Researcher TypeDr GradeEmail
آلاء أحمد المرزوقيAlmarzuqi, Alaa AhmedResearcherMaster 

Files

File NameTypeDescription
 42450.pdf pdf 

Back To Researches Page