Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/1112
Title: Sensor-Supported Psychomotor Learning System
Authors: Aisha Essa Bourahma 
Supervisor: Dr. Maha Faisal
Degree Awarded: M. Sc Degree in: Computer Engineering
Keywords: Sensor; E-learning; Reference Model; Learning System; Gesture recognition.
Issue Date: 2019
Publisher:  Kuwait university - college of graduate studies
Abstract: Sensor-supported systems have been adapted for use in various domains, such as health, sports, and education. In this thesis, we proposed a reference model for sensor-supported learning systems and implementation of a sensor-supported psychomotor learning system for Islamic prayer. The reference model was based on a review of recent articles in this area and identification of different observable properties of learning systems, the way in which different sensor types (including motion, audio, identification [ID], image, environment, biophysical, and software sensors) are used to support learning, and educational objectives (cognitive, psychomotor, and affective). The proposed reference model could be used to represent the capabilities of current sensor-supported learning systems and provide a basis for the development of future ones. The Islamic prayer learning system was implemented using the proposed reference model to reach optimality in building a psychomotor digital system using the input sensing device, Kinect, as a motion tracking device. It gives targeted users (new Muslims) the ability to correctly learn Islamic prayers without a teacher. The system prototype evaluation provided high positive feedback from the test participants (new Muslim converts and teachers of Islamic studies). A gesture detection module for Islamic prayer was required to build the application prototype, which led to provision of a distinctive attribute collection for defining complex body gestures, such as bowings and prostrations. Using this body attribute collection instead of raw Kinect data solved the need for massive training data in order to achieve good detection accuracy in motion tracking systems. Evaluation results demonstrated high gesture classification accuracy (91.6%) using a random forest algorithm to classify 12 gestures depending on only 300 records.
URI: http://hdl.handle.net/123456789/1112
Appears in Programs:0612 Computer Engineering

Files in This Item:
File Description SizeFormat 
my thises 21.pdf7,59 MBAdobe PDFView/Open    Request a copy
Show full item record

Page view(s)

24
Last Week
5
Last month
checked on Mar 29, 2020

Download(s)

3
checked on Mar 29, 2020

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.