Development of a Framework for a Java-based Signal Processing E-Learning Platform

Abstract

The design and implementation of an interactive, but easy to extend and to maintain e-learning platform is a complex task. In order to achieve this, existing learning platforms have been reviewed concerning learning principles applied in them. Then, use cases have been defined and a prototype of a web based learning platform has been built. Out of this prototype creation process, the software architecture of the learning platform as well as a model for creating exercises have been developed. As a result of this thesis project, a well-structured, JavaServer Faces based distributed e-learning software system has been developed. This software system integrates MATLAB® functions and provides a well-structured user interface. The user is able to configure the input signals as well as the digital signal processing algorithms freely, which gives her the possibility to study the algorithm in a way she desires. The results of the algorithms are presented using interactive charts, which can be saved to local disk for later reference. Printable question sheets are attached to the exercises, which guide the student towards to learning goals defined in advance. By this thesis project it has been shown, how modern web technologies like JavaServer Faces, jQuery and Highcharts are used to create an e-learning platform with MATLAB® as a back end. The Model-View-Controller based software architecture of the learning platform allows to separate responsibilities and thus keeps the code understandable and clean. Due to its flexible software architecture, the learning platform can be extended by other exercises, but also by other back ends like GNU Octave. In the future, it shall be investigated how the e-learning platform can be extended to two and three dimensional signals like images and videos. Also, an integration of simple vector- and matrix exercises is desirable.