Automatic Pseudo-Random Regression testing for GUI-centric embedded software

Embedded systems are implanted in all intelligent devices and are made usable to their customers through graphical interfaces. This paper proposes a novel platform independent automatic GUI testing framework for GUI-centric embedded systems through the use of Pseudo-Random Regression(PRR) testing methods. Random test sequences are generated from a meaningful set of possible events in-order to exercise all the possible transitions in GUI-centric applications. Our results prove the effectiveness of using XML documents to define user interfaces and the high percentage of state and transition coverage while ensuring failure recovery, in-case of transitions to unexpected menu states.