VR-Based Motion Detection and Behavior Understanding

 < Day Day Up > 



Students in a virtual university have individual learning profiles, which may include exam records, Web site navigations, chat room discussions, and even their behaviors in a virtual campus. Motion detection of students in a 3D campus can be easily obtained. However, it is difficult to analyze the semantics of student motion. We aim to develop a behavior supervision machine, based on finite state automata, to properly guide students while they are on the campus.

In the virtual campus, five scenes are constructed. Theoretically, it is possible for students to go from one location to another without any restrictions. In the real world, learning activities may proceed according to some templates of actions. For instance, after a student attended a class, he or she may go to the library to check out a reserved reference assigned by the instructor. And, the student needs to make an appointment with the instructor before he or she can solve a question. Finally, the student will turn in homework before the next class starts. These actions can be captured by action control buttons in each scene. A nondeterministic finite state automata (NDFA, see Figure 8) is used to represent a possibility of student actions. A NDFA can be converted to a deterministic finite state automata (DFA) that can precisely monitor student actions. In this DFA, each student may navigate with a different pass. Thus, a language is derived from the DFA to enable student navigation on the virtual campus. There will be more than five states of the DFA according to the conversion from a NDFA to a DFA. Each of these states represents an intermediate navigation status. On the other hand, each action control button represents a link between states. The traversal of DFA w. r. t. a student can be recorded, as well as the immediate states. In each immediate state, there is a set of rules (see discussion of the third year proposal) associated. These rules are used as the base for a virtual patrol (see next year’s proposal) to restrict the navigation of a particular student. For instance, if a student goes to the open plaza and talks to other students all the time, without attending the class, the virtual patrol can either send a message to the student or simply bring the student to the classroom the next time the student logs in to the system. Campus regularity can be designed to monitor student behaviors in this virtual university. In order to realize the virtual campus state machines, we define a language and implement an interpreter to monitor student behavior. The language and machine will be discussed in the proposal of the third year.

click to expand
Figure 8: A Nondeterministic finite state automata for the virtual campus



 < Day Day Up > 



Designing Distributed Environments with Intelligent Software Agents
Designing Distributed Learning Environments with Intelligent Software Agents
ISBN: 1591405009
EAN: 2147483647
Year: 2003
Pages: 121

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net