Alireza Tavakkoli, Ph.D.

Assistant Professor of Computer Science

Associate Director of Faculty Computing
Digital Gaming and Simulation Program Director

                    Office:                 UC 109-B
                    Phone:            (361) 570-4204
                    Email: tavakkolia [at] uhv . edu

Skip Navigation Links
Image Pending Approval

Integrated Virtual Reality for Human Robotics Interaction

This project aims to develop a fundamental framework for establishing an immersive virtual reality environment for robust and scalable human robotics interactions in a heterogeneous intelligent architecture. This integrated environment will allow for human robotics interaction with a large number of robotics agents through a virtual reality environment. A number of industry level applications from the manufacturing industry, the gas and oil sector, as well as medical sector will be able to utilize the proposed architecture for the purpose of training, simulation, and testing. A set of novel frameworks will be developed to establish an efficient human-robotics collaborative environment in a fully immersive Virtual Reality (VR) world.
The proposed architecture may be utilized to create digital libraries for a set of industry applications for training and simulation purposes. The investigation results in developing computational and algorithmic frameworks for establishing such environments. Furthermore, original models and libraries for human-robotics interaction could be developed based on the proposed algorithms to devise efficient infrastructures for a targeted educational or industrial application.

Approved and Pending Funding Sources

  • NSF 12-580, Information and Intelligent Systems (IIS)-Human Centered Computing ($470,073), Currently Under Review.
  • Alcoa Foundations ($45,000), Digital Simulation and Robotics for Engineering, 2011-2013.
  • UHV Internal Research Award ($6,000), Detection of Patterns for Thread Avoidance with Multiple Robotic Agents – A Unified Framework, 2012-2013.

Publications

  • K. Akintola, A. Tavakkoli, "A Novel Gait Recognition System Based on Hidden Markov Models", in the Proceedings of the 8th International Symposium on Visual Computing, Rethymnon, Greece, July 2012.
  • A. Kolawole, A. Tavakkoli, "Robust Foreground Detection in Videos using Adaptive Color Histogram Thresholding and Shadow Removal", in the proceedings of the International Symposium Visual Computing, Las Vegas, (NV), September 2011.

Efficient Video Surveillance

In this project a larger number of individuals from UHV faculty, staff and students, were asked to observe surveillance videos and detect events in them. The objective of this study is to evaluate the performance of human operators in visual surveillance tasks. We developed and impolemented a virtual factory-like environment in an advanced game engine called Unreal Engine.
A software tool is developed that helps detect potential events in video surveillance images. Upon the detection of potential events, the area in which the event may occur may be highlighted to direct the attention of human observers. The performance improvement achieved by employing such tools will be investigated throughout this study.
We will further investigate the limitations of human visual perception in observing simultaneous images that contain many visual stimuli. This will lead us to develop more efficient tools and techniques. These algorithms will further enhance human visual surveillance experience by addressing cognitive and visual challenges human operators face in such applications.

Funding Sources

  • UHV Junior Faculty Research Award ($10,000), Evaluating the Performance Improvement of Visual Surveillance Tasks by Employing Intent Recognition, 2010-2011.

Publications

  • A. Tavakkoli, D. Loffredo, “Can Human Visual Surveillance be Improved with Intent Recognition?”, Journal of Systemics, Cybernetics, and Informatics, Vol. 11, No. 1, pages 43-50, January 2013.
  • A. Tavakkoli, Donald Loffredo, “Improving Human Video Surveillance Performance with Intent Recognition”, International Journal of Human-Computer Studies, (Under Review).
  • A. Tavakkoli, D. Loffredo, "Efficient Video Surveillance with Intent Recognition", in the proceedings of the International Symposium on Models and Modeling Methodologies in Science and Engineering, Orlando, (FL), July 2011. (Best Session Paper/Presentation)
Photo from LOTRO.com

Game Cultures – Virtual Life in MMORPGs

A large number of participants have been recruited to participate in a research study to help our group investigate how cultures and language form and play a role in the age of digital media and virtual worlds. Participants in this research study are asked to play in a Massively Multiplayer Online Role Playing Game (MMORPG). Our research groups is analyzing the ways in which human players communicate with other players or Non Player Characters (NPCs), perform certain tasks and in general socialize in a virtual environment.
The study investigates the hypothesis that players in Massively Multiplayer Online Role-Playing Games (MMORPGs) communicate with other players not only to complete tasks set by the game design but to socially construct virtual communities with identifiable subcultures. The participants' interactions in the game will be qualitatively analyzed to elicit the dynamics of the virtual subcultures that players have formed. Such discourse analysis is a widely accepted method for ethnographic and anthropological research.

Publications and Funding

This research is currently undergoing. A list of publications and proposals for funding will be provided shortly.

Interactive Level Desgin in Mobile Education

The core concept of the project is a gameplay environment involving two players that have full control over creation and modification of levels. This level design mechanism was implemented in an iOS-based game in the area of genetics and based on an existing written assignment. The game includes support for both instructors, who have the ability to create and post assignments and students, who can take the assignments. Two trials of the iOS application consisted of in-class testing of 21 students.
Students first took the original paper assignment, followed by the iOS version. Start times, end times, and grades were recorded for both versions. A comprehensive study of the grades and times for the iOS version of the assignment versus the paper version was conducted. Our objective is to demonstrate that the inherent differences between the delivery of educational materiasl using mobile computing devices to traditional pen-and-paper methods.
We are interested in investigating the infrastructre and architectures to develop strong delivery mechanisms needed to ensure student learning outcomes and completion of the educational material will not be adversly affected.

Publications and Funding

This research is currently undergoing. A list of publications and proposals for funding will be provided shortly.
This page has been created and maintained by Alireza Tavakkoli. Last Modified January 2013©.