Research Projects:
   
  1. Understanding and quantifying the impact of moving target defenses on computer networks.
  2. Human Robot Teams Informed by Human Performance Shaping Factors. 
       








      • Supervisor-based Human Robot Team Run-time Task Reassignment. [Youtube]
            



       

    1. Robot Communication using arm gestures. [Youtube]    
    This project is intrigued by the idea that a team of scout robots need to use only gestures to communicate with each other under specific battlefield conditions.






    Selective Projects:    
    1. Extended kalman filter SLAM. [Youtube] [Youku]           
               
    This project shows an implementation of extended kalman filter SLAM(simultaneous localization and mapping) based on ROS. It has been done in one of my CIS890 
    course with Dr.Gustafson. RANSAC algorithm has been implemented to extract point landmark from laser scan. Key SLAM processes implemented include state prediction acorrding to odometry, state correction based on measurements of landmarks and state updating with landmark association. A visualization component has been implemented to display the whole process in Rviz.


    Legend information can be found in video description.

    1. Shape-based Image Retrieval
       

    2. Voice Recognition, Color Tracking, Obstacle Avoidance by Lego Robot. [Youku, passcode: 555666]
              

       








    1. Microsoft Imagine Cup Project Hoshimi Demo [Youtube]
           
    Project Hoshimi is an artificial intelligence programming contest. Each game has two players and each player (a team) needs to design and implement their own AI strategies to build and control a team of heterogeneous robots (up to 40) to accomplish a sequence of missions to maximize their rewards as well as restrain and minimize opponent's score. Concrete missions include navigating to specific points to build needle robot, find out the closest location of mines and send robots to collect and transport minerals to needle robots, destroy bad factories. Different mission has different reward points. 
    A good implementation should properly consider how to program robots to find the best landing point, explore the map, plan the optimal paths, reconnaissance opponent robots' operation and launch attack or defence to win the game.