CS&E Profile: Stephen Guy

Motion planning, Computer Animation, Robotics, Complex Systems

Ph.D. 2012, Computer Science, University of North Carolina – Chapel Hill

M.S. 2009, Computer Science, University of North Carolina – Chapel Hill

B.S. 2006, Computer Engineering, University of Virginia

About

Stephen J. Guy is an assistant professor whose research focuses on interactive computer graphics (e.g., real-time crowd simulation, path planning, intelligent virtual characters) and multi-robot coordination (e.g., collision avoidance, sensor fusion, path planning under uncertainty). Stephen has served on the program committee for international conferences, his work on crowd simulation has been nominated for multiple best paper awards, and his work on distributed motion planning has been licensed for use in commercial software.

Research

My research focuses on studying complex systems with several independently moving actors. A common example of this is human crowds – each person in a crowd moves independently, but we see consistent patterns in crowds that emerge across a variety of situations. I study these types of systems with two themes in mind:

1)/Modeling/: Can we produce simulations that display the interesting behaviors seen in these systems?For example, can we produce models of human crowd motion that is believable enough for games and movies or even accurate enough to assist in architectural design and evacuation planning?

2)/Analysis/: What can our models tell us about real-world behavior and how can we transfer our knowledge from studying one domain to solve problems in a new area? For instance, we have applied our models of human motion planning to robots, and have shown how this enables them to naturally move around each other while avoiding any collisions.

This type of research builds strongly on many areas. For example, solving long term analysis goals such as real-time video monitoring of crowd safety involves bringing together aspects of computer vision, big-data, machine learning and AI to extend our models to be responsive to real-world data in real-time.

I am also interested in other areas that blend these same types of research fields such as machine-learning for robotics, and intelligent agents in virtual environments.