CS&E Profile: Ioannis Karamouzas

Motion planning, Crowd simulation, Computer Animation, Robotics

Ph.D. 2012, Computer Science, Utrecht University, The Netherlands

M.Sc. 2005, Computer Science, The University of Manchester, U.K.

B.Sc. 2004, Applied Informatics, University of Macedonia, Greece

About

Ioannis Karamouzas joined the department of Computer Science and Engineering in November of 2012. Before coming to Minnesota, he obtained his Ph.D. from Utrecht University, in the Netherlands, where he worked under the supervision of Mark Overmars and A. Frank van der Stappen. Ioannis serves as a regular reviewer for a number of international conferences and journals. His research focuses on the development of motion planning algorithms for autonomous virtual humans, robots and virtual crowds.

Research

My main research interests lie in the areas of human motion planning and robotics. Regarding the modeling of human motion, I am particularly interested in first studying how humans behave and interact in real life and then trying to capture their dynamics in a simulation. To this end, my current focus is on analyzing large amount of data involving interactions of pedestrians in public spaces (e.g. shopping streets, malls, parks) and in controlled motion capture experiments. Such analysis will allow us to objectively evaluate existing crowd simulation methods and gain more insight into their relative strengths and weaknesses. In addition, it will help us develop new models of crowd motion that are able to overcome the shortcomings observed in existing techniques. Better crowd models can have a broad impact on our society in many ways. Applications areas span a wide range, from the evaluation of emergency escape routes to the planning and design process of public pedestrian facilities, and from training first responders to simulating compelling crowds in computer games and movies.

In robotics, my research mainly focuses on multi-robot planning and coordination (e.g., collision avoidance, path planning in dynamic environments, formation control for multi-robot teams) and on human-robot interaction (e.g., design of socially intelligent robots that move and interact in a life-like manner). I am also interested in blending techniques from other research fields, such as computer animation and machine learning, with motion planning algorithms to address challenging problems that lie at the intersection of these areas.