Note: This information was taken from my old UU research page. I will update it soon.
Computer games are getting bigger and more dynamic, and they feature increasingly large crowds of virtual characters who perform increasingly complex tasks. In other areas, there is an increasing desire to simulate real-life crowded events before they take place, e.g. to predict and prevent safety problems.
How should we represent a virtual world such that large crowds of characters can find and traverse visually convincing paths in real-time?
Our research aims to create a single framework that can answer navigation-related questions for a wide range of virtual environments and characters, while requiring as little extra manual work as possible.
For a recent summary of our research, please read our 2015 overview paper.