Research in virtual human simulation has seen a dramatic rise in recent years with widespread applicability in games, visual effects, education and urban simulations. The next generation of these applications demand functional, purposeful, heterogeneous autonomous virtual avatars, that exhibit rich, believable interactions with their environment and other characters with the far-reaching goal of complete immersion for the end users.
This course will focus on simulating and animating autonomous virtual humans for real-time applications. Students will be introduced to techniques for representing, controlling, animating, and authoring autonomous virtual humans and will develop their very own virtual human simulator by applying techniques from artificial intelligence, motion planning, physical simulation, and machine learning to character animation. The course will culminate with a final project where students will identify and address shortcomings in existing solutions.
The final deliverable will be a technical report (of publication quality) and a 5 minute video demonstration of the developed application, supplemented with regular in-class updates and demonstrations.
- Discrete Representations of 3D Environments for Navigation
- Anytime Dynamic Planning Algorithms for Motion Planning
- Multi-agent collision avoidance
- Agent-based techniques for simulating crowds
- Sound propagation, perception, and localization for autonomous agents
- Data-driven and procedural solutions for character animation
- Computer-Assisted Authoring Tools for Interactive Narratives
- Statistical Analysis and Optimization of Multi-Agent Simulations
- Regular mini-projects that will lead towards developing a complete virtual human simulation framework.
- Final project where students will work as individuals or in teams of 2 to identify and address shortcomings in current state-of-the-art solutions for character animation.
Exams and Grading
Final project deliverable: Technical report of publication-level quality describing final project, 5 minute video demonstration, and regular in-class demonstrations.