Future Game Animation (Beyond Ragdolls)

An example of emergent evasive behavior generated by our method.

Project Overview

The project has three main goals:

1. Reduce the time and skill needed in animating by developing AI (optimization and machine learning) that synthesizes physically valid animation based on goals and constraints. For example, one defines the starting pose and kicking pose of a jumpkick, and the AI generates the needed steps, jump and landing.

2. Further reduce animation cost by optimizing animation interfaces using recent 3d input devices like Kinect, PlayStation Move and Leap Motion

3. Explore what kinds of novel game mechanics can be designed and implemented using a real-time version of the AI.

Here are some videos of the results:

The project is headed by prof. Perttu Hämäläinen, and funded by Tekes (Finnish Funding Agency for Innovation) and 7 game and animation companies for the period June 2013-Dec 2015. Contact prof. Hämäläinen if you are interested in doing a Master's thesis or pursuing a doctoral degree in the project or a related field.


Kytö, M., Dhinakaran, K., Martikainen, A., Hämäläinen, P. (2016) Improving 3D Character Posing with a Gestural Interface, To appear in IEEE Computer Graphics and Applications.

Naderi, K., Rajamäki, J., Hämäläinen, P. (2015) RT-RRT*: a real-time path planning algorithm based on RRT*. In Proceedings of the 8th ACM SIGGRAPH Conference on Motion in Games (MIG '15). PDF, Video

Hämäläinen, P., Rajamäki, J., Liu, C.K. (2015) Online Control of Simulated Humanoids Using Particle Belief Propagation. To appear in Proc. SIGGRAPH 2015. PDF, Video 1, Video 2, Source code

Hämäläinen, P., Eriksson, S., Tanskanen, E., Kyrki, V., Lehtinen, J. (2014) Online Motion Synthesis Using Sequential Monte Carlo. Proc. SIGGRAPH 2014. PDF, Video, Source code (no Unity examples yet, the .zip only includes the sampler and a basic inverted pendulum example.)

Perttu Hämäläinen, May 7 2015