An Argument for Motion Intelligence

[embed]

Our demos are often designed to show how intelligent simulation can synthesize even the most complex physical skills — from creating digital Basketball players to simulated stunt men. However, our most recent experiment focuses on simulating something more subtle: body language. Specifically, argumentative gesturing in a multi-character scenario.

Each of the characters in the video above is trained using five reference motions (i.e. motion data clips used to teach simulated characters motor skills); there are three separate argument reference motions, each about half a minute long, as well as two getting up motions, triggered after a character falls.

With only a few learned skills, the character controllers used in this demo are relatively simple, as are the mechanics of the simulation. When the simulation is played, each character transitions between the three argumentative gestures at random, creating the appearance of a lively debate. When the characters experience external force from a user, the environment, or another character, they fall and then promptly get up. The primary goal is to continue arguing.

This experiment is in its early stages, but nonetheless yields rich results. What excites us about this demo is the ability to transform a small number of motion clips into an interactive, dynamic scenario that — as paired down as it is — can excite the imagination and inspire open-ended entertainment.

*Read more stories like this on the DeepMotion Blog.*