How it Works
Dynamic Motion Synthesis is based on Oxford University's research on the control of human and animal body motions. In essence, we build a physical, biomechanically-realistic model of a character (e.g. a human or a dinosaur), implant an appropriate brain structure (usually a neural network), and use optimisation techniques (such as artificial evolution) to create the desired adaptive behaviour.
The following video demonstrates how this works in practice. The example shown is a simple biped which learns how to walk using artificial evolution.
The process starts with random walkers, none of which can walk properly. The best ones (those that make at least one step without falling over) are allowed to produce offspring, which are again selected according to how far they walk. This selection is repeated over a number of generations. At the end of process the biped can walk without falling over.
