A Face-Encoding Grammar for the Generation of Tetrahedral-Mesh Soft Bodies

Many of the most profound works of artificial life have emerged through the composition of physical simulation and generative representations. And yet, while physics engines are becoming more realistic, and generative representations are growing more powerful, they are still predominantly used to simulate rigid objects. The natural world and its organisms are, by contrast, soft, and full of much more interesting (and complex) interactions than those which can be faithfully reproduced by rigid body dynamics. In this work we describe and implement a grammatical encoding capable of generating large, complex, and multi-resolution soft structures which can be natively simulated by the state-of-the-art hardwareaccelerated physics engines. The structures generated by the encoding exhibit all the benefits (structural modularity, l argescale co-ordinated change) of more conventional rigid-body generative encodings.