A Unified Approach to Grown Structures
MIT Media Lab
We explore a computational approach and associated protocol, which emulates biological growth by developing complex geometries over multiple iterations. The general framework for the generation of grown structures utilizes a hybrid approach to the simulation of evolving interfaces. A geometric input representation—phenotype (e.g. a triangle mesh, a set of line segments, or a point cloud) is transformed into an intermediate representation—genotype. Data gathered from these three representations is then used to deform the initial geometric representation. Lastly, the deformed initial representation is topologically changed to react to the deformation of the object. This is done iteratively, such that results given for input representations are continuously deformed and refined. As the process repeats, the deformations aggregate into the growth of a coherent form. By altering the geometric genotype and phenotype, a broad variety of different structures can be ‘grown’. For an exhaustive presentation showcasing 15 unique derivations of this process, we refer to following video.
2016 Grown, Printed and Biologically Augmented