Differentiable Stripe Patterns for Inverse Design of Structured Surfaces

Stripe patterns are ubiquitous in nature and everyday life. While the synthesis of these has been thoroughly studied literature, their potential to control mechanics structured materials remains largely unexplored. In this work, we introduce Differentiable Patterns---a computational approach for automated design physical surfaces with stripe-shaped bi-material distributions. Our method builds on work by Knöppel colleagues [2015] generating globally-continuous equally-spaced stripe patterns. To unlock full space, propose a gradient-based optimization tool automatically compute that best approximate macromechanical performance goals. Specifically, model combines solid shell finite elements XFEM accurate fully-differentiable modeling elastic surfaces. resolve non-uniqueness problems original method, furthermore robust formulation yields unique differentiable We combine components equilibrium state derivatives into an end-to-end pipeline enables inverse mechanical demonstrate our diverse set examples illustrate as space materials. simulation results experimentally validated prototypes.