IST Lunch Bunch

I will present my recent research on modeling
large-deformation elastodynamics using the Finite Element Method.
While FEM and PDEs of elasticity are well-understood, they cannot
easily match the observations of real complex elastodynamic systems
such as the human body, plants and virtual characters. Common
obstacles in practice include unknown nonlinear materials,
inconsistent meshes that self-collide already in the neutral
configuration, and stringent computational speed requirements. I will
present techniques to set nonlinear material properties to match
real-world observations. I will also discuss how to simplify complex
elastodynamics using model reduction. Finally, I will discuss our
recent technique that uses geometric and algebraic topology to
generate tetrahedral meshes that mesh the volume occupied by the
object despite self-collisions in the input mesh.