Predicting the trajectory of objects moving freely in a fluid is a fundamental problem with applications across the natural sciences: From molecules to intracellular processes across plants and animals to geophysics and vehicle fuel efficiency. Inversely designing an object to take a specific path is equally important, e.g., for targeted drug delivery, aerial leaflets, and drone design. Settling trajectories, however, are extremely difficult to predict because the body interacts with itself via the flow. These “fluid-structure interactions” impose a basic indeterminacy on the problem, which has limited progress on several fundamental questions. These include understanding the origin of symmetry in plant leaves and engineering settling objects to follow specific paths.

The intrinsic unpredictability has led to state-of-the-art experiments and theories focusing on individual trajectories of highly idealized objects (sphere, disk, strip) and trial/error design. Most real objects, however, are not ideal but contain defects. This renders current methods inadequate.

To break the stalemate, I will create a new kind of on-demand experiment with complete geometric freedom while delivering data orders of magnitude faster than current methods. An automatically fed laser cutter (or 3D printer) placed above a settling tank is at the system's core. One setup can map several 100 trajectories daily. We can thus obtain more data in a single day than included in a typical publication. Critically, we can measure and model how shape and defects impact the statistics of objects moving freely in fluids, thus moving beyond the classical deterministic interpretation.

To systematics the link between form and function, we map the performance of morphologies evolved via natural selection: this “Aerodynamics Genome” will be openly accessible. This is the first step towards abandoning our tradition of trial-and-error, and with this comes the potential for a new era of aerodynamics.