Traditional midsoles use uniform foam density, but our feet don’t load uniformly. This project explores computational lattice structures for performance footwear that integrate lattice geometry, data-driven stiffness gradients, and human-centered performance tuning.
‍
By evaluating six lattice typologies and two material gradient approaches, the work reveals how micro-scale lattice morphology influences macro-scale comfort, stability, and weight —informing design decisions for different athletic applications.
‍

Two gradient approaches were explored:
1) Design intuition meets biomechanics: beam thickness increases with expected load performance
- Toe 0.7mm (Flexible and Printable)
- Forefoot 1mm (Propulsion)
- Arch 1.3mm (Support)
- Heel 1.6mm (Impact absorption)

2) Pressure data informed: the more load, the thicker of the beam

A uniform-radius lattice was used as a baseline to evaluate how topology, voxel resolution, and beam radius affect midsole performance. Millipede and Opossum plugins evaluated deflection and weight across all candidates.
‍
Key findings:
1) Re-entrant lattices consistently achieve the best comfort–weight tradeoff.
2) 0.8–1.4 mm beam radii provide stable, printable structures.
3) Lower X/Z voxel densities reduce over-stiff regions and improve compliance.
‍
These results establish the parameter range used for the subsequent gradient-thickness studies.
‍

The optimal parameters were then applied to the lattice structures that using gradient beam radii
‍
Key findings:
1) Regional gradients:
- Forefoot-biased & All-rounder = softest/lightest;
- Heel-biased = stiffest/heaviest;
‍
2) Data-driven gradients: Always achieve better comfort–weight balance than intuition or uniform radius.
‍
3) Structure effect: Voxel lattices outperform Voronoi for cushioning and targeted stiffness.
‍

This project demonstrates how computational lattice exploration and FEA validation can guide design decisions prior to physical iterations—though real-world testing remains essential to verify TPU behavior, layer adhesion in complex geometries, and comfort metrics that simulation cannot capture.
‍
Attempts at physical prototypes revealed manufacturing realities: while technically printable via SLS or MJF (and, to a lesser extent, SLA/FDM), single-unit production of fine lattice structures (min-2mm beam thickness) proved cost-prohibitive due to slow print speeds, complex support removal, and inability to amortize setup costs.
‍
This gap between computational design and accessible manufacturing could be addressed through maker space partnerships or in-house printing, as well as design-for-manufacturability refinements (e.g. optimized print path and support structures).
‍
As a result, this project emphasizes data-driven design, digital validation and parametric logic, with physical prototyping deferred to contexts where resources or production volumes make lattice printing economically viable.
‍