Adhesion & AI-driven design

Bio-inspired dry adhesion

Dry adhesion is achieved with micro-patterned polymer interfaces that mimic the reversible grip of insect, spider, and gecko toe pads. Adhesion arises from van der Waals forces amplified by hairy surface structures, enabling strong, repeatable attachment without glue. Our group develops mathematical and computational models to optimize the geometry, chemistry, and mechanics of these biomimetic adhesives.

AI-driven design

AI-guided design of fibrillar adhesives builds on our mathematical and computational models by introducing machine learning to discover optimal geometries. Instead of exploring shapes and patterns through trial-and-error, we train AI algorithms on large datasets of adhesion performance to identify the subtle interplay of fiber spacing, aspect ratio, and orientation. This approach enables us to predict and program the “perfect” design of dry adhesives—structures that maximize grip when needed, yet release effortlessly. By combining biomechanics with AI, we accelerate innovation in biomimetic adhesion, bridging natural principles and engineered solutions.