Learning to self-fold at a bifurcation.

Area of Science:

• Materials Science
• Mechanical Engineering
• Soft Matter Physics

Background:

• Disordered mechanical systems deform along complex pathways, often involving bifurcation points.
• Designing these pathways computationally requires advanced algorithms and precise control over geometry and material properties.

Purpose of the Study:

• To explore a physical training framework for controlling folding pathway topology in disordered sheets.
• To investigate how changes in crease stiffnesses, induced by prior folding, can guide material behavior.
• To study the robustness of this training for different learning rules.

Main Methods:

• Investigated a physical training approach where material properties adapt based on deformation history.
• Studied various "learning rules" quantifying how local strain affects local folding stiffness.
• Experimentally demonstrated the concept using sheets with epoxy-filled creases whose stiffness changes post-folding.

Main Results:

• Demonstrated that plasticity in materials enables learning of nonlinear behaviors through deformation history.
• Showcased a robust method for guiding folding pathways in disordered sheets.
• Validated the effectiveness of the physical training framework.

Conclusions:

• Specific forms of material plasticity can be harnessed for adaptive control of mechanical systems.
• Physical training offers a viable alternative to computational design for achieving desired deformation pathways.
• This approach provides robust control over complex material behaviors based on their history.