Rapid Design and Fabrication of Body Conformable Surfaces with Kirigami Cutting and Machine Learning

TL;DR

This work tackles the challenge of mechanical mismatch in wearable skin interfaces by integrating Kirigami patterning with a data-driven surrogate model to enable rapid, personalized knee patches. It combines 3D knee scans to capture in-plane skin deformation, a FEA-calibrated Gaussian Process surrogate to map kirigami cuts to deformation, and CMA-ES optimization for inverse design, culminating in laser-cut patches fabricated within about 6.5 hours. The approach yields patches that conform to over 75% of the knee surface across three subjects and demonstrates a protective knee cap with integrated foam, illustrating practical applicability. The framework is generalizable to other deforming body surfaces and offers a scalable route to patient-specific wearables, protective gear, and body-conformable electronics.

Abstract

By integrating the principles of kirigami cutting and data-driven modeling, this study aims to develop a personalized, rapid, and low-cost design and fabrication pipeline for creating body-conformable surfaces around the knee joint. The process begins with 3D scanning of the anterior knee surface of human subjects, followed by extracting the corresponding skin deformation between two joint angles in terms of longitudinal strain and Poisson's ratio. In parallel, a machine learning model is constructed using extensive simulation data from experimentally calibrated finite element analysis. This model employs Gaussian Process (GP) regression to relate kirigami cut lengths to the resulting longitudinal strain and Poisson's ratio. With an R2 score of 0.996, GP regression outperforms other models in predicting kirigami's large deformations. Finally, an inverse design approach based on the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) is used to generate kirigami patch designs that replicate the in-plane skin deformation observed from the knee scans. This pipeline was applied to three human subjects, and the resulting kirigami knee patches were fabricated using rapid laser cutting, requiring only a business day from knee scanning to kirigami patch delivery. The low-cost, personalized kirigami patches successfully conformed to over 75 percent of the skin area across all subjects, establishing a foundation for a wide range of wearable devices. The study demonstrates this potential through an impact-resistant kirigami foam patch, which not only conforms to dynamic knee motion but also provides joint protection against impact. Finally, the proposed design and fabrication framework is generalizable and can be extended to other deforming body surfaces, enabling the creation of personalized wearables such as protective gear, breathable adhesives, and body-conformable electronics.

Figures (6)

• Figure 1: Flowchart of the rapid design and fabrication pipeline of personalized Kirigami knee patch. Note that the second step --- measure skin deformation --- is completed manually in this study due to the need to manually transfer data between different software. Developing an integrated and automated software package could further speed up this process.
• Figure 2: Scanning and measuring the subjects' knee surface deformation. (a) Reconstructed knee surfaces of the 3 subjects (i.e., S1, S2, and S3), using the 3D scanning data. (b) Heatmaps showing the longitudinal, lateral strain, and Poisson's ratio distribution from the 0° to 90° knee flex. The unit cells with the maximum and minimum values in each heatmap are labeled, further highlighting the variety between different subjects. Note that the length variables (e.g., $\widehat{op}$ and $\widehat{q'r'}$) in the strain calculations are geodesic distances on the 3D knee surface.
• Figure 3: Populating the design space of Kirigami with experimentally-calibrated finite element analysis (FEA). (a) Comparing the Kirigami sheets' deformation pattern between FEA and experiment (with top end displacement U=70mm). (b) Comparing the Poisson's ratio between FEA and experiment. Here, the solid lines are averaged test data, and the shaded bands are standard deviations from the readings of 12 unit cells in the middle of the samples. The dashed lines are FEA predictions. (c) Populated design space, where each marker corresponds to a unique combination of Kirigami cut design and longitudinal strain, and the color of the marker represents the corresponding Poisson's ratio. (d) A close-up view of the Kirigami design variables. (e) A subsection of this design space, showing a few slices corresponding to U=10, 50, 120, 250mm. (f) Detailed FEA simulation outcomes, where [i-iv] and [1-4] highlight two Kirigami designs.
• Figure 4: Elements of the inverse design methodology. (a) The input and output architecture of the surrogate modeling. (b) L2-Normalization for design candidate selection. In the design space plot on the left, we selected 6 design candidates on the optimal curve (i.e., P1-P5 and the optimal candidate). Then their longitudinal strain vs. Poisson's ratio responses are plotted on the right. (c) The flow chart of the optimal Kirigami design: The initial data was sampled randomly at first, as shown by generation 1. After a few iterations, it converges to our optimal solution at the 26th iteration.
• Figure 5: The final, personalized Kirigami patches for the three subjects. (left) In this plot, each star-shaped marker represents the longitudinal strain--Poisson's ratio combination of a grid cell on the subject's knee surface. They are the same data from the heatmaps in Figure \ref{['fig: scan']}(b), but plotted differently. Meanwhile, each thin curve in this plot is the longitudinal strain--Poisson's ratio response corresponding to a unique Kirigami cut design. These curves are generated by the surrogate model, which is based on the FEA simulation data in Figure \ref{['fig: FEA']}(c). Essentially, the optimization procedure laid out in Figure \ref{['fig: optimization']}(c) finds the curve that best matches each star-shaped marker. (right) The final Kirigami patch designs. The images in the left column are the reconstructed 3D scanning data from the subjects wearing their Kirigami patches, and the images in the right column are the corresponding Kirigami designs.