Harnessing Deep Learning of Point Clouds for Morphology Mimicking of Universal 3D Shape‐Morphing Devices

Shape‐morphing devices, a crucial branch in soft robotics, hold significant application value in areas like human–machine interfaces, biomimetic robotics, and tools for biological systems. To achieve 3D programmable shape morphing (PSM), the deployment of array‐based actuators is essential. However, a critical knowledge gap in 3D PSM is controlling the complex systems formed by these soft actuator arrays to mimic the morphology of the target shapes. This study, for the first time, represents the configuration of shape‐morphing devices using point cloud data and employing deep learning to map these configurations to control inputs. Shape Morphing Net (SMNet), a method that realizes the regression from point cloud to high‐dimensional control input vectors, is proposed. It has been applied to 3D PSM devices with three different actuator mechanisms, demonstrating its universal applicability to inversely reproduce the target shapes. Further, applied to previous 2D PSM devices, SMNet significantly enhances control precision from 82.23% to 97.68%. In the demonstrations of morphology mimicking, 3D PSM devices successfully replicate arbitrary target shapes obtained either through 3D scanning of physical objects or via 3D modeling software. The results show that within the deformable range of 3D PSM devices, accurate reproduction of the desired shapes is achievable.