Methodology for the Design and Control of a Soft Finger Based on the Mullins Effect and Material Training

The advancement of soft multi-digital grippers with both adaptive grasping and in-hand manipulation capabilities remains a challenging issue for the development of human-like dexterous manipulation. Despite four decades of research, the most advanced grippers remain encumbered by excessive complexity and a lack of robustness. The field of soft robotics presents a promising avenue for reducing the level of complexity and enhancing the safety of grasping and interaction with the environment. This work presents a methodology for the design and control of a soft finger, with the objective of ultimately developing a highly dexterous gripper. To address these challenges, it is essential to master the design, fabrication process, and behavior of the finger’s soft material. The iterative design and fabrication process requires a comprehensive understanding of the theoretical and experimental aspects, as detailed in this paper. Given that the finger is constructed from silicone, the proposed methodology and outcomes demonstrate the importance of accounting for the Mullins effect and conducting finger training prior to controlling the pneumatic soft finger. The proposed hard real-time control architecture guarantees the robustness of the analysis and control of the finger’s behavior, while also offering perspectives for coordinated multi-fingered manipulation.