FEM-Based N-Net Control of a Flexible Dexterous Finger

Abstract Owing to their light weight, compactness, and good adaptability, soft robots are promising for applications with safe human-machine interactions. Compared with traditional rigid robots, the distributed actuation of soft robots with soft bodies leads to complex nonlinear deformation, making it challenging to model its dynamics. In this paper, a Finite Element Method (FEM)-based fast neural network open-loop control method is presented to control the flexible dexterous hands in their task-space using the finite element method. The main idea is to use a simple neural network to find the mapping between the end-effector control values and the displacement. In this way, the control of complex structures can be simplified. This model is especially suitable for complicated braking structures that cannot be accurately sensed. Finally, the effectiveness of the proposed method is verified by experiments, and the results are convincing.