KineFormer: Solving the Inverse Modeling Problem of Soft Robots Using Transformers

Soft robotic manipulators provide numerous advantages over conventional rigid manipulators in fragile environments such as the marine environment. However, developing analytic inverse models necessary for shape, motion, and force control of such robots remains a challenging problem. As an alternative to analytic models, numerical models can be learned using powerful machine learning methods. In this paper, the Kinematics Transformer is proposed for developing accurate and precise inverse kinematic models of soft robotic arms. The proposed method recasts the inverse kinematics problem as a sequential prediction problem and is based on the transformer architecture. Numerical simulations reveal that the proposed method can effectively be used in controlling a soft arm. Benchmark studies also reveal that the proposed method has better accuracy and precision compared to the baseline feed-forward neural network.