Physical Modeling and Offline Parameter Identification-Based Method for Shape Control of Deformable Linear Objects

Like the mathematic model which plays great role in the rigid body motion controller design, accurate models also are crucial for the manipulation of deformable linear objects (DLOs). However, the modelling process for DLOs has bigger challenge than for rigid bodies, due to difficulties within the complex physical property and high-dimensional state space description, as well as non-ideal condition involving disturbances. In this paper, the problem of obtaining accurate models for DLOs is investigated by introducing a deformable body parameter identification method based on a modified mass-spring system (MSS) model and a Radial Basis Function Network (RBFN). The MSS model of DLOs is firstly developed at the first step. Then the parameter set of the model is fine-tunning by applying RBFN offline. Finally, an ADRC controller is designed for the shape control task. Simulations experiments show that the proposed identification method gives more accurate model, which simplified the controller designer and improve the manipulation performance of DLOs. Furthermore, the method is simple in structure and easily applied to various DLOs.