Machine Leaning-Based Method for Kinematics Parameters Identification of Twin-Pivot Cable-Driven Continuum Robots

Zheshuai Yang,Yu Lan,Dong Yang,Laihao Yang,Yu Sun,Xuefeng Chen
DOI: https://doi.org/10.1109/icsmd57530.2022.10058226
2022-01-01
Abstract:Twin-pivot cable-driven continuum robots have been widely employed in complex and unstructured scenarios, benefiting from their compliant and torsion-resistant properties. However, significant numbers of sections lead to a challenging problem in solving the inverse kinematics/statics. And the conventional Jacobian-based method suffers from complex computation, which is time-consuming. This paper proposed a machine learning-based method to identify kinematics parameters. First and foremost, an accurate static model is established. And then, the multi-layer perceptron (MLP) is employed to learn the inverse statics based on the large numbers of samples generated by the proposed static model. Finally, the verification of the proposed models is performed. The experimental results indicate that the mean absolute percentage error of cable lengths is within 2.2%, and the tip position error is within 1.6%.
What problem does this paper attempt to address?