Evaluation of a Data-Driven Model for Deformable Linear Objects Manipulated by Two Robot Manipulators Using Experimental Data

Deformable objects such as cables, ropes, and cloth-like objects are more commonly used in our daily lives rather than rigid bodies. However, manipulating these deformable objects with a robot is challenging due to their high degrees of freedom and complex deformations, making it difficult to generate dynamic models using traditional analytical methods. Data-driven models have been employed to develop dynamic models for such deformable objects. However, to learn the models from experimental data, it is recommended to reduce the number of variables. In this paper, we established an experimental environment where deformable linear objects (DLOs) can be manipulated using a dual-arm robot. We evaluated the performance of the data-driven model using the positional information of the objects obtained in this setup. We achieved an average position error of 0.010 m for forward prediction and 0.073 m for long-term prediction. In future research, we intend to employ this data-driven model to manipulate cables into desired shapes.