Learning quasi-static 3D models of markerless deformable linear objects for bimanual robotic manipulation

Abstract

The robotic manipulation of Deformable Linear Objects (DLOs) is a vital and challenging task that is important in many practical applications. Classical model-based approaches to this problem require an accurate model to capture how robot motions affect the deformation of the DLO. Nowadays, data-driven models offer the best trade-off between quality and computation time. This paper analyzes several learning-based 3D models of the DLO and proposes a new one based on the Transformer architecture that achieves the lowest median relative prediction error among baselines. Moreover, we introduce a zero-motion, physics-consistency data augmentation technique that improves the prediction performance of almost all considered DLO data-driven models. Thanks to this technique, even a simple Multilayer Perceptron (MLP) achieves close to state-of-the-art performance while being significantly faster to evaluate. We further show that a simple length-scaling heuristic is particularly effective for the Transformer when transferring between the tested cable lengths. In the experiments, we compare the performance of the learning-based 3D models of the DLO on several challenging datasets quantitatively and demonstrate their applicability in the task of shaping a DLO.

The authors' abstract, as published at the source. Expert Systems with Applications, 2026 · DOI ↗