{"ID":3210762,"CreatedAt":"2026-06-06T23:32:07.996611237Z","UpdatedAt":"2026-06-06T23:32:07.996611237Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2410.09740","arxiv_id":"2410.09740","title":"Gaussian Splatting Visual MPC for Granular Media Manipulation","abstract":"Recent advancements in learned 3D representations have enabled significant progress in solving complex robotic manipulation tasks, particularly for rigid-body objects. However, manipulating granular materials such as beans, nuts, and rice, remains challenging due to the intricate physics of particle interactions, high-dimensional and partially observable state, inability to visually track individual particles in a pile, and the computational demands of accurate dynamics prediction. Current deep latent dynamics models often struggle to generalize in granular material manipulation due to a lack of inductive biases. In this work, we propose a novel approach that learns a visual dynamics model over Gaussian splatting representations of scenes and leverages this model for manipulating granular media via Model-Predictive Control. Our method enables efficient optimization for complex manipulation tasks on piles of granular media. We evaluate our approach in both simulated and real-world settings, demonstrating its ability to solve unseen planning tasks and generalize to new environments in a zero-shot transfer. We also show significant prediction and manipulation performance improvements compared to existing granular media manipulation methods.","url_abs":"https://arxiv.org/abs/2410.09740v3","url_pdf":"https://arxiv.org/pdf/2410.09740v3","authors":"Wei-Cheng Tseng, Ellina Zhang, Krishna Murthy Jatavallabhula, Florian Shkurti","published":"2024-10-13T06:18:53Z","has_code":false}