Celebi's Choice: Causality-Guided Skill Optimisation for Granular Manipulation via Differentiable Simulation

Minglun Wei, Xintong Yang, Junyu Yan, Yu-Kun Lai, Ze Ji

Keywords: deformable object manipulation, differentiable physics simulation, robot learning, causal inference

TL;DR: We present Celebi, a method that integrates differentiable physics simulation with causal inference to achieve stable and efficient optimisation for robotic excavation and levelling tasks.

Abstract: Robotic manipulation of deformable objects is inherently challenging due to their complex, nonlinear, and often unpredictable dynamics under force. Granular materials such as soil present an even greater challenge, as they exhibit both particulate and continuum behaviours that complicate modelling and control. As a representative case, soil manipulation is central to automated container-based agriculture, where robots must perform precise excavation and levelling. However, physical trials are costly, and standard reinforcement learning (RL) methods often suffer from sample inefficiency and unstable dynamics in such settings. To address these challenges, we propose Celebi, a causality-enhanced optimisation method that integrates differentiable physics simulation with adaptive step-size adjustments based on causal inference. To enable gradient-based optimisation, we construct a differentiable simulation environment for granular material interactions. We further define skill parameters with a differentiable mapping to end-effector motions, facilitating efficient trajectory optimisation. By modelling causal effects between task-relevant features extracted from point cloud observations and skill parameters, Celebi selectively adjusts update step sizes to enhance optimisation stability and convergence efficiency. Experiments in both simulated and real-world environments validate Celebi’s effectiveness, demonstrating robust and reliable performance in robotic excavation and levelling tasks.

Submission Number: 7