Geometrically Plausible Object Pose Refinement using Differentiable Simulation
This work addresses the critical issue of physical plausibility in object pose estimation for robust in-hand manipulation in robotics, representing a strong specific gain rather than a foundational advancement.
The paper tackles the problem of geometrically infeasible object pose estimates in dexterous manipulation by proposing a multi-modal refinement approach that combines differentiable simulation, rendering, and visuo-tactile sensing to optimize for spatial accuracy and physical consistency. The result is a reduction in intersection volume error by 73% with accurate initial estimates and over 87% under high uncertainty, outperforming ICP-based baselines while also reducing translation and orientation errors.
State-of-the-art object pose estimation methods are prone to generating geometrically infeasible pose hypotheses. This problem is prevalent in dexterous manipulation, where estimated poses often intersect with the robotic hand or are not lying on a support surface. We propose a multi-modal pose refinement approach that combines differentiable physics simulation, differentiable rendering and visuo-tactile sensing to optimize object poses for both spatial accuracy and physical consistency. Simulated experiments show that our approach reduces the intersection volume error between the object and robotic hand by 73\% when the initial estimate is accurate and by over 87\% under high initial uncertainty, significantly outperforming standard ICP-based baselines. Furthermore, the improvement in geometric plausibility is accompanied by a concurrent reduction in translation and orientation errors. Achieving pose estimation that is grounded in physical reality while remaining faithful to multi-modal sensor inputs is a critical step toward robust in-hand manipulation.