Compounded impact of pose-estimation and reconstruction errors on grasping success

Determine how errors arising from 6D object pose estimation and 3D object model reconstruction compound and propagate to influence the success of downstream robotic manipulation tasks, specifically grasping.

Background

The paper highlights that 6D object pose estimation and 3D reconstruction are typically evaluated with separate, task-agnostic geometric metrics (e.g., ADD for pose estimation and Chamfer distance for reconstruction). This separation leaves a gap in understanding how combined perception errors affect physical interaction.

To address this gap, the authors introduce a physics-based benchmark that evaluates the functional efficacy of pose estimation and reconstructed meshes in robotic grasping. The stated uncertainty motivates their end-to-end evaluation of how pose and geometry errors manifest in grasp outcomes.

References

However, this decoupled evaluation creates a significant gap, and it is unclear how errors from pose estimation and geometric reconstruction compound and propagate to affect the success of downstream manipulation tasks like grasping.

— Benchmarking the Effects of Object Pose Estimation and Reconstruction on Robotic Grasping Success  (2602.17101 - Burde et al., 19 Feb 2026) in Introduction

However, the synergy basis is extracted from hand postures rather than from the objects being grasped: the same basis applies regardless of object shape, which leaves open how synergies cope with errors in the object description.

— Grasping by interconnection: robust closing motions from coarse object templates  (2609.19228 - Vanderheyden et al., 16 Sep 2026) in Section 2.3, Human-inspired grasp synthesis