Search papers, labs, and topics across Lattice.
This paper introduces a modular framework for pose estimation that decomposes the process into failure detection, error attribution, and targeted recovery. The framework identifies specific error sources affecting pose estimation and applies tailored mitigation strategies, improving robustness and efficiency. Experiments using Iterative Closest Point (ICP) in real-world robotic grasping demonstrate that this approach achieves competitive performance with foundation models, while using a simpler and faster pose estimator.
Decomposing pose estimation into error attribution and targeted mitigation allows simple algorithms like ICP to rival the robustness of complex foundation models, but with significantly less computation.
Robust estimation of object poses in robotic manipulation is often addressed using foundational general estimators, that aim to handle diverse error sources naively within a single model. Still, they struggle due to environmental uncertainties, while requiring long inference times and heavy computation. In contrast, we propose a modular, uncertainty-aware framework that attributes pose estimation errors to specific error sources and applies targeted mitigation strategies only when necessary. Instantiated with Iterative Closest Point (ICP) as a simple and lightweight pose estimator, we leverage our framework for real-world robotic grasping tasks. By decomposing pose estimation into failure detection, error attribution, and targeted recovery, we significantly improve the robustness of ICP and achieve competitive performance compared to foundation models, while relying on a substantially simpler and faster pose estimator.
