A framework for histograms of object pose distributions for robotics

Project Overview

This page is a collection of the research performed in the project "A framework for histograms of object pose distributions for robotics". This project presents a comprehensive framework for representing and manipulating object pose distributions using histogram-based methods. The framework is designed to address the challenges of pose estimation and tracking in robotic manipulation and perception tasks.

By utilizing histogram representations, the framework provides efficient storage and retrieval of probabilistic pose information, enabling robots to make informed decisions in uncertain environments.

Current Status

In the current stage of the project, we have developed a method for estimating SO(2) revelution and reflection uncertainty in point clouds, and applied it in bin picking. We are currently working on extending the usage of this method to a modular bin-picking framework that jointly handles both grasping and object pose uncertainties using interchangeable modules.