Let's Handle It: Generalizable Manipulation of Articulated Objects

TL;DR

This paper introduces a framework for developing generalizable manipulation controllers that process raw point cloud data to perform household object manipulation, demonstrated on multiple skills within a robotics challenge.

Contribution

It presents a traditional robotics-based framework that integrates point cloud processing, trajectory planning, inverse kinematics, and behavior trees for manipulation tasks.

Findings

Successfully applied to four household manipulation skills

Demonstrates generalization across different objects

Achieved competitive performance in SAPIEN ManiSkill challenge

Abstract

In this project we present a framework for building generalizable manipulation controller policies that map from raw input point clouds and segmentation masks to joint velocities. We took a traditional robotics approach, using point cloud processing, end-effector trajectory calculation, inverse kinematics, closed-loop position controllers, and behavior trees. We demonstrate our framework on four manipulation skills on common household objects that comprise the SAPIEN ManiSkill Manipulation challenge.