Autonomous object model acquisition with parallel jaw gripper

Granting robots the ability to identify and interact with any object in a cluttered, real-world environment requires solutions that tackle partial visibility and object novelty. Humans overcome this challenge with experience, which we emulate by developing a method to build a database of known 3D object models over time. We present an autonomous object registration pipeline deployed on a self-contained camera and robotic arm hardware system that produces 3D object point clouds. Experimentation evaluates the quantitative and qualitative aspects of the produced point clouds, and we show that our methods produce representations comparable with state-of-the-art methods. It is our hope that the solution presented in this thesis will aid in perception-based tasks that require identifying and interacting with objects in an unstructured environment.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Colthurst, Victoria Ruth

Contributors dc:contributor

• Kim, Joohyung

• Robotics
• Robot
• Robots
• 3d
• Three
• Dimensions
• Autonomous
• Object
• Model
• Acquisition
• Gripper
• Grasp
• Grasping
• Geometric
• Grasps
• Geometry-based
• Geometry
• Registration
• Point
• Cloud
• Clouds
• Reconstruction
• Cluttered
• Rotation
• Papras
• Plug
• Robotic
• Depth
• Scan
• Scans

Statement dc:rights

• Copyright 2023 Victoria Colthurst

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/120333

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F120333