Predicting Robotic Grasps Using Surrogate Datasets

Honors College Thesis

Público Deposited

Baixar PDF

Citeable URL: https://ir.library.oregonstate.edu/concern/honors_college_theses/9880vs81z

Descriptions

Attribute Name	Values
Creator	Unrath, Matthew Paul Nishita
Abstract	One of the tasks that continues to prove difficult in robotics is the ability to grasp objects of varying shapes. It is time-consuming to acquire large amounts of real-world data in order to train accurate classifiers that can predict the success or failure of a grasp. To solve this issue, we examine using artificially generated surrogate, or substitute, datasets as replacement training data for more expensive physically-tested training data. By dividing up the grasp space using kd-trees, we demonstrate that surrogate datasets can be efficiently leveraged to produce high-precision data in specific areas of the grasp space. This greatly eases the burden of collecting data by only requiring physical testing in areas where surrogate datasets have little expertise. Key Words: robotic grasping, kd-tree, Gaussian Process, Logistic Regression, classification
License	All rights reserved
Resource Type	Honors College Thesis
Date Available	2014-09-29T14:42:20+00:00
Date Issued	2014-06-02
Degree Level	Bachelor's
Degree Name	Honors Bachelor of Science (H.B.S.)
Degree Field	Computer Science
Degree Grantor	Oregon State University
Commencement Year	2014
Advisor	Wong, Weng-Keen
Non-Academic Affiliation	Oregon State University. Honors College
Declaração de direitos	In Copyright
Publisher	Oregon State University
Peer Reviewed	No
Language	English [eng]
Replaces	http://hdl.handle.net/1957/52433

Parents:

This work has no parents.

Em Collection:

Miniatura	Título	Data de carga	Acesso	Ações
	Unrath_Matthew_UHCThesis.pdf	2017-07-29	Público	Baixar