Seattle, Washington
June 14, 2015
June 14, 2015
June 17, 2015
978-0-692-50180-1
2153-5965
NSF Grantees Poster Session
13
26.178.1 - 26.178.13
10.18260/p.23517
https://peer.asee.org/23517
551
Dr. Tucker holds a joint appointment as Assistant Professor in Engineering Design and Industrial Engineering at The Pennsylvania State University. He is also affiliate faculty in Computer Science and Engineering. He teaches Introduction to Engineering Design (EDSGN 100) at the undergraduate level and developed and taught a graduate-level course titled Data Mining–Driven Design (EDSGN 561). As part of the Engineering Design Program’s “Summers by Design” (SBD) program, Dr. Tucker supervises students from Penn State during the summer semester in a two-week engineering design program at the École Centrale de Nantes in Nantes, France.
Dr. Tucker is the director of the Design Analysis Technology Advancement (D.A.T.A) Laboratory. His research interests are in formalizing system design processes under the paradigm of knowledge discovery, optimization, data mining, and informatics. His research interests include applications in complex systems design and operation, product portfolio/family design, and sustainable system design optimization in the areas of engineering education, energy generation systems, consumer electronics, environment, and national security.
Object Mining by Co-robot Learning Systems in Engineering Laboratory Environments For Enhanced Student Feedback AbstractThe objective of this paper is to test the hypothesis that co-robot learning systems classifyobjects within an engineering laboratory environment with accuracies comparable to humanobservations. The term co-robot refers to a class of robots that work side by side humans, ratherthan being completely autonomous and isolated. In the process, there is an exchange ofinformation both from the human and co-robot with the ultimate goal of creating a symbioticlearning relationship, where both humans and co-robots learn from one another while performingspecific tasks of interest.Universities spend millions of dollars each year constructing new laboratory facilities ormaintaining existing ones. However students typically only have access to them during “normalbusiness hours”, as the time constraints of instructors and teaching assistants limit the availabilityof these resources. A successful integration of co-robots has the potential to expand theavailability of laboratory facilities by providing students with real time performance feedback,comparable to that of an instructor or teaching assistant. However, in the context of engineeringeducation, there exists a knowledge gap in terms of whether the integration of these co-robotsystems would have a meaningful impact on enhancing students’ performance during laboratoryactivities. In order for co-robots to provide meaningful performance feedback to students, theymust themselves acquire knowledge about the laboratory environments themselves. Specifically,co-robot learning systems must be able to detect objects within a laboratory environment thatstudents may use during laboratory activities. This would enable the co-robot systems to providefeedback to students while they used these objects.Technology advancements in computer vision and machine learning techniques are creatingopportunities for STEM researchers to integrate commercial, off-the-shelf technologies in thedesign and development co-robot systems in STEM education classrooms. In this work, theauthors employ the Microsoft Kinect to serve as the computer vision system to observe objects inthe laboratory environment. An object detection algorithm is proposed, based on machinelearning principles to enable the co-robot system to detect objects in the laboratory environmentand classify them, based on the knowledge existing in its knowledge repository. N-fold crossvalidation is employed to determine the co-robot’s object classification accuracy, compared to abaseline rate set by an instructor.The knowledge gained from this research has broad impacts within engineering education andbeyond, as researchers seek novel technology solutions that have the potential to transform themanner in which students learn and receive feedback, towards more customized modes of STEMeducation delivery.
Tucker, C., & Kumara, S. (2015, June), An Automated Object-Task Mining Model for Providing Students with Real Time Performance Feedback Paper presented at 2015 ASEE Annual Conference & Exposition, Seattle, Washington. 10.18260/p.23517
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2015 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015