Humanoids are bipedal robots engineered to mimic human locomotion, balance and coordination. The Honda ASIMO and KAIST HUBO are adult-sized humanoids that have captured public interest and have given researchers insight on issues ranging from balance disorders to cognition and perception. This 5-year PIRE project teams U.S. and Korean universities (Drexel University, University of Pennsylvania, Virginia Tech, Swarthmore College, Bryn Mawr Women's College, Korea Advanced Institute of Science and Technology (KAIST), Seoul National University and Korea University) to advance humanoids.

The critical technical gap that prevents a vertical advance in robotics is the lack of universally available platforms to reproduce results and validate hypotheses. This PIRE's goal is to provide humanoid platforms to a wide audience of researchers by developing a 3-tier tool set based on KAIST's HUBO humanoid: (1) virtual-HUBO is a free and open emulator for testing AI and IT concepts; (2) mini-HUBO is a low-cost 20-inch tall version of the full-sized humanoid for implementing algorithms; and (3) online-HUBO is a tethered version of the full-size humanoid that is accessible for researchers over the internet. This 3-tier approach provides little to no barriers to entry to humanoid research. These platforms will provide U.S. scientists and engineers the opportunity to leverage U.S. leadership in artificial intelligence (AI) and information technology (IT) to advance humanoid abilities in perception, cognition and social interaction.

To reach the next generation of robotic scientist and engineers, the PIRE team is working closely with the Philadelphia Please Touch Museum (PTM) to design exhibits featuring HUBO to inspire and motivate students to pursue science and engineering careers.

This PIRE project engages each member's unique resources, including electro-mechanical design (Korean collaborators), virtual-HUBO (Bryn Mawr), online-HUBO and co-op program (Drexel), mini-HUBO (Virginia Tech), advanced locomotion (UPenn) and human-robot interaction (Swarthmore). The 6-month co-op cycles (twice per year) at KAIST provide 20 U.S. undergraduates with an international research experience to cultivate skills and appreciation for effective global teaming and research. Shorter but more frequent visits by graduate students and faculty serve similarly functions but also ensure research goals and objectives are met. Lastly, Drexel's School of Education and Senior Personnel assess student performance, global teaming, and engineering skill acquisition.

This PIRE is supported by the Office of International Science and Engineering (OISE) and the Robust Intelligence (RI) Cluster of the Division of Information and Intelligent Systems (IIS) found within the Directorate for Computer and Information Science and Engineering (CISE).

Project Report

The goal of our work was part of a larger NSF project to advance the state-of-the-art of humanoid robotics in the US. Prior to 2007, there were few research groups in the US studying or developing humanoid robot systems. Almost all of the research and development was taking place in Korea and Japan. Working with a partner at KAIST, the premier technical university in Korea, our group of researchers developed both large-scale and small-scale humanoid robots we could build ourselves in the US. This was the first time US researchers could get access to high-quality humanoid robots for study and algorithm development. Having the actual robots, however, is only the first step in developing useful robot systems. In order to make humanoid robots do something useful, like interact with people, we had to give them brains, eyes, and a reason or purpose. The part of Colby College in this project was to develop the eyes of the robot and enable it to see the world through one or more cameras. In particular, we wanted the robot to be able to find people, or faces, see when somone was moving, and to identify other obstacles or objects in the environment. Only be being able to see its environment would a humanoid robot be able to start to move and interact safely and intelligently with the world around it. In addition, Colby College students developed the brains for the robot to enable it to play games. In particular, the students wrote the algorithms necessary for the robot to play both Simon Says and Red Light/Green Light with a person. In the Simon Says game the robot was Simon, and it would watch the person playing to the game to see if the person was making the proper actions. If the person made a mistake, the robot would catch the mistake and the game would be over. Creating a game like this demonstrated that it was possible for a humanoid robot to properly and safely interact with people in the real world. This is an important step to enabling robots to assist humans in their everyday lives, making people more productive and providing important services. Colby College students also developed algorithms to enable the robot to participate in the Robocup soccer tournaments. These included the ability to find and track an orange ball, to find and identify markers on a soccer field, and to find and identify goal posts. The students hope to get a whole soccer team playing together in the future. Soccer provides many real-world challenges such as motion, navigation, teamwork, planning, and visual understanding, all of which are critical to making real and useful humanoid robot systems. This project achieved its goals of improving and enhancing the humanoid robot infrastructure in the US. While there were no humanoid robots being developed in the US in 2007, there are multiple humanoid robots available in 2013, including several directly developed or inspired by this NSF grant. The US now has a thriving industry and research community for humanoid robotics. This NSF funded project enabled ten undergraduates to participate in computer science, including five women. All of these students have gone on to learn more about CS, with most becoming majors or minors. These types of experiences help to improve the diversity of computer science and open up this exciting and rapidly expanding field to everyone.

Agency
National Science Foundation (NSF)
Institute
Office of International and Integrative Activities (IIA)
Application #
0730206
Program Officer
R. Clive Woods
Project Start
Project End
Budget Start
2007-09-01
Budget End
2013-08-31
Support Year
Fiscal Year
2007
Total Cost
$2,516,000
Indirect Cost
Name
Drexel University
Department
Type
DUNS #
City
Philadelphia
State
PA
Country
United States
Zip Code
19104