As of May 2009, a 5-year, $5 million dollar international research initiative into humanoid robots has begun thanks to a partnership between KAIST (Korean Advanced Institute of Science & Technology) and Philadelphia’s Drexel University, funded by the NSF (National Science Foundation).
Researchers from The University of Pennsylvania, Colby College, Bryn Mawr College and Virginia Tech in the United States, and KAIST, Korea University and Seoul National University will be working on the newer version of KAIST’s HUBO humanoid (called HUBO 2). First demonstrated at Robo World 2008, it is a sleeker version of the KHR-3 HUBO developed in 2005 and has been given the nickname “Jaemi” (a gender-neutral prefix roughly translating to “American-born Korean Humanoid Robot”) for its American incarnation. Videos and more after the break…
The project’s goals are to improve Jaemi HUBO’s ability to walk on uneven terrain in unstructured environments, handle objects, and interact with human beings. This version of HUBO measures 4’3″ tall and weighs 95 lbs, consisting of an aluminum endoskeleton and polycarbonate frame, costing roughly $400,000. However, the team was able to purchase the parts for only $80,000.
“With Japan and Korea aggressively pursuing robotics because the field is one of the top 10 technology areas considered as engines for their economic growth, the U.S. is in danger of losing its leading position,” said Paul Oh, associate professor at Drexel’s College of Engineering and Director of the Drexel Autonomous Systems Lab (DASL).
“Even if support existed for the United States to build its own humanoid, the learning curve and building time would likely yield models that continuously lag behind those from Asia,” said Oh. “The critical gap that prevents a vertical advance in humanoids is the lack of platforms in the U.S. and infrastructure to globally consolidate knowledge and benchmark performance.”
Dennis Hong of Virginia Tech will develop a low-cost mini HUBO for the broader robotics community to apply and test algorithms, which will stand approximately 2′ tall and cost $10,000. Once an algorithm has been properly tested on a virtual model and the mini HUBO, it can then be proven on the full-sized version.
NSF | Daniel M Lofaro | Aving News Network