IRES: US-Korea Collaborative Software- and Hardware-Design of Humanoids for Real-World tasks in Human-Centered Environments

点击翻译按钮获取中文摘要

The Fukushima nuclear disaster was an awakening moment for robotics. Despite years of research and a wide range of robots, performing tasks like turning valves and attaching hoses, could not be done. If such tasks were executed in the earliest stages of the accident, much of the aftermath could have been averted. The fact that radiation levels were too high, forced rescue workers to wait and the disaster grew. Power plants and industrial sites are human-centered environments that are often characterized by doors, hatches, stairs and ladders that need to be entered or climbed. Additionally such environments feature valves, hoses and switches that need to be manipulated and connected. Robots however are not mature enough to efficiently function and effectively operate in such environments. This 3-year IRES project partners American and Korean roboticists who respectively have expertise in the software and hardware design of humanoids. The objective is to collaboratively design humanoids to operate in human-centered environments. This is both important and urgent, addressing both national and international needs. Such service and mitigation by robots demands the coupled development of software and hardware where the robot?s form must follow its function. The project?s intellectual merit stems from the formulation of a unified algorithmic framework. Current humanoids have the form to mimic a person?s motions but lack the software to function and operate in human-centered environments. Through the IRES, at least 12 American students will each spend 6-months in Korea working at the KAIST Hubo Lab. This partnership enables the American software developers to drive Korean hardware efforts and together design future versions of KAIST?s world-class humanoid. The net effect will be a robot that can efficiently function and effectively operate in human-centered environments like power plants and industry sites for collaborative tasks, humanitarian assistance and disaster response.The project broadly impacts research, education, business development and K-12 curricula: humanoid design will advance research in legged locomotion, whole-body motion planning, perception and dexterous manipulation; the 6-month experience at KAIST educates US students in international research collaboration; the involvement of Korean robotics companies helps US students develop products; and the after-school programs with a Philadelphia High School will yield robotics-based teaching modules. The net effect is a set of broader impacts which are enabled by the IRES and sustainable well after the project completes.