This award provides funds to build a state of the art humanoid robot experimental facility at CMU's Robotics Institute. Developing humanoid robots pushes robotics and automation technology forward. Researchers build humanoid robots to test and explore their theories about how humans work. Humanoid robots are beginning to play an economically important role in the service segment of our society. Primarily, humanoid robots inspire. For example Japan is publicizing its humanoid robotics research to get more children interested in science and engineering, currently a major concern in the U.S.
The PIs will purchase an upper body humanoid from Sarcos Research Corporation as well as two smaller robots and an advanced interface device. Using the proposed equipment, the PIs will undertake a number of projects which fall into four broad classes: core robotics and vision research that substantially increases the capabilities and skills of robots (Atkeson, Hebert, Rizzi, Pollard), authoring tools that will make it easier for people who are not roboticists and perhaps even not programmers to develop new applications on robots (Hodgins, Hudson, Kuffner), application prototyping and assessment of humanoid robots in various application domains (Pausch, Kiesler), and unifying applications, which involve a large number of interdisciplinary researchers. In addition to the expected robotics researchers, the team includes researchers from HCI and sociology to work on how these robots might be made easy to use and how they will fit into society. Taken together, these projects will significantly expand the ability both to control humanoid robots and to develop applications on them. These projects will also increase fundamental understanding of the power of the human form in a robot, and our fundamental understanding of human motion. The PIs will encourage other researchers at CMU and in the U.S. to use the facility, taking advantage of strengths in vision, speech, and reasoning.
Finally, the PIs are using humanoid robots to explore computational models of how human behavior is generated. Using a humanoid robot as a research tool forces the researchers to deal with a complex physical apparatus and complex tasks. The knowledge required to recreate human motions will also serve to increase our understanding of how people move and may lead to advances in physical rehabilitation and sports medicine. In contrast to most biomechanical research, the PIs will attempt to refine scientific understanding of both the functional components of human motion such as locomotion and grasping and also the subtle stylistic components that make it appear normal, humanlike, and expressive. Combining these two components should make the models useful for analyzing the performance of aspiring athletes as well as diagnosing individual problems with locomotion and manipulation.
