Humanoid Robotics: Robots that look, think and act in a way that we perceive as human or intelligent. Dexterous, mobile robots that resemble human beings are a reality. In my research, I strive to develop algorithms for planning and control that would enable these robots to autonomously perform human tasks.
Human tasks, from everyday housekeeping to complex search and rescue, almost always require significant interaction with the environment. Pushing objects out of the way, using tools and supports are all examples of the immense possibilities for environment contact. Humans not only excercise such contact routinely, but do so intuitively - often subconsciously. I believe that choosing meaningful contacts and performing useful manipulation that is unspecified in a robot's task would not only expand the skillset of current robots, but also demonstrate progress towards Artificial Intelligence in the physical world.
In contrast to humans, robots have no intuition. At least not yet. Real-world settings could contain hundreds of objects, not to mention the unlimited possibilities for support. How would the robot decide which environment interactions are useful? How would it plan to perform the interactions with guarantees of safety and stability? What would it do if something went wrong? These are the questions I address in my research.
|Research Interest Keywords|
|artificial intelligence, control, humanoid robotics, legged locomotion, manipulation, motion planning|
|The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.|
Contact Us | Update Instructions