Research » Research
My primary interest is to bring machine learning to high-dimensional complex robots, in particular, humanoid robots. This research has two major goals: (a) to make robot control autonoumous, i.e., to develop algorithms which take the 'smart human' out of the design-and-test loop. (b) The second goal is to understand how human learning of motor control can be understood by evaluating human motor control algorithms and underlying cost functions. This involves research in several areas:
Machine Learning: Good machine learning algorithms which scale into the dimensionality of humanoid robotics, i.e., having at least 21 dimensionsional inputs (7 degrees of a human arm with velocity and acceleratio)
'''Control: