I research physical human-robot interaction, with an emphasis on theoretical contributions.
In some of my earlier work, I focused on robots teaching humans: how robots should interact in order to teach users about the task, the environment, or the user’s performance. My projects involving robotic rehabilitation and series elastic actuators fall under this teaching category. Within robotic rehabilitation, I studied how robots can teach natural motions to people such that (a) the human does not become dependent on the robot’s assistance and (b) the human retains what they have learned after the robot is removed. My work on series elastic actuators explored what range of environments these devices could teach to the human, with an emphasis on ensuring human safety during interaction.
More recently, I have started to research how robots can learn from humans during physical human-robot interaction. This body of work, learning from physical interaction, treats the human’s interactions as intentional; the human interacts because the robot is doing something incorrectly, and the human’s interaction is meant to correct that mistake. A key insight here is that the robot should not keep behaving the same (incorrect) way after the human interacts—instead, the robot should update its behavior to try and match what the human wants. My work studies ways in which the robot can learn from the human and update its behavior based on physical human interactions.
To read more about these projects, click one of the links above!