Due to licensing restrictions, all of my results can be found in the various papers in my publications. This webpage is meant to help guide you through the papers and think about these problems in a clear, concise way.
Current and Freshly Completed Work
Ceiling Bot [Closed]
Ceiling bot was a project in *always on* ambient intelligence. In other words, how do you build a robot that uses wall power and not batteries, while still remaining mobile and actuated. The key insight being is that we pick the most architecturally vacant part of a building.
Motion Tracking for Local Warming [Ongoing]
This project is an ongoing endevour by the Sensable Cities group here at the Media Lab to address the energy demands of our growing world by applying new technology. Local warming is the idea that we can track people and keep them warm by actuating small warming lamps.
Augmented Tangibles [completed]
This project was for a class taught by Dr. Ken Perlin and Dr. Hiroshi Ishii on a concept called Eccescopy developed and pioneered by Dr. Perlin. This project, jointly investigated with Anthony DeVincenzi, focuses on the interplay between virtual reality and reality when the two become fluidly intertwined.
Mars Escape [completed]
Authoring autonomous behaviors can be seen as *the* problem to solve in smart, interactive robotics. Systems like Façade require complex behavior authoring that can take years to fully build. Mars Escape was inspired by Orkin's Restaurant Game in which everyday users take on the roles of agents in an interaction and provide authoring data that can be mined for behaviors that an autonomous agent can leverage. This was a novel project that leveraged virtual characters online to learn new behaviors that the robot could use in the real world.
Work performed during my Masters
Transparency in Learning and Interaction
Transparency, in general, allows a robot reveal its inner thoughts through nonverbal signalling and gesture. Much of my work at Georgia Tech, both in class and for my Master's Thesis explores the idea of making state well understood externally.
Social Learning Mechanisms
Social learning mechanisms as defined by Tomasello in The Cultural Origins of Human Cognition employ more than imitation learning, which is the most common type of learning in robotics. It can also be defined as learning through emulation, stimulus enhancement, and mimicking. We define novel differentiations between these mechanisms that were modeled after Tomasello's definition and characterize their effects to understand their role in social learning.