Research Interests
Mobile Robotics
How do we effectively program mobile robots to accomplish tasks in the real world, despite uncertainty, noisy sensor measurements, and a changing environment? How do we design and implement robust, easy-to-use software infrastructures for robots that operate in the real world? How do we integrate techniques from artificial intelligence and machine learning into the control software of these robots, to make them more "intelligent", flexible, and adaptive in the face of a complex, changing environment? How do we get these robots to work cooperatively with humans to accomplish tasks that we care about?Projects: Software for Robots, The Robot Photographer, Remote Robotic Exploration and Experimentation, Robots and Theatre.
Human-Robot Interaction
How can we design effective interactions between humans and robots? I am particularly interested in the case where the robots are not very expressive (ie don't have human-like faces and features), and the humans do not know much about robots. How can we get the robots to understand what the humans are trying to do, and to react accordingly? How can we get these non-expressive robots to give the humans around them an idea of "what they're thinking", so that the humans feel more comfortable? We are taking ideas from psychology, philosophy, and the performing arts to allow us to build better, more natural human-robot interfaces.Projects: The Robot Photographer, Remote Robotic Exploration and Experimentation, Robots and Theatre.
Technology and the Arts
I'm interested in exploring how we can use robots, sensors, actuators, and computation in an artistic context. What does adding robotic actors to a theatrical performance add to the piece? What roles do the robots play in such a piece, and how does the audience receive them? How can we provide tools for artists to make it easy for them to integrate robots, and robotic technologies, into their work in a meaningful way? How can we get engineers, scientists, and artists to "talk the same language" and collaborate on an equal footing?Projects: Action Jackson, The Robot Photographer, Robots and Theatre.
Brain-Computer Interfaces
How do we decode the electrical activity in the human brain, and use the resulting signals to control a computer or a robot? I'm particularly interested in applying machine learning techniques to decode cortical signals recorded using electrocorticography (ECoG), associating these signals with body movements, and using them to control robot devices. How can we learn to adapt our algorithms to a particular persons brain anatomy quickly, with minimal training? How can we harness the inherent plasticity of the human brain to work in tandem with our machine learning algorithms? How can we develop effective interfaces that allow high-fidelity control of robotic devices, despite the inherent limitations of our cortical recording techniques? This work is being done with colleagues from the Department of Neurological Surgery at the Washington University School of Medicine.Projects: Control of a Robotic Hand Prosthesis Using Cortical Signals.
Machine Learning
How do we get computers to learn from experience, adapt to new situations, and extract useful information from mountains of data? I'm particularly interested in learning to control high-dimensional, complex non-linear dynamical systems using techniques from reinforcement learning. How do we practically control systems with hundreds of degrees of freedom? How can we learn from human examples and demonstrations? How can we take what we learn on one problem and apply it to another, perhaps unrelated, problem?Projects: Non-bypassable Kernel Services for Execution Security, Optimization of Complex Biomechanical Systems.
