Aaron Gage, Ph.D. Vitae: HTML, PDF

Research:

My research has focused on sensing and affect (emotions) as they relate to mobile robots. The research falls into the following areas:

• Multi-robot Task Allocation (MRTA)
• Sensor Utility Metrics
• Sensor Allocation
• Affective Computing

My dissertation dealt with using the OCC emotional model (after Ortony, Collins, and Clore) as the basis for a multi-robot task allocation solution. The approach was a contract-net protocol based negotiation where the process of choosing which robot to use for a task was distributed among the entire team, and robots would eventually be shamed into helping after initially refusing. The practical advantages of this technique (over, say, Gerkey's MURDOCH) was a statistically significant reduction in the number of communications required for large teams. This relates back to sensing because the task for which the robots are being recruited will impose certain requirements on which robots are eligible. Not only must a robot be chosen, but the best robot for the task should be.

My Master's thesis was a little different, dealing with the allocation of finite sensing resources to the behaviors on a mobile robot with incomplete information (that is, no knowledge of what will happen in the future, or how long the current demands on the system will last). The same question arises in this work: what is the best sensor for a particular perceptual task? How do you objectively measure the utility of a set of sensors to a task?

I'm still working on that part.

Publications: