Michael is currently a PhD student in the Cognitive Science Department at RPI. Michael received a BS in Brain & Cognitive Sciences at the University of Rochester. His interests include agent learning, planning, and social interaction between agents and humans. His current research revolves around agent planning and behavior in interactive narrative and storytelling environments.
“Interactive Storytelling with Cognitive Robots” is a project directed to have robots, each with their own goals, behavior, and emotions adaptively convey a story to users. Interactive narratives are typically designed for computer users. The virtual environment and virtual characters are displayed on a computer monitor or projected on a big screen, and the user interacts with the virtual characters indirectly through a screen using keyboard, mouse or microphone. Cognitive robots are designed to create a more immersive experience as they can embody virtual characters, and make the interaction more natural and intuitive.
Currently, in our research, we examine the way movements of the robots are used to express emotions, illustrate locations, and suggest the social relationships among the characters. As an example of this functionality, if the user is disengaged from the conversation and turned away from the robot, the robot will move to the front of the child and try to initiate another conversation. The agent design has to be tolerant to potential errors from the robots’ vision and motor systems. Unlike digital avatars, whose verbal behaviors and body movements can be precisely coordinated by the agent, the agent has to consider the physical capacity, e.g. moving speed of the robot when planning the character’s behaviors.