Machine Perception Lab, Institute for Neural Computation, UC San Diego

La Jolla CA

The goal of the MPLab is to develop machines that can perceive people's behavior (e.g. facial expressions) and properly interact with humans.  This is achieved through computational analysis of real time, natural social interaction.  The MPLab is home to sociable robots that learn to interact with people.  The Lab is currently exploring the possibilities of social robots as teaching tools in early childhood education.


My role is to bring an HCI perspective to the RUBI project. I have lead 3 main projects during my time thus far at the MPlab; robot-child ethnography, robot's user-manual construction, and website architecture and flow.






The goal of the RUBI is to accelerate progress in the development of social robots.  The focus of this project has been on social robots that interact with toddlers as apart of their daily activity at UCSD's Early Childhood Education Center (ECEC). At ECEC, I directly observed robot-child interaction noting both general interactions and micro-level coding. Some key findings included comparing give-and-take exchanges between child-robot and child-teacher, and identify patterns of social relationships between child and robot. Strong tendency of children’s behaviors and interactions were noted and are currently going into consideration while we develop the next generation robot, RUBI 6.  

Child behavior library created to categorize particular child behaviors 

Child behavior library created to categorize particular child behaviors 



Medium- level operating diagram

Medium- level operating diagram



Rapidly growing popularity of RUBI prompted us to develop a manual in which other researcher labs can re-create and study a RUBI robot for themselves. Our goal was to create a manual for any user type (engineer, researcher or teacher) Profiling each user was achieved through user research, investigating common issues and gauging levels of ability during interaction with the robot.  Complex problems were broken down into smaller components and technical terms were defined that allowed for a more diverse set of robot creators.  With different user considerations in mind, I proposed a series of manual prototypes catered to multiple users.




Similar to the manual, I created a ‘RUBI Project’ webpage prototype that focuses on developing an interactive experience that effectively communicates structure, organization and information custom to RUBI in a sustainable way that can be understood by all user types.