Computer Science Thesis Proposal

Tuesday, December 13, 2016 -
9:00am to 10:00am


8102 Gates & Hillman Centers



For More Information, Contact:

We believe that it is essential for robots that coexist with humans to be able to interact with their users in a seamless way.  This thesis advocates the use of language as a rich and natural interface for human-robot interaction.  Previous work on language-based human-robot interaction has extensively focused on enabling robots to understand spatial language in the context of users giving step-by-step directions to robots.  We assume a  mobile  service  robot,  like  our  CoBot,  is equipped  with  a  map  of  its environment and is able to autonomously navigate to a desired goal position.  This thesis will address the problem of user-to-robot interaction, but is going to assume users provide a high-level specification of the task (e.g.,  "Take  the package  to  the  small-size  lab  and  then  bring  me  tea") rather than step-by-step navigational instructions.  Furthermore the thesis will focus on a novel robot-to-user interaction where the robot is able to adapt to different users, to answer user queries about its state (present, past or future), and to proactively take information-providing actions (i.e., reporting on the outcome of a task after finishing its execution).  Summing up, this thesis will contribute a novel language-based bidirectional interaction approach for mobile service robot:  from users to the robot and vice-versa.  We will evaluate the work in real and extensive real-data constructed simulated environments. Thesis Committee: Manuela Veloso (Chair) Jaime Carbonell Stephanie Rosenthal Xiao-Ping Chen (University of Science and Technology of China) Copy of Thesis Summary 


Thesis Proposal