sec_related_work.tex

\section{Related Work}
	\label{sec:related_work}
	
	From some years ago, there has been plenty of research on fully autonomous robots capable of performing
	tasks in structured environments (kitchens, offices, etc.), as the one presented by Blodow et al~\cite{Blodow}
	or Beetz et al~\cite{Beetz}.
	For that purpose many different control architectures have been proposed, some of them are described by
	Medeiros~\cite{Medeiros}.
	For non-structured environments there is one basic paradigm called \emph{supervised autonomy},
	proposed first by Cheng and Zelinsky~\cite{Cheng}, which has become the current state of art for
	the DARPA Robotics Challenge (DRC) since the Trials and also for the Finals.
	During these competitions, each team relied on an Graphic User Interface (GUI) showing a 3D model of the
	robot and the environment, in such a way that the operator(s) could control the robot beyond the joint-level,
	by specifying task-level commands that were robot-centric and/or object centric, as described by the teams
	Tartan Rescue~\cite{Dellin}, MIT~\cite{Fallon}, RoboSimian~\cite{Hebert} and ViGIR~\cite{Romay}.
	In this paper, we describe our implementation of this approach.