Decision-Making in Robotics

This paper presents, as a tutorial, an overview of the current state of the decision-making problem applied to robotics. The study is approached in a comprehensive and inclusive way and, therefore, trying not to go into specific solutions for specific problems. The article is focused primarily on high-level decisions to be taken by a robot, not lower-level decision-making problems that are solved using traditional control techniques or with very specific algorithms. We refer to "decisions" of a robot in the broad sense of determining the activities performed by it. That is, without any a priori exclusion based on issues such as the decision-making strategy used, the type of robot, the performed tasks, etc. The article is structured in a number of sections that addresses various topics of interest in robotics, from the perspective of decision-making: autonomy, intelligence, goals, high-level decisions, decision-making strategies, control architectures, perception, human-robot interaction, learning and emotions. Copyright (C) 2010 CEA.