New trends in robotic reinforcement learning: Single and multi-robot case

Abstract

A rather general approach to learning control is the framework of Reinforcement Learning, described in this chapter. Reinforcement learning offers one of the most general framework to take traditional robotics towards true autonomy and versatility. Single robot reinforcement learning as well as Multi-robot reinforcement learning are a very challenging areas due to several issues, such as large state spaces, difficulty in reward assignment, nondeterministic action selections, and difficulty in merging learned experiences from other robots. There are still many difficulties in application iof robotics reinforcement learning and in scaling up the multi agent reinforcement learning to multi-robot systems. After reviewing important approaches in this field, some problems and promising research directions will be given. © 2009 Springer-Verlag Berlin Heidelberg.