Learning to Navigate Using a Lazy Map
S.H.G. ten Hagen, S. H. G. Hagen, B.J.A. Kröse · 2003
In this paper we present a novel framework that allows a mobile robot to represent the environment and learn a task. The representation is a `lazy map' in which observation are stored. Reinforcement learning is used to enable a human teacher to train the robot using rewards and punishments.