Dynamic world model with the lazy potential function

Abstract

One of the fundamental skills of an autonomous mobile robot is its ability to determine a collision-free path in a dynamically changing environment. To meet this challenge, robots often have their own world model - an internal representation of the environment. Such a representation allows them to predict future changes to the environment, and thus to plan further moves and actions. This paper presents an agent-oriented dynamic world model built on top of asynchronous non-homogeneous cellular automaton, equipped with the new collision-free path finding algorithm based on a lazy potential field function. The presented abstract model is preliminarily verified using a specially designed middleware library supporting cellular model simulations. © 2011 Springer-Verlag.