Models for map building and navigation

In this paper we present several models for solving map building and navigation problems. These models are motivated by biological processes, and presented in the context of artificial neural networks. Since the nodes, weights, and threshold functions of our models all have physical meanings, we can easily predict network topologies and avoid traditional trial-and-error training. On one hand, this makes our models useful in constructing solutions to engineering problems (problems such as those that occur in robotics, for example). On the other hand, this might also contribute to the ability of our models to explain some biological processes, few of which are completely understood at this time.<
>