A Bio-Inspired Model of Navigation in a Multi-Chamber Maze

Abstract

We present a model of spatial memory inspired by rodent hippocampus. Hippocampus is a part of brain involved in spatial orientation. The model provides a robust and efficient method of storing several navigational maps in a single attractor neural (Hopfield-like) network. It models such phenomena as place cells, long-term potentiation, long-term depression, path integration, inhibitory interneurons etc. Structure of the model conforms to functional schema of hippocampal formation. In order to validate the model we've developed a spatial task within a maze of several chambers. The robot has to follow a complex trajectory passing gateways and avoiding barriers. The algorithm could provide a potential foundation for a future robotic use.