Human-Like Local Navigation System Inspired by a Hippocampal Memory Mechanism

We aim to develop a robust and flexible navigation system having adaptability against changes in circumstances like a human being. A practical local navigation system inspired by the structure of the entorhino-hippocampal loop circuitry is described. The local navigation system consists of a landmark extraction unit, a learning-type matching unit and a local route memory unit. In the local route memory unit, the sequence-learning mechanism of the entorhino-hippocampal system is implemented using a fully-connected-type network. This system has two operation modes. In the learning mode, a sequence of landmarks is represented by enhanced loop connections in the connection matrix of the network. In the recalling mode, the system traces the stored route comparing current landmarks with the stored landmarks using the landmark extraction and learning-type matching units. The effectiveness of the proposed system was demonstrated using an autonomous mobile robot having the proposed local navigation system.