A Proposal on Self-Organizing Model of Minimal Physiological Neural Network Capable of Logical Operation

Makoto Kikuchi

Keywords

Physiologic neural network, controller, modeling, selforganizing, unconscious rapid calculations

Abstract

A mathematical model of a minimal physiologic neural network that includes factors such as physiological characteristics and genetic inheritance is proposed to generate logical operation functions. Then, the mathematical models of neurocyte can form small-neural networks using self-organization algorithm. The results indicate that it is possible to create a minimal physiologic neural network capable of logical operations, including numerical calculations, which previously was believed to require higher-order functions of the brain. In particular, this mathematical model is able to design the small-neural networks that can mimic functions of insect's simple-nervous systems autonomously. There is an expectation that the results can be applied to the improvement of the self-growth miniature robot's controller in the future. However, this model exhibits limitations in that the computing time to look for a network that satisfies the necessary conditions increases when a more complex neural network is required.

Important Links:



Go Back