Framework for Hopfield Network based Adaptive routing - A design level approach for adaptive routing phenomena with Artificial Neural Network

TL;DR

This paper proposes a framework utilizing Hopfield Neural Networks to enable adaptive routing in satellite constellations, addressing the unique geometric and dynamic challenges of LEO satellite networks.

Contribution

It introduces a novel design framework for applying Hopfield Networks to adaptive routing in satellite constellations, considering their specific network geometry and environment.

Findings

Framework effectively models adaptive routing in satellite networks

Hopfield Network-based approach improves routing efficiency

Addresses non-conventional routing challenges in LEO satellites

Abstract

Routing, as a basic phenomena, by itself, has got umpteen scopes to analyse, discuss and arrive at an optimal solution for the technocrats over years. Routing is analysed based on many factors; few key constraints that decide the factors are communication medium, time dependency, information source nature. Parametric routing has become the requirement of the day, with some kind of adaptation to the underlying network environment. Satellite constellations, particularly LEO satellite constellations have become a reality in operational to have a non-breaking voice/data communication around the world.Routing in these constellations has to be treated in a non conventional way, taking their network geometry into consideration. One of the efficient methods of optimization is putting Neural Networks to use. Few Artificial Neural Network models are very much suitable for the adaptive control mechanism, by their nature of network arrangement. One such efficient model is Hopfield Network model. This paper is an attempt to design a framework for the Hopfield Network based adaptive routing phenomena in satellite constellations.