Evidence Hub

Spatiotemporal Prediction of Cosmic-Ray Density for Satellite Communication Systems Using Physics-Informed Neural Networks

Labeb AHM, Abdelraouf OAM

2026

Abstract

Cosmic rays represent a significant environmental hazard for modern internet satellite systems, as their interaction with onboard electronics can induce several critical effects, including single-event upsets, total ionizing dose accumulation, electronic noise, signal interference, and long-term degradation of power systems. Accurate characterization of cosmic-ray intensity variations with altitude and time is therefore essential for assessing radiation risks and improving the reliability of space-based communication infrastructure. In this study, a physics-informed neural network (PINN) framework is developed to model the spatiotemporal variation of cosmic-ray intensity by integrating governing physical equations with observational data from the Neutron Monitor Database (NMDB). Two complementary PINN models are constructed. The first model predicts the variation of cosmic-ray intensity with altitude over a range of approximately 4000 m, while the second model captures the short-term temporal evolution of cosmic-ray intensity over 3 days. By embedding physical constraints into the learning process, the proposed framework ensures physically consistent predictions while effectively leveraging observational data. The altitude-dependent model achieved a low prediction error with a mean squared error (MSE) of 9.5 × 10⁻⁴, whereas the temporal model produced an MSE of 3.12 × 10⁻², demonstrating the capability of the proposed approach to accurately capture both spatial and temporal…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free