Impact Assessment of Solar Activity on Low Earth Orbit Satellite Systems Using Advanced Machine Learning Techniques

This paper presents a comprehensive analysis of the impact of solar activity on the operational stability of low Earth orbit (LEO) satellites, utilizing advanced machine learning techniques. We introduce Polaris ML 2.0, an enhanced platform that integrates telemetry data from the SatNOGS network with solar activity indices to predict and mitigate satellite system failures. The study employs XGBoost algorithms, optimized through Bayesian methods, and introduces a novel connectivity graph model to map the interdependencies between satellite telemetry parameters and solar phenomena. Our findings demonstrate significant improvements in anomaly detection and system resilience, offering a scalable solution for real-time monitoring and adaptive management of satellite operations in the face of space weather challenges.