Propellant-less Orbital Maneuvering System with Superconducting Magnet Control and Physics-Informed Neural Networks for Autonomous Satellites

This whitepaper presents a revolutionary propellant-less orbital maneuvering system (OMS) enabling unlimited satellite operational lifetimes through the integration of superconducting magnetic coils, piezoelectric actuators, and physics-informed neural networks (PINNs). Based on German patent application DE 10 2025 002 893.8, the system achieves sub-milliarcsecond attitude control (≤ 0.001°) and 100 m/year Δv capability without any propellant consumption.The innovation combines • YBCO superconducting coils operating at 77 K, generating 10 000 A·m² magnetic moments,• PMN-PT piezoelectric actuators with 10 nm precision, and• an eight-layer neural network (8.4 M parameters) trained using physics-informed learning to honor fundamental conservation laws.High-fidelity simulations demonstrate a 50-year operational lifetime for a 25 kg microsatellite at 450 km altitude, consuming only 48 W average power. Key Contributions • First operational YBCO superconductor implementation for spacecraft control.• First physics-informed neural network application to autonomous satellite operations.• Synergistic multi-physics force integration achieving ≈ 40 % performance improvement.• Energy harvesting from orbital perturbations for sustained system efficiency. Economic Impact A 10× lifetime extension reduces mega-constellation replacement costs by ≈ 90 %, yielding a net present value of ≈ $106 billion for Starlink-scale deployments.Applications include constellation sustainability, military reconnaissance, formation flying, debris removal, and deep-space exploration. This work is part of the FOSS INTELLIGENCE HOLDINGS LTD. research portfolio and aligns with its long-term goal of developing quantum-resilient, propellant-independent control systems for next-generation aerospace platforms.Citation:Foss, D. T. (2025). Propellant-less Orbital Maneuvering System with Superconducting Magnet Control and Physics-Informed Neural Networks for Autonomous Satellites.Zenodo. https://doi.org/10.5281/zenodo.17466156 License: CC-BY 4.0Author: David Tom Foss (2025), FOSS INTELLIGENCE HOLDINGS LTD.