Evolving Spacecraft for the Free-Molecular Frontier: AI-Optimized Design for High-Drag VLEO Missions

Very Low Earth Orbit (VLEO, ~150-450 km) has become an increasingly popular region of space for both commercial and scientific satellites, as operating in this orbital regime provides benefits such as lower communications latency and improved Earth imaging capabilities. However, due to higher atmospheric densities at these altitudes, spacecraft are subjected to increased drag forces, resulting in rapid orbital decay. To optimize spacecraft designs for VLEO, we implement an Evolutionary Algorithm (EA), that draws on principles of biological evolution to solve optimization problems. We present an EA that evolves spacecraft designs from primitive shapes and uses Vehicle Environment Coupling and Trajectory Response (VECTOR) software to simulate atmospheric drag. We also report a heuristic study on surface feature sensitivity to aerodynamic coefficients, allowing us to identify practical size scales and shape classes for geometric primitives. This work presents optimized spacecraft designs evolved by the algorithm and discusses surface feature classes that disproportionately impact drag susceptibility. Spacecraft optimization is an expansion of our EA, which was used in previous works to develop antenna models optimized for sensitivity. Future work involves representing spacecraft designs as 3-D point-clouds in lieu of geometric primitives for higher-fidelity and resolution of small-scale features.