As satellite fleets expand, autonomous systems are becoming important for managing spacecraft when communication with Earth is limited or unavailable. CUS-GNC has developed SpacePilot, reusable onboard software that combines guidance, navigation and control (GNC) algorithms with trained neural networks to help spacecraft make mission-related decisions and respond to changing conditions. A related application aims to address equipment failures by assessing how a spacecraft can operate with its remaining capabilities.
The company uses Mathworks MATLAB and Simulink to design algorithms, model spacecraft dynamics and simulate missions, while Aerospace Toolbox helps model the space environment. Monte Carlo simulations test the neural networks, and representative onboard computer testing and flight telemetry help validate and refine the software, which is already operating on multiple satellites. The approach aims to let fleets grow without enlarging control teams.
Image generated by: ChatGPT