Cracking the Black Box: Causal RL for Resilient Satellite Autonomy

This framework moves beyond traditional black-box AI by combining causal inference with reinforcement learning for space operations. It ensures that autonomous satellites can not only respond to anomalies under strict power constraints but also provide interpretable justifications, which is vital for long-distance mission reliability.