Reinforcement learning-based attitude control for aerospace vehicles using small-world spiking neural networks