Adaptive Optimal Control of Unknown Constrained-Input Systems Using Policy Iteration and Neural Networks | Synapse