Coordinated Reinforcement Learning-Based Distributed Optimal Locomotion Control for Biped Reconfigurable Robots Under Gait Task Constraints | Synapse