Modular Deep Reinforcement Learning for Continuous Motion Planning with Temporal Logic | Read Paper on Bytez