[RL] Quadruped Blind Trot Walking with Reinforcement Learning in "Mini-Pongbot"