Text this: Multi-UAV Reinforcement Learning With Realistic Communication Models: Recent Advances and Challenges