Latency Optimization for UAVs Based Emergency Communication

Due to the importance and urgency of unmanned aerial vehicles (UAVs) in emergency communication and disaster relief tasks, this paper optimizes the UAV latency in emergency communication. We propose a system model of UAVs in the emergency communication based on the edge computing and formulate the communication latency optimization problem into a non-convex optimization problem. Reinforcement learning is used to solve the non-convex optimization problem in line-of-sight (LoS) and non-line-of-sight (NLoS) situations. The simulation results show that different methods can significantly shorten the UAV communication latency, and our proposed Improved Q-learning has a faster convergence speed and smaller communication latency compared with other methods.