Joint Speed Control and Energy Replenishment Optimization for UAV-Assisted IoT Data Collection With Deep Reinforcement Transfer Learning

Unmanned-aerial-vehicle (UAV)-assisted data collection has been emerging as a prominent application due to its flexibility, mobility, and low operational cost. However, under the dynamic uncertainty of Internet Things energy replenishment processes, optimizing performance for UAV collectors is very challenging task. Thus, this article introduces novel framework that jointly optimizes flying speed each significantly improve overall system (e.g., usage efficiency). Specifically, we first develop Markov decision process help automatically dynamically make optimal decisions dynamics uncertainties environment. Although traditional reinforcement learning algorithms, such $Q$ -learning deep -learning, can obtain policy, they often take long time converge require high computational complexity. Therefore, it impractical deploy these conventional methods on UAVs with limited computing capacity resource. To end, advanced transfer techniques allow “share” “transfer” knowledge, thereby reducing well improving quality. Extensive simulations demonstrate our proposed solution average up 200% reduce convergence 50% compared those methods.