Research On Energy Efficiency Optimization Of UAV Communication System Based On Reinforcement Learning

UAVs have played an important role in military reconnaissance,traffic monitoring,agricultural planting,and other fields due to their fastmoving speed and flexible deployment characteristics.Therefore,the integration of future communication networks and UAVs has attracted much attention.However,the limited energy of UAVs has caused many difficulties in the communication system.Therefore,according to the actual communication scenario,the power,trajectory,and other variables of the UAV are optimized based on the reinforcement learning algorithm to improve the energy efficiency in this thesis.The main innovations of this thesis are listed as follows:1.An optimization scheme for UAV-enabled video transmission policy is proposed.A UAV-enabled secure video transmission system model with eavesdroppers is constructed,and the system goal is to maximize the energy efficiency under the constraint of video leakage probability.A safe deep q-learning network(Safe-DQN)algorithm is proposed to construct a safe policy space.The simulation results show that the proposed algorithm is more secure and energy-efficient than other optimization algorithms.2.A joint optimization scheme for UAV-enabled video transmission policy and trajectory is proposed.In addition to the interference noise of the macro base station,the UAV moves away from the eavesdroppers to reduce video leak probability further.The optimization problem is modeled as a constrained Markov decision process to maximize energy efficiency,and the Safe-DQN algorithm is applied to solve it.As a result,compared with other algorithms,the video leakage probability of the proposed algorithm decreases significantly under the same energy efficiency.3.An energy efficiency optimization scheme of a UAV-enabled video transmission system based on mobile edge computing is proposed.The constructed system includes several observation UAVs capturing videos at fixed trajectories and a relay UAV equipped with a mobile edge computing server.In order to improve the energy efficiency of the system,a UAV decision-making algorithm is proposed based on deep deterministic policy gradient.The power allocation,the computing resource allocation,the trajectory and the transcoding policy are optimized by the proposed algorithm.The simulation results show that the energy efficiency of the proposed algorithm is higher than other algorithms under different environmental parameters.