The Scheme Of Drone Deployment And Energy Optimization In Aerial-ground Collaborative Networks

Due to the flexible mobility,autonomous coordination and rapid networking capabilities of drones,they have high application value in the fields of intelligent transportation systems.The Flying Ad hoc Networks(FANET)consisting of a group of drones can provide seamless,multi-dimensional flexible wireless coverage for the Vehicular Ad hoc Networks(VANET),thereby enhancing the responsiveness and emergency response capability of the ground transportation network.However,the rapid topology changes of drones and vehicle nodes pose serious challenges to network coverage and optimization.In addition,the drone capacity is limited by the endurance,and reasonable network deployment is particularly important.In view of the above-mentioned multi-objective nonlinear optimization problem and considering the low complexity requirements,this paper proposes the deployment and energy optimization mechanism of the drones under aerial-ground collaborative networks.The main research contents are as follows:(1)The aerial-ground collaborative networking architecture is proposed,and the regional filling theory is used to study the location deployment method of multiple drones under the architecture.Through the cooperative interaction between drones and drones,drones and cloud computing centers,drones and vehicles,the space-time three-dimensional position of the drones is optimized to achieve optimal coverage of the drone.Since both FANET and VANET are dynamic time-varying systems,it is necessary to determine and update the location of the drone based on the actual environment and user needs,and to meet the path loss.The coverage of the drone is optimized to provide better network coverage and higher data transfer rates for the VANET network.(2)The energy optimization mechanism of drones is studied by using Pareto optimal trade-off(POT).The energy consumed by the drone includes communication energy and mobile energy.The mobile energy consumption of the drone is proportional to its deployment height,and the communication energy consumption is inversely proportional to the deployment height.As a result,the reduction in drone communication energy consumption is at the expense of increased mobile energy consumption,which results in energy trade-offs in the air-ground cooperative networking.In order to characterize this trade-off,the optimal transmission theory is used to derive the communication energy consumption and mobile energy consumption model of the drone.With Pareto optimal trade-offs,two different drone speeds can be obtained,as well as the Pareto boundary for moving energy and communication energy consumption under different data transmission requirements.(3)The dynamic position-energy joint optimization mechanism of multiple drones based on deep reinforcement learning(DRL)is studied.The use of deep Q-learning reflects the perception of deep learning and the ability to make decisions for intensive learning.According to the current coverage and motion state of the drone,the next action is predicted,and the energy efficiency of the drone is set as a reward function to evaluate the cost of performing this action.Finally,a simulation verification platform for drone and vehicle interaction based on deep reinforcement learning is built.The proposed drone's position and energy joint optimization mechanism is simulated and compared.The simulation results show that compared with the optimal transportation theory and particle swarm optimization algorithm,the deep reinforcement learning algorithm has significantly improved the average coverage,average total energy consumption and average energy efficiency of the drone.Guide the drone's independent decision-making in a low-complexity,low-cost way,deploy the best location,and improve the performance of the drone's effective coverage,energy consumption and energy efficiency.