Research On Vehicular Network Coverage Enhancement Mechanism Based On Unmanned Aerial Vehicle

With the rapid development of key communication technologies such as mobile communication and Internet of Things,vehicular network will gradually be popularized and applied to real scenarios.Some high computing power and low-latency communication services in the vehicular network require intensive infrastructure deployment to provide support.Considering the flexibility,mobility and fast networking capabilities of unmanned aerial vehicle(UAV),it is a good solution to apply it to the vehicular network.However,the deployment location of UAVs is closely related to the service performance of vehicle users,and UAVs are limited by energy constraints.In order to enhance the coverage performance,we deploy UAVs flexibly and cooperatively in the vehicular network scene.The static and dynamic UAVs coverage enhancement strategies are studied respectively.The main content of this research and specific contributions in these two parts are as follows:In this thesis,a low-latency static UAVs coverage enhancement strategy is proposed.Considering that there are many unpredictable factors in the real environment,especially when the number of UAVs is not enough to directly cover all target areas,the autonomous coordination of multiple UAVs is still challenging.Therefore,an optimal deployment scheme of UAVs in vehicular network is proposed,which takes advantage of UAV’s flight capability and aims at minimizing total service delay.This scheme divides ground vehicles into two categories to calculate delays separately: For vehicles in the UAV’s coverage area,the service delay is equal to the communication delay,while vehicles without UAV’s coverage area will add additional flight delay.Under this total delay,the K-means algorithm is used to obtain the deployment range of the UAV,and then the optimal height of the UAV is analyzed,and the genetic-based heuristic algorithm is used to obtain the optimal deployment result.The evaluation results show that when the number of UAVs is limited,the optimized model has a lower service delay than the model with the maximum number of covered vehicles.This thesis proposed a trajectory optimization method for UAV,which is based on DDPG(Deep Deterministic Policy Gradient)to enhance the network coverage of vehicles on roads with incomplete infrastructure coverage,in which we consider two different cases,namely,the static state of the vehicle(caused by natural disasters)and the dynamic state of the vehicle(normal driving).In our test,multiple UAVs are used for coordinated deployment to provide high-quality services for vehicles through trajectory movement.and the trajectory optimization strategy makes use of the three-dimensional trajectory model,communication model and energy consumption model of the UAV.In order to maximize the energy efficiency of the trajectory movement of multiple UAVs and ensure the fairness of the real-time information rate of the service vehicles,a multi-UAV trajectory optimization strategy using the continuity control algorithm DDPG is proposed,which includes maximizing energy efficiency model,state space,action space and reward function.Simulation verifications were carried out for the two scenarios of vehicle static and vehicle dynamics.The results show that the proposed energy-efficient multi-UAV trajectory movement scheme based on DDPG can learn the vehicular network environment and its dynamic characteristics,and provide the vehicular network with effective coverage and superior to alternative methods in terms of information rate and energy efficiency.