Research On Trajectory Design And Resource Optimization Of Unmanned Aerial Vehicle Cooperative Computing Network

With the development of communication technology and the popularity of mobile devices,a large number of computing tasks appear in mobile devices.However,mobile devices are limited by factors such as size and power,resulting in insufficient resources to meet demand.In order to complete tasks with limited resources,mobile devices adopt the cloud computing mode of offloading tasks to servers with sufficient resources to complete.However,in traditional cloud computing architectures,the distance between servers and users is relatively long,resulting in high communication time costs and unfriendly to timesensitive tasks.In order to improve the quality of service,the concept of mobile edge computing(MEC)emerged.By deploying servers at the edge of the network to provide computing services to users,it can effectively reduce energy consumption and delay.Based on the advantages of mobility and network coverage,unmanned aerial vehicle(UAV)as a new mobile platform has attracted great attention.Utilizing the characteristics of strong mobility and low cost of UAV,establishing a MEC network based on the UAV platform can increase the coverage area while meeting the requirements of low delay.However,in practical applications,limited by factors such as power,the load capacity and running time of UAVs cannot fully meet the requirements.Therefore,in this thesis,a UAV cooperative network with stronger computing capabilities and wider coverage is used to serve ground devices.In addition,in practice,as the number of devices increases,there are more and more task types.UAVs must cache computing programs to complete corresponding types of tasks.Therefore,the offloading and resource allocation schemes for the UAV cooperative network in multi type task scenarios have great research value.This thesis studied an offload scheme for the cooperative network composed of base stations and multiple UAVs,and established an optimization problem for joint offload decision-making,task scheduling,location deployment,and resource allocation.Subsequently,an optimization scheme was proposed to achieve minimum completion delay while satisfying UAV resource constraints.Simulation results show that this scheme has better performance in reducing delay.In addition,UAVs with rapid mobility capabilities can play a significant role in areas lacking infrastructure.In these areas,UAVs are not sufficient to simultaneously cover all ground devices.Therefore,this thesis studied a trajectory optimization and resource allocation scheme for multi-UAV cooperative network based on task awareness.In this scheme,multiple UAVs moved along their respective trajectories to provided communication and computing resources for ground devices.To minimize the completion time,an optimization problem for joint UAV trajectory design,access decision,and resource allocation was established.Finally,to solve this problem,a trajectory design and resource allocation scheme based on deep reinforcement learning was proposed.Simulation results show that the proposed algorithm effectively reduces the task completion time of multi UAV cooperative networks.