Research On Communication Resource Allocation And Load Balancing Algorithm Of Unmanned Aerial Vehicle Aerial Base Station

The Unmanned Aerial Vehicle(UAV)is characterized by low cost,strong flexibility and easy operation.It can be equipped with communication equipment as an aerial base station to meet the communication needs of users on the ground.But the coverage range and communication resource of single UAV are severely limited,which makes it difficult to provide services for more users in a larger area.The cooperative deployment of multiple UAV base stations can solve this problem.However,how to allocate communication resources and achieve load balancing for multi-drone aerial base stations has become an urgent problem to be solved.Therefore,combined with the project "Research and Development and Test of Prototype of Air-to-Ground Broadband Data Link Communication System",the following research has been done to allocate communication resources of UAV base station and achieve load balancing in the target area:(1)Considering the scene where UAV Inter Channel Interference(ICI)exists and the ground user location information is known,a mathematical model is proposed to maximize the minimum user data transmission volume in a unit cycle time by optimizing the communication resource allocation plan.This problem is a mixed-integer non-convex optimization problem,which is difficult to be solved.Therefore,the original problem is solved step by step.The first step is to explore the optimal deployment coordinates of UAVs by using clustering algorithms,and the ground user nodes in each cluster are associated with the UAV aerial base station in the cluster center.According to whether the number of drones deployed is adaptively determined by the clustering algorithm,the mean shift algorithm or the improved K-Means algorithm is used for deployment UAVs to solve the integer constraint and the safety distance constraint between drones;the second step,fix the transmission power of each UAV after deploying the UAV and obtaining the associated plan,and then use the convex optimization theory to optimize the number of time slots allocated to each user;the third step,fix the time slot resource allocation plan,optimize the transmission power of each UAV,this non-convex problem is approximately solved as a convex problem;the fourth step is to optimize the number of time slots allocated to each user and the transmission power of each UAV by using the iterative optimization method of block coordinate descent to obtain the final communication resource allocation plan.The simulation results show that the proposed joint optimization algorithm guarantees convergence.Compared with other algorithms,it greatly improves the minimum data transmission volume of users in a unit cycle time when the number of users,user distribution,and regional edge length change.(2)Considering the scenarios of unknown location information of ground users and the fixed deployment of UAV aerial base stations,the optimal correlation scheme is proposed to solve the load balancing problem.Considering multiple network performance indicators,a multi-objective optimization problem is constructed.It is difficult to solve the multi-dimensional combinatorial optimization problem by traditional algorithms.Heuristic algorithms can solve this problem but the time complexity is too high.Thus,an artificial neural network(ANN)optimized by the Fruit Fly Optimization Algorithm(FOA)is proposed to solve this problem,that is,the ANN-FOA algorithm,and then analyze the complexity of each algorithm,the relationship between the performance of the ANN-FOA algorithm and the number of training samples and the network structure is discussed.The simulation results show that the proposed ANN-FOA algorithm is significantly better than other algorithms of the considered network performance indicators,and it is verified that the ANN-FOA of single training sample and single hidden layer structure is sufficient to meet the load balancing requirements.