Throughput Optimization Of Unmanned Aerial Vehicles-Ground Nodes Collaborative Self-Organizing Network

In natural disasters,military operations and other scenarios,the fixed communication infrastructure is destroyed,causing the ground personnel communication blocked.It is difficult for these people to carry out joint search and rescue and achieve cooperative operations.With the maturity of unmanned aerial vehicles(UAV)technology,it is possible to build a high performance air-to-ground self-organizing network using unmanned aerial platform,which is of great importance to improve the efficiency of disaster relief,guarantee life safety and win the information war.This thesis focuses on the optimization of the throughput of UAV-ground nodes collaborative self-organizing network by UAV deployment.The main work and innovations are as follows:1.We designed a multi-level distributed structure UAV-ground nodes self-organizing network system,and established the mathematical model of network throughput.In the static scene of fixed ground nodes location,firstly,UAV is assumed to master the whole network node locations.The gradient-based UAV positioning algorithm based on global information is proposed to optimize the average end-to-end throughput of the system.In view of the non-smoothness of the objective function,we introduced the generalized gradient,and proved the convergence of the algorithm by using the related theory of the non-smooth dynamic system.Secondly,we proposed the gradient-based positioning algorithm based on local information under the condition that the UAV only grasp the node positions in the range of one hop.The simulation results in 40 sampling scenarios show that the performance of two gradient-based algorithms is excellent,and the mean deviation of the throughput optimization results compared with the optimal value are 0.027%and 0.817%,respectively.In the end-to-end throughput fairness,the latter is better than the former.2.In dynamic scene of motional ground nodes,in order to balance the competition of communication resources among users in different UAV subnetworks,firstly,we studied the user grouping algorithm based on KM algorithm,and the uniform grouping in number is realized on the premise of minimizing the distance between ground nodes and UAV.Secondly,a UAV speed control strategy based on line search is proposed,which can make the UAV respond to the network topology changes quickly and reduce the flight oscillation at the non-smooth boundary of the objective function.Finally,a collision avoidance strategy based on artificial potential field is proposed,in which repulsion field is introduced between UAVs to avoid getting too close to each other and influencing the flight safety.Combined with the ad hoc network node mobility model,the simulation results showed that the algorithms in the dynamic scene can be effectively combined with the gradient-based UAV positioning algorithm to achieve real-time optimizing of the system throughput.3.We built a simulation platform of UAV-ground nodes coorperate self-organizing network system,and developed the relevent softwares.The algorithms studied in this thesis are verified in the indoor simulation experiments.Simulation results show that the algorithms can keep good cooperation between UAV and ground nodes,and have the ability of adaptively optimizing the system throughput,which proves the feasibility of building high throughput UAV-ground nodes self-organizing network system based on this series of algorithms.