Key Issues In UAV-assisted MANET

In dynamic application environments, many concurrent control parameters maylead to the complicated efects on the system performance of Mobile Ad hoc Network(MANET). There are still huge challenges to achieve a practicable MANET. This thesisaddresses the key issues of network connectivity, network performance, link dynamicsin UAV-assisted MANETs, and investigates relay strategies under diferent optimizationobjectives.First, A Cooperative Movement Strategy based on Algebraic Connectivity Gradient(CMS-ACG) is proposed. The simulation results show that the CMS-ACG is efectivein maintaining the connectivity among the UAVs. Algebraic connectivity arise as thenumber of network nodes and the connection strength threshold increase, while the degreeof improvement of algebraic connectivity decrease as the number of nodes increases.Second, we proposed a Cooperative Movement Strategy based on Weighted Cen-troid of Load and Priority (CMS-WCLP). The problem of the UAVs movement is formu-lated to minimize the sum of the weighted square of distance between the UAV and thenetwork nodes, in which the node weight is defined as the function of load and packetpriority. A network simulation environment is realized based on a system model of UAV-assisted MANETs. The results show that CMS-WCLP can efectively improve the rateof packets successfully transmitted and reduce the average packet delay.Third, we proposed a Cooperative Movement Strategy based on Weighted Centroidof Communication Duration (CMS-WCCD) for optimizing the intra-domain communi-cation link. We also derived the analytical expressions for the probability of the up-downlink and the intra-domain communication link maintenance. The results show that CMS-WCCD, in which the node weight is defined as the function of communication duration,can potentially be used to improve the performance of the intra-domain communicationlink, and the performance of the intra-domain communication link declines consistentwith the analytical expression value as the number of nodes gradually increases.Finally, the macroscopic behavior of urban mobile users are explored using methodsof random matrix theory. Based on the data of mobile users in urban grids, the deviationsof mobile users data matrix from random matrix in spectral distribution are analyzed,which reflect the macroscopic behavior of urban mobile users. The results indicate the temporal trend of the mean of correlation coefcient and the largest eigenvalue are inadvance than the temporal trend of mobile users data. We have also tried to give a rea-sonable explanation for this phenomenon.