| UAV Ad Hoc Network(UAVNET)has been in the spotlight for its low cost,high flexibility and cooperative capability to carry out complex tasks.In addition to the dynamic characteristics of general UAVNET,the distance between nodes of large-scale and high dynamic UAVNET often reaches hundreds of kilometers.For the sake of supporting longdistance communication,UAV nodes in such kind of network are necessarily equipped with directional beam antenna to obtain high channel gain.The existing MAC protocol,however,cannot be directly applied to such scenario.In this thesis we study and design a directional MAC protocol suitable for this scenario for a large-scale and high dynamic UAVNET.The main work are as follows:First,the design of directional MAC protocol based on TDMA.In the case of no satellite navigation,the beam tracking algorithm is proposed in this protocol to update the beam direction in current communication through the information obtained from the previous communication period.The network is initialized by the application of designated master node or competing master node,and the control channel as well as data channel is adopted to manage the network topology for data communication.For the data channel,we propose a general fixed slot and frequency allocation mode based on N nodes and ??N2??frequencies,which enables each possible link to have resource allocation in each cycle,so that the link can be reconstructed or the link interruption can be found in time when there is no service data transmission.A “Beam-Inference” mechanism without satellite navigation is proposed that when the attitude of the node changes,the beam direction of the directional antenna is corrected to maintain the original link.Then,based on the OPNET simulation platform,the problem of beam direction and beam angle convergence in OPNET is solved.Finally,the simulation results show that the network access time,throughput and time delay meet the actual needs.Compared with the initialization mode of the designated master node,the competing master node can accommodate more nodes but with longer time to access the network.Second,the optimization of the above mentioned beam tracking algorithm.When there is a long distance between the nodes,it is not easy for the nodes in the above beam tracking algorithm to move out of the beam range in a period.However,this will not be held so when the two nodes are not far away from each other,which may cause communication failure between nodes.Therefore,the beam tracking algorithm based on Kalman filter is proposed,which uses the estimated value at the previous moment and the observed value at the current moment to estimate the beam direction at the present moment.Analysis results show that the communication success rate is increased by about 70% compared with the original method with 8° narrow beam.Third,the improvement of the above-mentioned network initialization mechanism.Due to the fixed slot and frequency allocation mode,there is a large correlation between the network access time of competing master nodes and the distribution range of nodes.The larger the distribution range is,the longer the network access time is.In the improved mechanism,the scanning,control as well as data beams are separated from the frequencies,and the slot allocation mode for the control link is proposed to ensure the real-time performance of the information receiving and transmitting when each possible link has resource allocation.The simulation results indicate that the network access time of nodes is less than that in the original initialization mechanism,and it is no longer restricted by the node distribution range. |
