【作者】 陈希；

【导师】 黄传河；

【作者基本信息】 武汉大学 ， 计算机系统结构， 2021， 博士

【摘要】 空天地一体化信息网络是由各种轨道卫星构成的天基网络、飞行器构成的空基网络以及蜂窝网等传统地基网络构成的一体化网络,具有覆盖范围广、部署灵活度高、成本低等优势,是未来无线移动通信的发展趋势。作为其关键技术之一,一体化网络的接入控制方法很大程度上决定了空天地一体化信息网络的通信效率。与传统移动通信不同,空天地一体化网络更远的通信距离和快速变化的拓扑结构使得现有的接入控制协议不能直接适用,因此亟需新的方法来解决诸如飞行器与地面基站、卫星与飞行器等远距离无线链路的高效通信。本文以无人机网络为研究对象,包括作为数据接收端的基站以及数量不等的无人机。这些无人机通常以侦察、勘测为目的,飞行距离长,覆盖范围广,通信实时性要求高,能够与基站直接进行通信。本文分别从无人机的接入请求、通信资源分配等方面进行研究,提出相应的解决方案,研究内容包括以下三个方面:(1)针对远距离无线网络中信道不稳定、随机请求信号易冲突的问题,提出一种信号传播延迟感知的随机接入方法。在网络环境未知的情况下,需要申请资源的无人机根据其与基站的距离,选择合适的时间点发送请求消息以减小与其它竞争节点信号的冲突概率。随机接入不可避免的信号冲突以及远距离链路信道的不稳定性降低了请求消息的成功率,因此允许竞争节点在不同的时间点发送多个请求副本,并提出了一种传播延迟感知的时隙选择算法。针对发送端和接收端分别设计了时隙选择算法和信号恢复算法,并对接收端的请求接收成功率进行了理论分析。仿真实验表明提出的算法在吞吐量、可靠性方面优于随机时隙选择算法,并通过仿真结果对复杂的接收成功率进行了曲线拟合,得到了更为精简的数学表达式。(2)针对远距离链路信号传输效率低、重传代价大的问题,提出了一种传播延迟感知的无冲突时隙分配算法,保证公平性的前提下最大化系统吞吐量。短距离无线网络通常忽略传播延迟,并假设同一时间同一信道上只有一个数据信号在传输,而多个信号同时传输则会产生冲突。在远距离通信中,由于传播延迟较大,这种情况下,与接收端距离不同的多个发送端则可以在同一时刻发送信号并且还不产生冲突。本文基于无人机与基站的距离差异,设计了针对竞争成功的无人机的时隙分配算法,以满足不同节点对时隙资源的需求并提高系统吞吐量。通过将系统吞吐量最大化的问题形式化,并利用传播延迟感知的时隙分配算法得到近似解。对网络负载饱和与不饱和两种场景下的吞吐量进行了理论分析,证明了提出的算法的次最优性,并通过仿真验证了其性能优于传统接入方法。(3)针对远距离无线通信环境下单信道正交接入信道利用率不高的问题,提出了一个多信道、功率约束的非正交多址接入优化算法,在满足接入的节点数量不超过单信道最大承载量、多个并发信号解码成功率的前提下,实现系统容量最大化。该方法首先根据无人机与基站之间信道增益的差异,对无人机进行分组,将信道增益差异较大的无人机分为一组,即可以在同一信道上传输数据;接着对同一信道上的无人机的发送功率进行优化,利用拉格朗日乘子法将系统容量最大化,将混合整数非线性优化问题转换为一个凸优化问题,然后采用梯度下降法进行求解,并设计了相关算法。最后对算法的可行性和性能进行了分析,并通过仿真实验进行了验证。更多还原

【Abstract】 Space-air-ground integrated information network is an emerging network.It is integrated by satellite systems,air networks and terrestrial networks.It has the advantages of wide coverage,high flexibility and low cost,and is the development trend of wireless mobile communication in the future.As one of its key technologies,the access control method largely determines the communication efficiency of the space-air-ground information network.Different from the traditional mobile communications,there may be more long-distance links and fast changing wireless communication scenarios in the space-air-ground integrated network,which makes the existing access control protocols not directly applicable.Therefore,new methods are urgently needed to solve the efficient communications of long-distance wireless links such as aircraft and ground base stations,satellites and aircraft.This dissertation takes long-distance unmanned aerial vehicle(UAV)networks as the research objects.A long-distance UAV network includes a base station(BS)as the data receiver and a number of UAVs.These UAVs are usually for the purpose of reconnoiter and reconnaissance.They usually have long flight distances,wide coverages and high requirements for real-time communication.UAVs directly communicate with a BS without any relay.In this dissertation,resource requests and resource allocations in long-distance UAV networks are studied,and the corresponding solutions are proposed.The main contents include the following three parts.(1)This dissertation proposes a propagation-delay aware random access method to solve the problem of channel instability and signal conflicts in long-distance UAV wireless networks.When the network environment is unknown,the UAV that needs to apply for resources selects an appropriate time point to send the request in terms of its distance from the base station to reduce the conflict probability with other competing UAVs.Due to the inevitable conflict of random access and the instability of signals in long-distance link,competing UAVs choose to send multiple copies of requests at different time points in order to improve the success rate of requests.This dissertation proposes a propagation-delay aware time slot selection algorithm.The time slot selection algorithm and signal recovery algorithm are designed for the transmitter and the receiver respectively.And the success rate of request reception is analyzed theoretically.The simulation results show that the performance of the proposed algorithm is better than that of the random slot selection algorithm.At last,by the simulation results,the complex success rate of request is fitted with a curve,and a simple expression is obtained.(2)This dissertation proposes a propagation-delay aware slot allocation algorithm to solve the problem of low transmission efficiency and high re-transmission cost.The algorithm maximizes the system throughput under the premise of fairness.In shortdistance wireless networks,the propagation delay is usually ignored.It is generally considered that there is only one signal over the same channel at the same time,or multiple signals will conflict.While on the long-distance links,due to the large propagation delays,it is possible for multiple transmitters with different distances from the receiver to transmit signals at the same time without conflict.According to this idea,utilizing the distances between UAVs and base station,a propagation-delay aware slot allocation algorithm is designed for the UAVs to satisfy the requests and improve the system throughput.The problem of maximizing system throughput is formalized,and the propagation-delay aware time slot allocation method is utilized for local optimization.In this dissertation,the throughput of the network under saturated scenario and unsaturated scenario is analyzed theoretically.Besides the optimization of the proposed algorithm is proved.The simulation results show that the performance of the proposed algorithm is better than that of the traditional access methods.(3)This dissertation proposes a multi-channel,power constrained non orthogonal multiple access optimization algorithm to solve the problem of spectrum underutilization of single channel orthogonal multiple access in the long-distance wireless networks.Under the premise that the maximum UAVs transmitting over the same channel do not exceed a specific number of UAVs,and the transmission power differences of UAVs over the same channel is enough to decode,this dissertation addresses the system throughput maximization problem.First,according to the differences of the channel gains between UAVs and the BS,the UAVs are clustered.The UAVs with large difference would be divided into one cluster.That is,the signals in one cluster can be transmitted over the same channel.Then the transmission power of UAVs over the same channel is optimized.The mixed integer non-linear optimization(MINLP)problem of maximizing throughput is transformed into a convex optimization problem by using Lagrange multiplier method.Subsequently,the gradient descent method is used to solve the convex optimization problem,and the related algorithms are designed.The feasibility and performance of the algorithm are analyzed theoretically and verified by simulation experiments.更多还原