基于信息物理系统模型的车载自组织网络拥塞控制机制研究

发布时间：2018-03-29 14:41

本文选题：车辆自组织网络　切入点：拥塞控制　出处：《西安电子科技大学》2015年硕士论文

【摘要】：车辆自组织网络(Vehicular Ad-hoc Networks, VANETs)是一种特殊的无线自组网(Mobile Ad-hoc Networks, MANETs),它是由车辆之间(Vehicular to Vehicular, V2V)以及车辆和基础设施之间(Vehicular to Infrastructure, V2I)组成的无线自组织网络,车辆可以通过VANETs网络向周围车辆发送含有车辆速度、位置等基本信息,也可以在交通事故时向周围车辆发送警告信息以避免发生连续交通事故,同样也可以提供车载娱乐提高交通出行的趣味性。在车载自组织网络中,车辆之间信息的交换扮演着重要的角色,使得车辆之间可以相互感知并且为车辆的安全应用或非安全应用提供参考信息。因为车辆周期性的发送消息,当车辆过多时,如果不采用拥塞控制策略,信道就会变得拥塞,从而导致干扰增加,接收端信干比下降,严重影响车辆间的通信效率。现有的拥塞控制方法大都基于调整发送功率或发送速率,忽视了载波监听门限对网络拥塞的影响,而通过调整载波监听门限可以控制网络的时延、吞吐量等。本文的主要工作在于：第一,通过分析载波监听门限和竞争窗口对网络性能的影响,提出了一种载波监听门限和竞争窗口联合控制的算法,在网络中,载波监听门限和竞争窗口是两个很重要的参数,载波监听门限可以决定一个节点能否进行发送信息。而竞争窗口则决定着当节点侦听到信道处于忙时的退避时间。通过分析,我们使用一个能反映网络状态的参数将网络分为三种状态,并根据网络的不同状态设置不同的载波监听门限值和竞争窗口值,且采取不同的措施来自适应的调整载波监听门限和竞争窗口。通过分析表明,我们提出的载波监听门限与竞争窗口联合控制算法可以很好的提高网络的吞吐量、降低误码率、降低分组时延等。第二,提出了一种基于线性二次型高斯最优控制(Linear Quadratic Gaussian, LQG)的车辆网络拥塞控制方法。通过分析表明,载波监听门限值对网络的吞吐量有着重要的影响,但在实际中,网络的吞吐量往往和预期值有所偏差,因此我们可以根据实际网络吞吐量与预期网络吞吐量的差值动态的调整载波监听门限,提高网络的吞吐量,达到我们的期望值。我们利用线性二次型最优控制的方法来求得所需的载波监听门限。本方法可以分为两个步骤：(1)利用卡尔曼滤波器来消除获取载波监听门限时噪声带来的影响,以得到精确的网络吞吐量。(2)利用线性二次型最优控制方法求得最优的载波监听门限。求得的最优载波监听门限可使网络吞吐量达到预期值,可确保网络不发生拥塞。仿真表明,本文提出的基于LQG的拥塞控制算法可以有效的提高网络吞吐量,使网络即使发生拥塞的情况网络吞吐量也能保持在一个合理的水平。
[Abstract]:Vehicular ad hoc networks (Vehicular Ad-hoc, Networks, VANETs) is a kind of wireless ad hoc network special (Mobile Ad-hoc Networks, MANETs), which is composed of the vehicle (Vehicular to Vehicular, V2V) and between vehicles and infrastructure (Vehicular to, Infrastructure, V2I) composed of wireless ad hoc networks, the vehicle can contain the vehicle speed to send around the vehicle through the VANETs network, location and other basic information, also can be in the traffic accident to send warning information around the vehicle to avoid continuous traffic accidents, can also provide car entertainment improve transportation interest. In thevanet, exchange information between vehicles plays an important role, so between the vehicles to provide reference information and mutual awareness for the safety of vehicles or non application security applications can send messages. Because of the vehicle when the vehicle cycle. Over time, if not the congestion control strategy, the channel will become congested, resulting in increased interference, the receiver SNR decreased, seriously affect the efficiency of communication between vehicles. Most of the existing congestion control method to adjust the transmit power or the sending rate based on the effect of listening to the network congestion threshold wave load, and by adjusting carrier sense threshold can control time delay, the throughput of the network. The main work of this paper is: first, through the analysis of the influence of carrier sensing threshold and contention window on the network performance, proposed a carrier monitoring threshold and contention window combined control algorithm, in the network, a carrier sensing threshold and the contention window are two the important parameters of carrier sensing threshold can determine whether a node can send information. And the contention window determines when a node listens to the channel is busy when the backoff time . through the analysis, we use a network can reflect the state parameters of the networks are divided into three kinds of state, and according to different network conditions set different carrier sensing threshold and contention window value, and take different measures to adjust from the carrier sense threshold and competition window. The analysis shows that the carrier sense we propose a threshold and contention window control algorithm can improve network throughput, reduce the error rate, reduce the packet delay. Second, proposed a linear optimal control based on the two Gauss (Linear Quadratic Gaussian, LQG) of the vehicle network congestion control method. The analysis shows that the carrier sense threshold on the throughput of the network have an important impact, but in reality, the network throughput often and the expected value is different, so we can according to the actual network throughput and expected The difference of dynamic adjustment of the network throughput of carrier sensing threshold, improve the network throughput, meet our expectations. We use the method of linear quadratic optimal control in two to obtain the desired carrier sense threshold. This method can be divided into two steps: (1) to eliminate the influence of access threshold noise caused by carrier sense using the Calman filter, in order to get the accurate network throughput. (2) carrier sensing threshold with two linear optimal control method to obtain the optimal threshold is obtained. Optimal carrier sensing can make the network throughput reach the expected value, which can ensure the network congestion. Simulation results show that can effectively improve the network throughput of LQG congestion control algorithm based on this, the network throughput of the network even if congestion occurs can be maintained at a reasonable level.

【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;U463.67;TN929.5

【相似文献】