异构网络垂直切换技术研究

发布时间：2018-07-21 16:06

【摘要】：在过去的十几年中,无线通信网络技术得到了迅猛地发展,涌现出了许多制式各异的网络。如已经演进到第4代的蜂窝通信网络、组网灵活的个域网、高带宽的WLAN以及覆盖范围较广的WiMAX,这些不同技术规格的网络在空间中重叠覆盖,构成了一个异构网络环境。这些网络在覆盖范围、性能指标等方面各有不同,同时不同网络在业务支持性方面各有侧重。未来无线通信网络技术必然是朝着异构网络融合的方向前进,下一代网络将是各种不同网络并存、协同、互补地工作,这样才能满足用户越来越多样化的需求,发挥不同网络的长处为用户提供更高质量的服务。垂直切换作为异构网络融合的基础性技术,吸引了众多的学者参与研究。本文介绍了垂直切换技术的研究背景和国内外的研究现状,并分析了基于RSS的切换策略、基于QoS的切换策略等现有垂直切换算法的优缺点。在此基础上,本文针对网络发现、触发时刻和网络选择提出了以下改进算法:(1)提出了基于预测RSS变化率的分级周期扫描机制。垂直切换包括网络发现、切换触发、网络选择和切换执行四个阶段。在网络扫描阶段,算法通过预测RSS变化率来估计网络环境变化的趋势,并分级调整网络发现的扫描周期。(2)提出了基于预测驻留时间的切换触发算法。在切换触发阶段,算法对终端在网络中的驻留时间进行预测,并结合切换耗时决定切换的触发时刻。(3)提出了基于活动业务的网络选择算法。在单业务场景下,算法根据不同业务对于网络属性的要求,利用模糊层次分析法为各个单业务制定了相适应的属性权值,并通过收益函数方法来确定最优网络。在多业务重叠的应用场景下,算法在保证活动业务关键属性的权值不变的基础上,重新分配其他属性的权值,再利用收益函数确定重叠业务场景下的最优网络。本文最后利用MATLAB仿真平台对提出的改进算法进行模拟验证:(1)终端以0.5m/s至10m/s速度移动时,分级周期扫描机制平均可以比传统固定周期扫描机制减少66%的扫描次数;(2)基于预测驻留时间的切换触发算法的掉话率比基于RSS的切换策略低12%,比基于迟滞电平的切换策略低10%,与复杂度较高的基于位置感知的切换策略相同;(3)在多业务重叠场景下,本文基于活动业务的网络选择算法在语音业务和电子邮件业务分别作为最高优先级业务的场景下优于基于业务优先级的算法。而在视频业务作为最高优先级业务时,两个算法表现出相同的切换性能。
[Abstract]:In the past ten years, wireless communication network technology has developed rapidly, and many networks of different standards have emerged. For example, the cellular communication network has evolved to the 4th generation, the flexible personal area network, the high-bandwidth WLAN and the wimax with a wide coverage range. These networks with different technical specifications overlap in space and form a heterogeneous network environment. These networks have different coverage and performance indicators, and different networks have different emphasis on service support. The future wireless communication network technology is bound to move towards the direction of heterogeneous network convergence, the next generation network will be a variety of different networks co-exist, cooperate, complementary work, so as to meet the needs of users more and more diverse. Give play to the strengths of different networks to provide users with higher quality services. As the basic technology of heterogeneous network convergence, vertical handoff has attracted many scholars to participate in the research. This paper introduces the research background of vertical handoff technology and the research status at home and abroad, and analyzes the advantages and disadvantages of existing vertical handoff algorithms such as RSS based handoff strategy, QoS based handover strategy and so on. On this basis, this paper proposes the following improved algorithms for network discovery, trigger time and network selection: (1) A hierarchical periodic scanning mechanism based on prediction of RSS rate is proposed. Vertical handoff consists of four phases: network discovery, handoff trigger, network selection and handover execution. In the phase of network scanning, the algorithm estimates the trend of network environment change by predicting the rate of RSS change, and adjusts the scanning period of network discovery. (2) A switching trigger algorithm based on predictive residence time is proposed. In the phase of handoff trigger, the algorithm predicts the terminal residence time in the network, and decides the trigger time of handoff in combination with the handover time. (3) an active service-based network selection algorithm is proposed. In the single-service scenario, according to the requirements of different services for network attributes, the fuzzy analytic hierarchy process (AHP) is used to determine the appropriate attribute weights for each single service, and the optimal network is determined by the method of revenue function. In the multi-service overlapped application scenario, the algorithm reassigns the weights of other attributes on the basis of keeping the weights of the key attributes of the activity business unchanged, and then uses the income function to determine the optimal network in the overlapped service scenario. Finally, the improved algorithm is simulated and verified by MATLAB simulation platform: (1) when the terminal moves from 0.5m/s to 10m/s, The hierarchical periodic scanning mechanism can reduce the scanning times by 66% compared with the traditional fixed period scanning mechanism. (2) the handoff trigger algorithm based on predictive residence time is 12% lower than the RSS based handoff strategy and lower than that based on hysteresis level. The handoff strategy is 10% lower than that of the high complexity location-aware handoff strategy. (3) in the multi-service overlapping scenario, In this paper, the network selection algorithm based on active services is superior to the one based on service priority in the scenario where voice service and email service are respectively the highest priority services. However, when video traffic is the highest priority service, the two algorithms exhibit the same handoff performance.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN92

【参考文献】