车联网广义分组级信道建模方法研究

To fully understand the impact of high movement speed on vehicular network performance, channel modeling is always regarded as the core components of the vehicular networks. Due to the current channel modeling methods cannot capture the temporal statistical characteristics of fast time-varying channels and channels with random fading rates, system cannot adjust and optimize transmission schemes accurately and timely according to the dynamic variation of channels, which lead the low spectrum efficiency vehicular network with high-speed movement. This project proposes a generalized packet-level channel modeling methodology, firstly, the steady and temporal statistical characteristics of channels are described by the finite state Markov chain with any order; based on this fact, the relation between moving speed and memory length of channel is formulated as order estimation problem of Markov chain to quantitatively evaluate the impact of moving speed on memory length of vehicular channel; finally, aim to accurate describe the variation of temporal statistical characteristics and memory length of vehicular channels with any fading rates. As the assumption of that the channel state is captured by one order Markov chain is broken, the packet-level channel modeling method is generalized. Generally, the results obtained in this project will enrich the existing channel modeling methods, and provide a serial of new guidelines for the vehicular networks’ deployment.

为了充分认识高速移动对车联网性能所可能带来的影响，信道建模一直是车联网研究的关键内容。现有的信道建模方法无法准确刻画快速时变和衰落速率随机变化信道的暂态统计特性变化，造成系统无法跟踪信道的动态变化来及时地对传输方案进行优化调整，最终导致高速移动条件下车联网的频谱利用率严重下降。本课题提出一种广义的分组级信道建模方法，首先采用任意阶有限状态马尔科夫链来刻画信道的稳态以及暂态统计特性；在此基础上，将移动速度与信道记忆长度的关系转化为信道状态序列马尔科夫阶数求解问题，来定量描述车辆移动速度对于信道记忆特性的影响；最终旨在准确描述任意衰落速率条件下信道的暂态统计特性变化以及信道记忆特性的变化。本课题突破采用一阶马尔科夫链描述信道状态变化的假设，实现了信道建模方法自“狭义”向“广义”的转变。本项目的研究成果不仅能作为现有分组级信道建模方法的必要补充，也为未来车联网部署和优化提供重要的信道模型工具。

为准确刻画环境及移动速度对车联网信道质量的影响，本课题提出一种广义分组级信道建模方法。首先采用任意阶有限状态马尔科夫链来刻画信道的稳态以及暂态统计特性；在此基础上，将移动速度与信道记忆长度的关系转化为信道状态序列马尔科夫阶数求解问题，来定量描述车辆移动速度对于信道记忆特性的影响。经过理论分析和仿真及测试验证，新的建模方法在保证模型精度的前提下显著地降低了计算复杂度。本课题突破采用一阶马尔科夫链描述信道状态变化的假设，实现了信道建模方法自“狭义”向“广义”的转变。在圆满完成项目预期目标的基础上，本项目还将新的信道模型应用到了包括：车联网自适应传输、车联网资源管理及移动边缘网络计算迁移等系统性能易受到车联网信道快速时变特性影响的环节，进行性能分析和策略设计。本项目的研究成果不仅能作为现有分组级信道建模方法的必要补充，也为未来车联网新应用的部署和参数设置提供了明确的理论指导。