分布式移动通信场景下的参数化信道建模及预测优化机制研究

项目名称： 分布式移动通信场景下的参数化信道建模及预测优化机制研究

项目编号： No.61304132

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 周毅

作者单位： 河南大学

项目金额： 23万元

中文摘要： 新一代宽带移动通信网络将采用分布式协作架构，并结合多种异构接入机制和空域分集技术，提高整个网络的自由度和服务质量，从而保证任何用户在任何时刻都能接入最佳网络。复杂的分布式网络架构要求准确掌握底层多链路信道传播的特性，构建参数化仿真模型，为通信系统不同层次的算法设计与优化提供有力保障。本课题在传统的信道统计建模基础上，引入智能控制理论中的人工神经网络优化技术，针对传统信道建模中的数据分析难、计算复杂度高、模型灵活性低及其非线性等问题，利用基于几何随机图论和小波神经网络优化相结合的高精度分析方法，研究分布式移动信道的宽带多径特征、多普勒特性及多链路相关性，并利用粒子滤波器对移动信道的时变非稳态特性进行跟踪建模，构建一套适用于分布式移动通信的建模理论及仿真方法，并通过典型场景下的实测数据进行模型评估和优化。基于神经网络优化的信道模型能够适应复杂多变的移动传播环境，具有较好的理论意义和实用价值。

中文关键词： 分布式通信；移动信道建模；神经网络优化；参数估计；几何随机图论

英文摘要： Distributed cooperative communication framework will be applied in future broadband mobile communication networks. With various advanced wireless access schemes and spatial diversity technologies, distributed communications can improve the DoF and QoS of whole networks. That will ensure any user can access the best network at any time. Complicated distributed network framework demands more knowledge of the propagation channel of multi-links. Building an excellent parameterized model can provide stronger support for algorithm design and optimization in different layers of mobile communication systems. Based on traditional channel statistical modeling, this proposal introduces artificial neural network modeling scheme from the intelligent control theory. The proposed neural network optimization method can solve a series of problems in traditional channel modeling, such as difficult data analysis, high computational complexity, low modeling flexibility and non-linearity etc. Combined the geometric random graph theory with wavelet neural network, a high-resolution analysis method for distributed mobile channels is proposed in this project. Based on the proposed scheme, we will research on the broadband multipath fading, Doppler characteristics, and multi-link correlation in distributed mobile channels. The particle

英文关键词： Distributed Communications；Mobile Channel Modeling；Neural Network Optimization；Parameter Estimation；Geometrical Stochastic Graph Theory