自适应传输MIMO系统中的信道预测技术研究

项目名称： 自适应传输MIMO系统中的信道预测技术研究

项目编号： No.61301170

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 无线电电子学、电信技术

项目作者： 孙德春

作者单位： 西安电子科技大学

项目金额： 22万元

中文摘要： MIMO系统中，发端获知下行信道状态信息(CSI)是进行自适应传输的基础。但发端获取的下行CSI都是过时的，这将导致自适应传输性能的急剧恶化，这一问题在TDD系统中尤为突出，而信道预测技术可有效解决此问题。本项目首先基于TDD系统的帧结构特征，提出结合帧间隔和导频符号间隔预测的双尺度信道预测方法，接着针对非均匀采样CSI训练序列，提出基于迭代自适应频谱分析技术的信道预测方法；然后基于正弦和信道模型，提出结合子空间技术和信息论准则的MIMO信道预测方法；最后，基于非线性预测模型，提出结合非参数噪声估计和粒子群优化算法的最小二乘支持向量机信道预测方法。对无线信道进行预测可有效解决过期CSI所致的自适应传输性能恶化问题。同时，信道预测技术应用于TDD系统可以使TDD系统无需CSI反馈信道的天然优势得以保持，增强了TDD技术的竞争力，契合了我国以TDD为通信技术基础制式的发展思路。

中文关键词： 无线通信；信道状态信息；信道预测；自适应传输；

英文摘要： Downlink CSI at the transmitters is vital to the adaptive transmission in MIMO systems. But the Downlink CSI acquired by the transmitters is always outdated, which will lead to serious performance degradation. This problem is even more serious in the TDD systems, and can be effectively solved by channel prediction. This project first proposes a double-scale channel prediction scheme where frame-scale predictor and pilot-symbol-scale predictor are jointly used for channel prediction. Then a new channel prediction scheme based on iterative adaptive approach spectral estimation is proposed for non-uniformly-sampled CSI training sequences. Next based on the sum-of-sinusoidal channel model, a channel prediction scheme is proposed where the subspace technique and information theoretic criterion are combined to perform MIMO channel prediction. Finally a nonlinear channel prediction scheme using least square-support vector machine is propose based on the non-parametric noise estimation technique and particle swarm optimization algorithm. Channel prediction can effectively alleviate the performance degradation in adaptive transmission systems caused by outdated CSI. Meanwhile no dedicated CSI feedback channel being needed, which is considered to be the inherent advantage of the TDD systems, can be kept by using channel p

英文关键词： wireless communications；channel state information；channel prediction；adaptive transmission；