时变信道复指数基频率盲估计

Blind estimation of complex exponentials basis frequencies of time-varying channels

在基于基展开模型的时变信道盲辨识中,可通过估计复指数基频率将时变信道的盲辨识问题转换为线性时不变信道的盲辨识问题。为了更准确地实现复指数基频率的盲估计,针对两种不同输入条件下的观测数据,分别提出了利用特定二阶循环矩和四阶循环累积量方法来进行估计。特定二阶循环矩方法比传统方法更容易提取基频率分量;四阶循环累积量方法比四阶循环矩方法具有更好的抗噪声性能。理论和仿真分析证明了这两种算法的有效性。

In the blind identification of time-varying channels using the basis expansion models,the identification can transform to blind identification of time invariant channels by estimating the complex exponentials basis frequencies.In order to estimate the complex exponentials basis frequencies accurately,this paper aimed at two types of observed data under the condition of various input and proposed the methods of exploiting specific second order cycle moment and fourth order cycle accumulation for estimating the complex exponentials basis frequencies.Compared with traditional method,the basis frequencies components were much more easily extracted by the method based on specific second order cycle moment;otherwise,the method based on fourth order cycle accumulation had better performance resistance to noise than the fourth order cycle moment.Simulation results prove the effectiveness of the proposed two algorithm.