基于小波分解的时变信道盲辨识

针对传统的信道盲辨识算法很难适应时变信道模型的问题，提出了一种基于小波多分辨率分解的时变信道盲辨识算法．该算法首先应用小波多分辨率分解原理来近似建模时变信道；然后，利用信道输出序列的二阶统计量对时变信道的小波分解模型的小波系数进行两级盲辨识；最后，应用估计的小波系数即可重构时变信道的冲激响应．仿真结果表明，该算法可很好地完成周期时变信道的盲辨识与盲均衡任务．

Blind Identification of Time-Varying Channels Based on Wavelet Decomposition

The conventional blind identification algorithms cannot identify and equalize time-varying channels well. So a new blind identification algorithm for time-varying channels is proposed. The algorithm is based on wavelet multi-resolution decomposition principle. It first proposes an approximate model of time-varying channels by means of wavelet multi-resolution decomposition. Then it estimates the wavelet coefficients of time-varying channel wavelet decomposition model using the second-order statistics of channel output sequence. Finally it can construct the impulse response of time-varying channel. Simulation results show that the proposed algorithm can blind identify and equalize periodic time-varying channels well.