利用频谱特性的速度自适应信道估计方法

Spectrum-Based Velocity-Adaptive Channel Estimation

本文利用信道在移动环境下频谱带限和对称的特征,提出一种新的对移动速度自适应的信道估计方法。该方法通过寻找频谱自相关的峰值来确定频域滤波器的长度。相对传统速度自适应方法,该方法不需要估计多普勒频率或移动速度,降低了系统实现的复杂度;也不需要信道的长时统计量或信噪比信息,实时性较高。信号在频域的处理必然会有频谱泄漏问题,传统方法滤除噪声的同时也去掉了部分信号,对此本文同时还提出一种恢复泄漏的频域信号的方法,使估计性能有所提高。仿真结果显示,从步行到高速移动的环境下,本文提出的信道估计方法的均方误差(MSE)都保持在较低的水平,具有良好的对速度自适应的估计性能。

A velocity adaptive channel estimation method using a band-limited filter is proposed in this paper. The pass-band of the filter is determined by searching the second highest peak location of the spectrum autocorrelation, which changes with velocity. Com- pared to the usual velocity adaptive estimator, no velocity or Doppler frequency estimator is needed, which greatly lowers the complexity, and no channel＇s higher order statistics or signal-to-noise ratio （SNR） information is needed, which reduces delay. The components which are higher than the maximum Doppler frequency are forced to zeros because they are just noise if there is no spectrum leakage, which is in fact not the real case due to the truncation in the time domain. Therefore, a compensation for the removed leaking signal is made after zero forcing. Computer simulation shows that, in a wide range of mobile speed, the performance of the proposed correlation method in terms of the mean square error （MSE） of estimation, only degrades a little compared to the ideal case in which Doppler frequency is perfectly known, and the compensation for the leaking signals improves the accuracy of channel estimation.