一种新颖的基于信道近零点分布的盲辨识算法

A NOVEL CHANNEL BLIND IDENTIFICATION ALGORITHM BASED ON DISTRIBUTION OF CHANNEL NEAR-COMMON ZEROS

研究子空间算法在阶数过估计时的信道零点分布,发现在有噪情况下额外引入的"公共零点"呈单位圆聚集性,并给出相关的理论分析。基于此,提出一种简单新颖的基于信道近零点分布的盲辨识算法。结合同样基于子空间的NECOE改进算法所得到的过估计信息,从估计的信道零点中移除"公共零点",从而实现正确的信道估计。仿真验证了该算法在较低信噪比情况下性能良好,复杂度低。

By studying the distribution of channel zeros produced by sub-space algorithm in tile circumstance of over estimating the channel order, it is found that the additionally introduced common zeros in noisy environment are gathered around the unit circle in clusters, and the correlated theoretical analysis is presented in the paper as well. On this basis, a simple and novel algorithm of blind identification based on the distribution of the channel near-common zeros is proposed. Combined with the over estimation information obtained from a modified NECOE algorithm also based on sub-space, the correct channel estimation can be achieved by removing the common zeros from the estimated channel zeros. Simulation verifies that this algorithm has good performance with a low complexity under the condition of lower SNR.