基于FastICA的高精度多径时延估计算法

摘要

在高精度无线定位中,针对现有超分辨率时延估计算法在低信噪比时性能较差的问题,提出了一种基于FastICA的多径时延估计算法.利用噪声子空间对白化矩阵进行修正,并采用参考信号约束初始分离矩阵,算法可快速收敛至所需的首达径分量,最后利用参考信号提取首达径时延.仿真结果表明,在信噪比较低的多径环境中,与已有超分辨时延估计算法相比能够有效提高时延估计精度.%Aming at improving the performance in low SNR environment of super resolution time delay estimation(TDE) in high precision wireless position location, this paper peoposed a multipath TDE algorithm based on FastICA. By modifying the whitening matrix with noise subspace and applying constraint to initial unmixing matrix in FastICA with reference signal, the algorithm reduced the complexity to converge to the expected FAP component. Finally, the time of arrival(TOA) of FAP could be achieved with reference signal. Simulation results showed the improvement on TDE performance compared with current super resolution algorithms in low SNR environment.