基于对现有MIMO-OFDM系统信道估计技术及其理论的分析,提出了一种基于无迹卡尔曼滤波(UKF)的信道估计算法。该算法不需要信道的任何先验统计信息,从频域出发,首先利用最小二乘(LS)算法得到导频处信道的粗略估计值,再在时域运用路径捕获法和UKF算法得到信道的精确估计值。仿真结果表明,与LS算法、最小均方(LMS)算法等相比,该算法在较低运算复杂度下可以获得信道估计性能的明显改善。
Based on the technological and theoretical analyses of present channel estimation algorithms for MIMO-OFDM systerns, the paper proposes an unscented Kalman filter (UKF)-based channel estimation algorithm which needs no prior knowledge of channel statistics. The p algorithm firstly makes use of the least square (LS) method to get rough estimation information for pilot channels in the frequency domain, and then achieves precise channel information in the time domain by using the significant tap catching method and the UKF method. The simulation results show that the proposed algorithm outperforms the conventional channel estimation algorithms such as the LS method and the least mean square (LMS) method under the moderate computation complexity.