Sigma-Point卡尔曼滤波用于OFDM载波频偏估计

Application of Sigma-Point Kalman Filter to Carrier Frequency Offset Estimation of OFDM Systems

对于非线性的动态状态空间模型,扩展卡尔曼滤波(EKF)通过泰勒展开拟合系统的状态和观测方程,以获得对状态值的估计,但其存在估值波动大、收敛慢等缺点;而基于Sigma-point点的卡尔曼滤波方法,则是通过确定性采样实现统计特性上的近似,从而获得更为准确的高阶统计特性.为此,建立了正交频分复用(OFDM)载波频偏的动态状态空间模型,并将Sigma-point卡尔曼滤波用于其频偏估计.仿真结果表明,该类方法可以捕捉更为准确的高阶特性,其估值准确、收敛速度快、波动小、对观测噪声大小不敏感.

For the non-linear dynamic state-space model, extended Kalman filter （EKF） fits the system state and ob- servation equations to obtain the estimation of state, but it has deficiencies like apparent fluctuation and slow conver- gence. While the Sigma-point Kalman filters obtain the statistical characteristics based on deterministic samples, and accordingly better approximation can be achieved. In this paper, the orthogonal frequency-division multiplex- ing （OFDM） carrier frequency offset is described as non-linear dynamic state-space model （ DSSM ） , and the Sigma- point Kalman filter is applied to the estimation of the offset value. Simulation results show that the proposed filter perform better at capturing high order moments than EKF, with higher accuracy, faster convergence, smaller fluctua- tions and lower noise sensitivity.