Kalman filter for channel estimation in OFDM systems

摘要

A scalar Kalman filter is applied for a Least-Square estimated value H_s at s. The filter has an input for receiving H_s, a filter equation and an out for the corrected estimated value H_s^k for the k^th variable. The filter equation is H_s^k=K_gainS_n[k] wherein: correction S_n[k]=S+K_n(H_s-S); prediction of the correction S=K_aS_n[k]; Kalman filter gain K_n=P/(1+P); minimum predication MSE P=K_a²P_n[k]+K_b; minimum MSE P_n[k]=P (1-K_n); and K_a, K_gain and K_b are constants.