Optimal training sequences for channel estimation in MIMO OFDM systems in mobile wireless channels

This paper describes a channel estimation and tracking scheme for MIMO OFDM systems based on pilot tones. The mean square error of the proposed scheme is derived, and through this we derive optimal training sequences and placement of the pilot tones. It is shown in this paper that the minimum mean square error (MMSE) is obtained with equispaced pilot tones in frequency, equipowered, orthogonal and phase shift orthogonal training sequences. For channels with large excess delay the number of pilot tones required for training is quite large, though a scheme for channel estimation over multiple blocks is proposed. Through simulations it is shown that the optimal training sequences derived in this paper outperform both the random and orthogonal training sequences which appear to be suboptimal. Moreover, an RLS like method is also proposed to enhance channel estimation.