Optimal superimposed training design for mobile OFDM systems under colored noise

In this paper, we address the problems of superimposed training (ST)-aided channel estimation for mobile OFDM systems under colored noise. By using the convex programming, the training sequences are optimized in frequency-domain, and the channel estimation is then performed over multiple symbols. In addition, an interference cancellation procedure is introduced to iteratively mitigate the information sequence interference to channel estimation. Simulation results show that the proposed design of training sequence outperforms that of the conventional ST for channel estimation in the presence of colored noise.