Abstract:

Although Orthogonal Frequency Division Multiplexing (OFDM) system with superimposed training sequence is bandwidth-efficient , the effectiveness of the conventional channel estimation method by utilizing the firstorder statistic is restricted by the power allocation ratio and noise. In order to solve this problem , by considering the property that the training sequence and data signal are arithmetically added and experience the same channel , a maximum likelihood channel estimation scheme is proposed for OFDM systems using the superimposed training sequence. In the proposed scheme , the transmitted data signals are considered as Gaussian variables to establish the likelihood function related to the channel parameter , and an iterative maximum likelihood estimation (MLE) algorithm is derived by means of the cyclic minimizing technology. Afterwards , the lower bound of the variance is obtained and the convergence of the algorithm is analyzed. Simulated results show that the proposed scheme is of better performance in the mean square error and the symbol error rate due to the simultaneous use of the first-order and the second-order statistics of the received signals.

Key words: Orthogonal Frequency Division Multiplexing, channel estimation, maximum likelihood estimation, superimposed training sequence