Channel Estimation Based on Full ISI Interference Cancellation in TDS-OFDM System

By adopting the same block-type pseudo-random noise (PN) sequences as the guard interval (GI) as well as the training sequence in time domain synchronous orthogonal frequency division multiplexing (TDS-OFDM) system, the received PN signal will be corrupted by the adjacent OFDM symbol sequences interference. A novel channel estimation algorithm is introduced through the received PN data and the tail response. Based on the equalization of linear convolution and circular convolution and residual interference cancellation theorem, Overlap save (OLS) algorithm is utilized to reconstruct the circular convolution from the received PN signal and the initial channel impulse response (CIR) is obtained. With the same PN sequences inserted in TDS-OFDM, traditional frequencydomain equalization can also be employed. The OFDM symbol interference on the received PN signal can thus be estimated and mitigated to correct the received PN signal and update the channel estimation. Simulation results show that the proposed scheme provides large improvement over the conventional method and indicates robustness in channel estimation, even if in the presence of large channel delay. Meanwhile, the proposed algorithm has comparable complexity as the conventional method.