LOFDM Time Domain Correlation Channel Estimation Based on the Time-Frequency Subspace Projection

LOFDM (Lattice Orthogonal Frequency Division Multiplexing), which is proposed by Strohmer T and Beaver S in 2003, has higher spectral efficiency and better bit error rate (BER) performance compared with OFDM systems in the time-frequency dispersive channel. LOFDM systems need to adapt the parameters of signals’ time-frequency lattice (TFL) and shaping-pulse scale to the channel dispersion characteristics in the transmitters, while the exact channel estimation is required for coherence demodulation in the receivers. The time-frequency interleaving characteristics of LOFDM signals induce that the time domain channel estimation can’t be achieved via the received signals directly. In this work, we separate the odd and even carriers through the equivalent time-frequency subspace projection, and a low complexity time domain correlation channel estimation algorithm for LOFDM systems is proposed based on pilot design and frequency domain pilot time domain correlation method. The interference problem is analyzed and the Cramér-Rao bound (CRB) of the time domain channel estimation for LOFDM systems is also deduced. Simulation results demonstrate that the proposed algorithm has good mean square error performance between the estimated CIR (Channel Impulse Response) and the true one, while symbol error rate performance of demodulation through the estimated CIR is near with the ideal channel.