Sparse Recovery of Linear Time-Varying Channel in OFDM System

In order to improve the performance of linear time-varying (LTV) channel estimation, based on the sparsity of channel taps in time domain, a sparse recovery method of LTV channel in orthogonal frequency division multiplexing (OFDM) system is proposed. Firstly, based on the compressive sensing theory, the average of the channel taps over one symbol duration in the LTV channel model is estimated. Secondly, in order to deal with the inter-carrier interference (ICI), the group-pilot design criterion is used based on the minimization of mutual coherence of the measurement. Finally, an efficient pilot pattern optimization algorithm is proposed by a dual layer loops iteration. The simulation results show that the new method uses less pilots, has a smaller bit error ratio (BER), and greater ability to deal with Doppler frequency shift than the traditional method does.