Sparse Multipath Channel Estimation for SC-FDE System with Unknown Sparsity

摘要

The performance of Single Carrier Frequency-Domain Equalization (SC-FDE) system is affected by the precision of channel estimation results. For a pilot-assisted SC-FDE system transmitting over sparse multipath radio channels, we discuss that the sparse multipath channel estimation can be formulated to be an underdetermined compressed sensing (CS) problem or an overdetermined sparse system identification problem under different transmission parameters. For the sparse system identification problem setting, we propose to use Zardoff-Chu sequence as the pilot sequence to form a deterministic circulant Toeplitz observation matrix for signal recovery. To address the unknown sparsity in practical applications, Sparsity Adaptive Matching Pursuit (SAMP) algorithm is investigated to reconstruct the channel impulse response (CIR) instead of other greedy sparse recovery algorithms that need a priori knowledge of channel sparsity. The simulation results demonstrate that using the designed observation matrix and the SAMP algorithm for sparse channel estimation with unknown sparsity achieves better performance than the traditional Least Squares (LS) channel estimation algorithm with reduced length of pilot sequence in the SC-FDE system over 3GPP radio channels.