Sparse inverse fast Fourier transform‐based channel estimation for millimetre‐wave vector orthogonal frequency division multiplexing systems

Abstract

Millimetre-wave propagation is a promising broadband transmission technology for future fifth generation mobile communication systems. For a vector orthogonal frequency division multiplexing system, the authors investigate the millimetre-wave propagation through a sparse multipath channel in a sense that it has a large time delay spread but with only a few non-zero taps. By exploiting the sparse nature of millimetre-wave channel, any sparse multipath channel can be characterised by the multipath delays and their corresponding channel coefficients. They first study an ideal case that the pilot signals are transmitted through a sparse channel without noise, and an exactly sparse inverse fast Fourier transform (SIFFT) algorithm is performed to estimate the non-zero channel taps with reduced complexity. Then, they consider a more practical scenario that the pilot signals through a sparse channel with noise interference, and an approximately SIFFT algorithm is employed to estimate the effective channel taps, while the remaining small coefficients interfered by noise can be wiped out. Through numerical analysis, they demonstrate that the proposed SIFFT algorithms can reduce the computational complexity while keeping the root mean squared error of channel estimation at a low level.