Pilot decontamination in massive MIMO systems: Exploiting channel sparsity with pilot assignment

This paper deals with the pilot contamination problem for massive MIMO systems. Considering the sparse nature of channel impulse response inherent in wideband systems, the paths of channels of the desired and interference users hardly arrive at the same time, which allows most paths of desired channel to be distinguishable from the paths of interference channels in time-domain. Based on this observation, we first estimate the power delay profile (PDP) of the desired channel with the contaminated channel estimate, from which we acquire the delay of each path of the desired channel. By extracting the corresponding channel components from the contaminated channel estimate, a clean channel estimate can be obtained. To reduce the impact of pilot contamination on the estimated PDP, we propose a pilot assignment method among adjacent cells to randomize the interference over successive uplink frames. Simulation results demonstrate substantial sum rate gain of the proposed approach over existing methods.