Location-aware channel estimation and pilot assignment for massive MIMO

A location-aware channel estimation algorithm is proposed for massive multiinput multi-output (MIMO) systems, to reduce the inter-cell interference caused by pilot contamination. Specifically, by utilizing the property of the steering vector, a fast Fourier transform based post-processing is introduced after the conventional pilot-aided channel estimation for mitigating pilot contamination, which only imposes a modest extra computational cost. Our asymptotic analysis proves that this post processing is capable of removing pilot contamination from the interfering users with different angle-of-arrivals (AOAs). Since in practice the AOAs of some users may be similar, we further propose a location-aware pilot assignment method to improve the overall system performance. Simulation results demonstrate that the proposed scheme can dramatically reduce the inter-cell interference caused by the re-use of the pilot sequence and improve the overall system performance significantly.