Abstract:

In the Massive multiple input multiple output (MIMO) multiple cells system, the reverse training pilots sequences of terminals, which are from different cells, are non-orthogonal due to the reuse of these pilots between cells with the mode of time division duplex. The non-orthogonality of the training sequences leads to pilots contamination, which makes the forward channel estimation and pre-coding inaccuracy and blocks the improvement of the system. An approach is proposed, which is based on the cells crossed classification orthogonal identity and pilots power control, to reduce the pilots contamination. All the cells are classified crosswise and orthogonal identifications are distributed to the terminals and base stations in different kinds of cells. Further, terminals add the allocated identification to the head of their pilots so that the base stations can identify the sources of the received pilots. Pilots power control is applied to the terminals in the same kind of cells to reduce the influence for the base station in intersecting cells. Theory analysis and simulations results verify that the proposed approach is not only simple but also effective to alleviate pilots contamination.