DESIGN OF CHANNEL FEED-BACK CODEBOOK AND ADDRESSING POWER LEAKAGE PROBLEM IN MM- WAVE MASSIVE MIMO SYSTEM WITH LENS ANTENNA ARRAYS

Abstract

Objective The modernistic theory of beamspace MIMO, which involves millimeter wave (mm-Wave) and massive MIMO system depending on lens antenna array (LAA) can efficiently decrease the count of power consuming radio frequency (RF) chains. Hence, it is perceived as an optimistic approach for the 5G technology and its successors. Research in beamspace (BmSp) MIMO has yet to be taken into debate the issues of power leakage in channels, therefore resulting in remarkable deterioration in SNR and the sum-rate (SR). A precoding technique for beam aligning and to tackle power leakage issue has been proposed. Initially a network using phase shifters has been designed, which helps RF chains choose beams and collect leaked power in beamspace MIMO. An algorithm based on rotation for precoding, for the available phase shift networks is proposed to align the gains of the channels in the same direction to maximize SNR received by the users. In systems that employ frequency division duplexing (FDD), it is necessary that the channel is fed back to the base station (BS) via feedback based on codebook. There is no devoted codebook for the LAA dependent mm-Wave systems. The codebook design is proposed to address the gap for these systems. In this codebook, initially a large dimension vectors based on recent idea involving angle coherence time are generated. Then relying on obtained vectors in channel subspace, codebook is created by considering lens and beam selection. After that the channel is quantized and given as feedback to the BS. Results Simulation results conveys that the overhead in feedback is proportional to dominant paths for each user which are few and the proposed approach with uniform linear array (ULA) attains the near optimal performance in SR comparing to ideal case of zero power leakage and also obtains superior Energy efficiency (EE) than existing single beam precoding and Multiple beam via multiple RF (MBMRF) structure precoding methods.