Research On The Beam Selection Strategies For Orthogonal Random Beamforming In Sparse Networks

As the smart phone and mobile Internet attract more interests from the world, thedemands of multi-media services are growing fast. To further provide a much higher rate fordata transmission in the wireless communication system, the MIMO technology is applied tothe LTE project, which can multiple the transmission rate and improve the quality of the datalink without the additional power and spectrum resource. In the multi-user MIMO system,the downlink BC is considered to be the bottleneck of data transmission. To release theperformance of the BC, however, the traditional adaptive scheduling schemes fail to avoidthe requirement of the full CSI at the BS, which makes the feedback channel overloaded.Orthogonal Random Beamforming scheme requires each user just feeds the best SINRand the corresponding beam index back to the BS, which eases the system pressure by thepartial CSI and low-complexity calculation. Whereas for a large number of users thesum-rate of ORBF exhibits an identical growth rate as that of dirty paper coding, theperformance in sparse networks is known to be severely impaired. By constructing thesystem model and assessing its performance in terms of the sum-rate with asymptoticanalysis, we find that the ORBF scheme needs to schedule less active beams in theinterference-limited system to avoid the performance depressing.In order to improve the performance of ORBF scheme in the sparse networks, wepropese the new adaptive beam selection strategies based on the CSI reconfiguration. Byusing the proposed strategies, the BS is able to approximately recover the perfect CSI withlimited feedback, and selects a better active beam-user set for data transmission. Accordingto the simulation results, the proposed strategies can adaptively reduce the active beam insparse networks with reconfiguration CSI, which can increase the system throughput with agood anti-interference performance. Because the process of CSI reconfiguration is based onthe order statistic of the SINR with different beams, the BS decreases the requirement ofextra CSI feeddback, which keeps the superiority of ORBF on CSI feedback.