3D MIMO Beamforming In 5G Communication Systems

Mobile communication is coming to the fifth generation (5G). In the age of 5G,3D MIMO of massive antenna is one of the highlighted technologies. As massive 3D MIMO increases vertical dimension in terms of space domain and adopts massive antenna array, on the one hand, the resolution and anti-interference ability has been greatly improved, on the other hand, the challenge of beamforming is to be taken. Therefore, the beamforming of massive 3D MIMO has become a focus of technology. In this thesis, the beamforming based on wolf pack algorithm (WPA) is analyzed and studied under the background of the 3D MIMO in 5G.WPA is a kind of algorithm which is suitable for multi-objective optimization. And the weights of beamforming can be optimized efficiently by using WPA. While deriving optimal weights, we improve the WPA and propose an algorithm, the Improved WPA in this thesis. The main improvements are listed below:firstly, the convergence speed of WPA is accelerated by using good point set uniform initialization methods; secondly, the mobility of group is developed by using adaptive steps; finally, the diversity of group search is enhanced by moving the excessive concentration of the individual to diversify, which means local optimum can be avoided.The existence of steering vector error is another challenge in the beamforming process. If consider WPA in correcting steering mismatch factor, the steering mismatch is reduced to a great extend. Using above process as well as sidelobe constraint, sidelobe interference can be restrained effectively by beamforming.The joint optimization among beamforming weight and antenna array element spacing is completed based on WPA. Both access and disconnection of array element are controlled by switch to realize the adjustment of the array element spacing in the base station. In the angel domain beamforming, lower sidelobe jamming and higher accuracy pointing are get by using joint optimization based on WPA. While as in the channel precoding beamforming, more match between weights and antenna array is realized. Therefore, the channel correlation among different users is reduced, namely, reducing the co-channel interference.Finally, computer simulation experiments are performed and the results show that guide mismatch is effectively corrected, as well as weights and antenna are optimized jointly. Besides these improvements, the simulation results also demonstrate that sidelobe interference and co-channel interference have been reduced. All of the above means the effect of beamforming is improved by using the proposed algorithms.