Research On Out-of-band Information Aided Millimeter Wave Communication Technology

In order to meet the bandwidth requirements of the next generation mobile com-munication system,5G introduces the millimeter wave frequency band as one of the communication bands.The introduction of millimeter wave frequency band allows the large-scale antenna arrays to be applied,which can help meet the requirements of rapidly growing mobile data.However,due to the characteristics that the propagation of mmwave suffers severe loss and is effected greatly by natural environment,it is nec-essary to compensate the path loss with massive MIMO and corresponding technologies such as beamforming.These technologies can provide path gain,while a lot of computa-tion and time overhead are required to complete the entire process.The goal of existing research on beamforming and channel estimation is to balance communication perfor-mance and overhead,now we consider the role of out-of-band information in mmwave communication to reduce overheead or improve communication performance.Most of existing research on out-of-band information aided communication use the information directly without considering channel features.This thesis studied location information and sub-6G information aided mmwave communication algorithm based on channel spatial consistency and frequency domain correlation.Main work of this thesis is as follows1.We studied spatial consistency of millimeter wave channel,and applied lo-cation information in aiding mmwave beamforming to reduce overhead of algorithm.First,this thesis studied spatial consistency of millimeter wave channel and obtained the law how millimeter wave channel parameters change in space.According to the law,we analyzed the corresponding distribution of beamforming precoders and combiners.With this distribution,codewords selection problem of beamforming can be transformed as a space division one.Support vector machine was used to solve this problem,and we designed corresponding custom kernel function through the derivation and approx-imate calculation of beam sweeping optimization problem based on the theory of high dimension linear separable.And the kernel function was used in SVM,which formed the final beamforming algorithm.Simulation results showed that this algorithm could greatly reduce the overhead while keeping the communication performance almost the same.2.Utilizing the correlation of communication channel parameters in frequency do-main,we studied sub-6G aided mmwave channel estimation.Considering the character-istic that multi frequency band communication systems are joint deployed in 5G,sub-6G aided millimeter wave communication is practical.Study how communication chan-nel parameters varied in frequency domain,which is called frequency independency or frequency domain correlation,showed that some parameters varied little at different frequencies which can be leveraged in mmwave algorithm.We mainly considered com-pressed sensing channel estimation utilizing sparse feature of mmwave channel,and the corresponding orthogonal matching pursuit compression recovery algorithm.In order to apply sub-6G information,this thesis studied compression recovery algorithm based on prior information and the case applied in OMP,which is called LW-OMP.Then sub-6G information is extracted as prior information of mmwave channel according to frequency domain correlation,combined with LW-OMP algorithm we got the finnal channel estimation algorithm.Simulation results showed that with the same measure-ment times,the proposed algorithm can provide great gain at lower SNR.