Research On High Resolution DOA Estimation In Hybrid Massive MIMO Systems

Direction of Arrival(DOA)estimation is an important topic in the field of array signal processing.Most of DOA algorithms,including subspace algorithms,obtain the direction of arrival by the analysis of the spatial covariance matrix(SCM)of the received signal.In order to reduce the cost of radio frequency(RF)chains at millimeter-wave bands,analog beamformer is adopted for massive MIMO systems operating at millimeter-wave bands,which leads to a analog-digital hybrid massive MIMO structure.However,in hybrid massive MIMO systems,received signals are combined in analog domain before sent to the digital receiver.As a consequence,the receiver cannot obtain the received signals at the antennas.Therefore,the traditional sample average approach cannot be used to obtain the SCM,so that the super-resolution DOA estimation algorithms cannot be directly applied to the hybrid massive MIMO systems.To utilize the high resolution property,a beam sweeping algorithm for SCM recon-struction has been proposed in massive MIMO systems with single RF chain.Although it can show theoretical insights on the SCM reconstruction problem,practical issues have not been considered yet.In view of the research deficiencies,this paper studies high-precision DOA estimation in the hybrid massive MIMO systems.The specific work is listed below:(1)Based on the previous work,a novel algorithm is proposed for SCM reconstruc-tion in partially connected hybrid massive MIMO systems.The overall SCM is divided into a number of sub-SCMs to enable SCM reconstruction with multiple RF chains.Then,a low-complexity algorithm is presented based on a large number of repeated entries exist in the SCM.In addition,the selection of the predetermined DOAs has also been optimized and the matrix inversion in the low-complexity algorithm can be further avoided.Sim-ulation results have shown that the proposed approach can achieve better performance than existing baselines,and the performance of MUSIC algorithm can be also improved accordingly.(2)Furthermore,an adaptive reconstruction algorithm is proposed to avoid using di-rect matrix inverse and adapt to the time-varying systems.During the adaptive iterative procedure,the reconstructed SCM can be updated fast by few samples in each adaptation cycle.Moreover,it also performs well with a narrow range of sweeping angles.Simula-tion results show the SGD algorithm presents a satisfactory performance once appropriate provisos are selected.(3)In fully connected hybrid massive MIMO systems,a fast beam sweeping algo-rithm is proposed for SCM reconstruction.In the presence of multiple RF chains,the pre-determined DOA are divided into several sub-groups,with each sub-group corresponding to one RF chain.Then,by beam sweeping simultaneously on all RF chains,the predeter-mined DOAs can be swept in a parallel manner,so that the SCM can be reconstructed fast.Then,an iterative updating procedure is further developed,in which the reconstructed SCM can be updated immediately as new beam sweeping results are available.Com-puter simulation is also conducted to demonstrate that the proposed approach can greatly accelerate the reconstruction and updating procedures for SCM reconstruction.