Research On Efficient Beam Training And Beam Tracking Algorithms In Millimeter-Wave Communication

With continuous development of the Internet of Things wireless communication data traffic is growing rapidly and the millimeter-wave wireless communication technology can meet the needs in system capacity and data rate for future wireless communication systems.However,millimeter-wave have serious propagation loss,and it is necessary to use joint transceiver beamforming technology to form a narrow beam with strong directivity to provide a large antenna array gain.Beam directional transmission brings two new problems: the high directivity of the beam increases the delay of beam alignment during the initial signal access;the directional transmission link is easily interrupted by the dynamic change of the communication environment.Therefore,this paper studies the efficient beam training and tracking technology in millimeter wave communication.First of all,this paper summarizes the research results of existing beam training.In order to solve the problem of the time-consuming in beam training and low gain at the beam junction,a two-stage codebook design and beam training algorithm in2 D MIMO are proposed and extended to 3D MIMO scenarios.The two-stage codebook and algorithm are divided into primary and auxiliary two parts: the main codebook in 2D MIMO uses a hierarchical codebook,and main search uses an improved binary search algorithm based on signal-to-noise ratio judgment,which greatly reduces the beam search delay;Using the expanded auxiliary codebook,the auxiliary search adopts an improved exhaustive search strategy to refine the beam and improve the gain at the beam junction.The codebook and algorithm design in3 D MIMO is an extension of the two dimensions of elevation and azimuth angle: the3 D codebook design is decomposed into the Kronecker product of the codebooks in the two dimensions;the 3D beam search sequentially searches in two dimensions.In each dimension the search process performs a primary and an auxiliary search using an improved binary search algorithm.The main search and auxiliary search canreduce the search delay and increases the gain at the beam junction respectively.The simulation results show that the codebook design and beam training algorithm in 2D and 3D scenarios achieve higher beam gains with lower beam search delays and good beam search accuracy.Secondly,to solve the the problem that the millimeter-wave directional beam communication is prone to have beam misalignment in rapidly changing wireless communication environment,this paper proposes a beam tracking algorithm based on Adaptive Unscented Kalman Filter(AUKF)principle.The algorithm uses the basic Unscented Kalman Filtering principle to estimate the state of the system.By introducing an adaptive adjustment factor,the system can adaptively adjust the prediction and observation covariance matrix to improve the estimation accuracy and convergence speed when the system has abnormal disturbances.Simulation results show that compared with the existing beam tracking algorithms,the algorithm in this paper not only has lower complexity and smaller tracking error,but also can maintain better estimation accuracy when the beam angle change rate is larger.The proposed beam tracking algorithm is more suitable for dynamic millimeter-wave communication scenarios.