Research On DOA Estimation Algorithm For Massive MIMO System

With the rapid development and gradual popularization of the fifth-generation communication system,large-scale multiple input multiple output(MIMO)technology has also become the focus of research.Accurate estimation of the Direction of Arrival(DOA)is the key to realize beamforming technology,and the DOA estimation is also widely used in wireless communication,radar,positioning technology.However,a great quantity of antennas are deployed in massive MIMO environment,which brings extremely high complexity to the DOA estimation,therefore,reducing the complexity but also ensuring the accuracy become the research focus.This thesis will further research on the DOA estimation algorithm based on uniform linear array and uniform rectangular array,the main work of this thesis is as follows:First,by reading numbers of literatures,summarize the current research status of DOA estimation algorithms,and focus on the research and comparison of DOA estimation algorithm based on the estimation of signal parameters via rotational invariance techniques(ESPRIT)and the multiple signal classification(MUSIC)algorithm,and summarize the advantages and disadvantages of the existing algorithms.Secondly,the one-dimensional DOA estimation based on uniform linear array is studied,and a variety of ESPRIT-based DOA estimation algorithms are compared and the computational complexity is analyzed.An improved algorithm is proposed for the problem of high precision but high computational complexity of the MUSIC algorithm.The ESPRIT algorithm is used to estimate the initial angle,and then the local spectral peak search is used to update the initial angle estimate.The experimental results show that the performance of this algorithm tends to be consistent with that of the MUSIC algorithm,but to a great extent reducing the computational complexity.Thirdly,the two-dimensional DOA estimation algorithm based on uniform area array is studied.Aiming at the problem that the traditional ESPRIT algorithm is prone to angle mismatch and of high computational complexity,an improved algorithm is proposed.The algorithm does not need to reconstruct the received signal,but estimates the angle through the rotation invariance between the four sub-matrices.Experimental results have verified that the improved algorithm can achieve effective angle estimation when signal-to-noise ratio and snapshots are low,and under different antenna formations the improved algorithm can maintain great performance;Aiming at the high computational complexity of search in the cascaded MUSIC algorithm,an improved algorithm is proposed.First,use the improved ESPRIT algorithm to estimate the parameters of DOA,and then two times of one-dimensional local searches are performed to update the initial value.It is verified by simulation that the performance of the algorithm is consistent with the cascaded MUSIC algorithm,but the complexity is further reduced.Then,research on the DOA estimation algorithm of coherent source based on ESPRIT.In the one-dimensional case,to solve the problem of reducing the effective aperture of the ESPRIT-Like algorithm,an improved Toeplitz matrix reconstruction algorithm is proposed.Experimental results show that the improved algorithm has the best performance when estimating the DOA of fully coherent signal,and can estimate more coherent sources.In the two-dimensional case,the spatial smoothing technology is combined with the unitary-ESPRIT algorithm and the improved two-dimensional ESPRIT algorithm,and further use the propagation operator(PM)idea to reduce the computational complexity.Finally,the completed work in this paper is summarized and the future work is prospected.