In 5G/6G systems, massive Multiple Input Multiple Output (MIMO) and large bandwidth resources are enabled, which offer opportunities for precise azimuth angle of arrival, elevation angle of arrival and delay estimation. Thus, accurate outdoor localization can be attained based on the estimated azimuth-elevation-delay 3-tuple by geometric relationship. Su-per resolution delay and angle estimation can be achieved by tensor ESPRIT algorithm. However, the estimated azimuths, elevations and delays are not intrinsically paired in standard tensor ESPRIT algorithm. In this paper, a novel Joint Azimuth-Elevation-Delay Estimation (JAEDE) algorithm via Progressive Subspace Matching (PSM) is proposed to address the challenge of unpaired azimuth-elevation-delay estimations in standard tensor ESPRIT algorithm. Specifically, a new tensor unfolding technique is designed and the subspace structure of the unfolded matrix is studied. Then, the pairing among azimuth-elevation-delay estimations can be reformulated as bipartite matching problems based on the unfolded subspace structure, which can be progressively solved. In the numerical simulations, it can be observed that the parameter estimation error of proposed PSM-JAEDE algorithm is significantly reduced compare to the conventional algorithm, which indicate its promising application in massive MIMO system.
Joint Azimuth-Elevation-Delay Estimation in Massive MIMO via Progressive Subspace Matching
2024-06-24
774747 byte
Conference paper
Electronic Resource
English