A Low Complexity Adaptive Algorithm for Eigenspace-Based Two-Dimensional Direction of Arrival Tracking

In this paper, we present a low complexity, yet accurate adaptive algorithm for the tracking of two-dimensional (2-D) direction of arrival (DOAs) based on a uniform rectangular array (URA). The new algorithm is a novel hybrid of tracking and beamforming processes by making use of three stages of one-dimensional (1-D) DOA tracking algorithms — in a hierarchical tree structure — to determine the two DOA components iteratively in a coarse-fine manner. In between every other 1-D DOA tracking algorithm, a complementary orthogonal beamforming process is invoked to partition the incoming signals into appropriate groups to enhance the tracking accuracy. Since the new algorithm only involves the 1-D subspacebased DOA tracking algorithm, the overall complexity is substantially less than the direct two-dimensional (2-D) extension of the existing 1-D DOA tracking algorithms, which requires an update of higher-dimensional vectors followed by a higher-dimensional eigendecomposition or a 2-D search. Furthermore, with the tree-structured DOA tracking scheme, the tracked 2D DOA components are automatically paired without extra computational overhead. Furnished simulations show that the new algorithm can provide satisfactory tracking performance in various scenarios. key words: fast algorithm, 2-D DOA tracking, 2-D Unitary ESPRIT, azimuth-elevation tracking, smart antennas