This paper, alongside its companion paper, introduces a two-dimensional (2D) (spatial-temporal) smoothing approach within a recursive subspace framework. This approach aims to estimate the direction-of-arrival (DOA) of a moving target with high mobility in an empirical nonstationarity environment. A blockwise 2D smoothing method is described in the companion paper, where account is taken of the spatial and temporal selectivity inherent in signal channels between a fast-moving target and receiver equipped with multi-element array antennas. This scheme can improve the DOA detection accuracy, however, at the cost of increased computational effort. To reduce the computational burden, this paper proposes an efficient use of the low-rank adaptive filter (LORAF) for dynamic subspace tracking, further improving the DOA estimation accuracy. It is demonstrated that through numerical simulations of high-mobility setting, the LORAF technique-based DOA estimation can achieve superior performance. This showcases its potential for DOA estimation in high-mobility applications such as unmanned aerial vehicles and commercial aviation.
