Single Snapshot Direction-Of-Arrival Estimation via Signal Fitting
Hua Yang, Yu Zhang, Ke Miao · 2024
A novel single snapshot direction-of-arrival (DOA) estimation method is proposed in this paper. Note that the observation at hand is contaminated by noise, we propose a signal fitting framework by using least squares (LS) together with the decomposition property of the interested signal to first retrieve the noise-free interested signal. This signal fitting framework can be equivalently transformed into a least squares programming with low-ranked positive semidefinite Toeplitz-Hankel constraint, which can be efficiently implemented via the alternating direction method of multipliers (ADMM). Moreover, the unknown DOAs can be obtained from the constructed Toeplitz matrix by employing classical Vandermonde decomposition techniques. Compared with the existing compressed sensing based DOA estimation algorithms, the proposed method does not need the noise statistics. Numerical simulations demonstrate the effectiveness and outperformance of the proposed method.