Estimation of the spatial and temporal frequencies of multiple sinusoids using a sparse linear array

In this paper, we extend our previous sparse linear prediction (SLP) method [1,2,3] for two-dimensional (2-D) applications. The temporal and spatial frequencies of multiple sinusoids are simultaneously estimated in pairs, rather than separately, from the outputs of a sparse linear array. Possible ambiguities due to the large interelement spacings can be resolved by properly choosing the interelement spacings and by incorporating different SLP polynomials, the coefficients of which can be calculated in parallel. The temporal and spatial frequencies are simultaneously estimated from the minima of a 2-D objective function, which can be calculated using a 2-D DFT of the properly mapped 2-D SLP coefficient matrix