A computationally efficient method for joint direction finding and frequency estimation in colored noise

The problem of simultaneously estimating the temporal and spatial frequencies of multiple signals arriving at an antenna array is considered. The problem is relevant in a variety of applications, including pulsed Doppler radar, multipath parameter estimation and synthetic aperture radar imaging. The proposed approach has a wider applicability than existing schemes. In particular it accommodates the case of (1) arbitrary array structures, (2) temporally and spatially colored noise, and (3) general single-pole ARMA signals. Despite the generality of this approach, the computational complexity is comparatively small. The temporal signal model (e.g. frequencies) is first estimated using a state-space realization approach. The arrival angles are then found one-by-one using 1-D searches. The resulting estimates are automatically paired.