An efficient joint direction of arrival and frequency ML estimator

Joint space-time estimation is an area of current interest in array signal processing. By using both direction of arrival (DOA), and frequency as discriminants superior performance may be achieved in practice than if only DOA is used. However, superresolution algorithms when adapted to the space-time, case become very computationally expensive. The deterministic maximum likelihood estimator (MLE), is known to perform well, being both accurate and robust. A reformulation of the MLE, DOSE, has been derived which is unique in allowing the FFT lemma to be applied to the MLE. We show how the DOSE estimator may be adapted for joint space-time estimation, and the 2D FFT applied, resulting in a massive reductio of the computation required for a joint space-time estimator. We also present real data which has been processed with the 2D fast DOSE estimator, demonstrating it´s ability in the difficult HF environment.