Maximum likelihood estimation of the arrival direction of a signal with known waveform

The maximum likelihood estimate of the direction of arrival (DOA) of a single narrowband plane wave impinging on a uniform linear array is considered. It is shown that incorporating knowledge of the signal waveform in the estimator improves the accuracy of the direction estimate. Although for the general case of multiple signals, maximization of likelihood functions requires a computationally expensive multidimensional search, it is found that for the simple case where there is only one incident signal, a closed-form solution exists. For this case, the performance of two estimators-one that utilizes the known signal waveform and one that does not-is compared. The Cramer-Rao (CR) bounds for these two cases are presented, and the performance of the two estimators is compared to their CR bounds.<>