A beam space ML algorithm for radar low-angle tracking

A beamspace maximum likelihood (ML) algorithm is proposed in attempting to solve the radar low-angle tracking problem. A uniform linear radar antenna array is divided into several nonoverlapping subarrays with equal numbers of sensors and identical beampatterns. The algorithm is then applied to the sub-array output to estimate the directions of both the direct and specular signals. The key advantage here is that, since the directions of both the direct and specular signals change slowly with time (or distance) in the low-angle tracking situation, the estimates of the directions based on the previous block of array data can be used together with the current block of data in estimating the present signal directions. Thus no iteration is actually required. The computation is therefore greatly reduced. The performance of the algorithm was tested by computer simulations