ML method at subarray level with low calibration cost

Most of super-resolution direction finding methods presented until now are based on the element outputs of the array. 2-D subarray level super-resolution direction finding method has an important application in phased array radars. We study ML (Maximum Likelihood) method at subarray level for coherent signal sources and propose the corresponding signal model. Two kinds of array manifold are introduced which are denoted as TAM (True Array Manifold) and DSAM (Direct Simplified Array Manifold) respectively. ML method at subarray level based on TAM has to calibrate the whole array while the method based on DSAM only needs to determine the phase centers and gains accurately. Method based on DSAM can reduce the array calibration cost largely, but the available direction finding range is 3dB bandwidth of subarray centered at the look direction. We can estimate direction in any possible space range with beam scanning and reduce computation cost of multi-dimensional search of ML. Simulation results express the validity of the proposed methods.