The E ects of Array Calibration

This paper studies the eeect of array calibration errors on the performance of various DF (direction nding) based signal copy algorithms. Unlike blind copy methods, this class of algorithms requires an estimate of the directions of arrival (DOAs) of the signals in order to compute the copy weight vectors. Under the assumption that the observation time is suuciently long, the following algorithms are studied: classical beamforming, least squares, total least squares, linearly constrained minimum variance beamforming, and structured stochastic estimation. Expressions for the mean-square error of the signal estimates are derived as a function of the calibration errors for both the case where the DOAs are known precisely and for the case where the DOAs must be estimated.