Optimal Blind Calibration of Uniform Linear Arrays

Calibration of uniform linear arrays remains an important task that enables proper functionality of almost all spatial array processing downstream operations. We develop a computationally and statistically efficient method for blind calibration, i.e., without known calibration signals, that yields the maximum-likelihood estimate (MLE) of the gain and phase offset parameters. We use a computationally lean, but consistent initial estimator, and refine it via Fisher’s scoring algorithm, for which we derive the associated Cramér-Rao Bound (CRB) of the model. Simulation results demonstrate that our method converges to the MLE, attains the CRB, and is superior to the recently proposed reduced-maximum-likelihood optimally-weighted least squares method.