Fast Calibration of Beam Directions Based on RNN for Phased Array Antennas

The accuracy of beam directions is a key characteristic of phased array antennas, which requires precise calibration. Current calibration methods require sequential measurements of beam directions at all angles within the field of view by rotating the turntable mechanically, which is very time-consuming. This paper proposes a fast calibration method of beam directions based on Recurrent Neural Network (RNN) for phased arrays. The RNN is used to learn the characteristics of beam direction errors at measured angles and predict of those errors at other angles. Experiments show that, to meet the same calibration accuracy, the proposed method only measures beam directions at 1/10 of the angles that are measured in the current method. Hence, the proposed method can significantly enhance the efficiency of beam direction calibration.