Calibrating Reception Characteristics of Antenna Elements and RF-Frontend Within a Phased Array Using Machine Learning Techniques

Many spatial signal processing applications, such as Direction of Arrival (DoA) estimation or various beamforming techniques, require precise information about the gain and phase response of the multi-antenna RF system. Theoretically, this can be calculated deterministically from the spatial arrangement of the antenna elements and the receiver / transmitter. However, several types of gain and phase mismatches occur along the RF chain. Some types vary with time, others vary with the DoA. Calibration of these mismatches is a challenging problem. In this paper, a novel method is proposed to calibrate the reception characteristics of individual antenna elements while also addressing time-varving mismatches.