In-Orbit Estimation Method of NGSO Satellite-Borne Antenna Pattern

Non-Geostationary Satellite Orbit (NGSO) constellations play an increasingly important role in modern communication networks. Accurate estimation of onboard antenna patterns is essential for ensuring communication quality. Using the estimated antenna pattern, we can understand the radiation intensity of the antenna in different directions, thereby optimizing the antenna pointing and beamforming strategy to maximize the signal strength in the target area. This paper introduces a time-sharing multi-beam pointing pattern estimation method and uses the deep convolutional generative adversarial network (DCGAN) model to complete the sidelobe information of the antenna pattern. The on-orbit estimation results are in good agreement with the real values.