Synchronization of Distributed Space Communication Reflector Array Antennas via Centralized Stochastic Gradient Descent

Abstract

Deep space communications are often challenged by low signal-to-noise ratios (SNRs). While antenna arraying offers a promising solution for SNR enhancement, it faces dif ficulties in synchronizing distributed reflector antennas. Imperfect synchronization can significantly impair system performance. In this paper, we propose a centralized learn ing approach to synchronize received signals across an array of reflectors. A stochastic gradient descent (SGD) algorithm is employed to minimize the error between a reference antenna and the others, enabling coherent signal combination and resulting in substantial SNR gains. Theoretical analysis and simulation results demonstrate that the performance of the proposed array is comparable to that of a single large reflector antenna. Specifically, at 12 GHz, an array of four 10-m antennas achieves 65 dB gain, comparable to a 20-m reflector antenna operating at the same frequency. Notably, constructing such an array is generally more cost-effective than manufacturing a single large reflector.