大规模PAF接收机数字化设计与信号处理架构研究

相控阵馈源（PAF）通过电子扫描形成多个同步波束，能增大望远镜的视场，提高巡天效率。大规模PAF馈源排列密集、数量众多，如采用传统的超外差式设计，模拟前端将非常复杂。采用数字接收机设计，靠近PAF馈源数字化，可以减少模拟链路、减小空间和重量、降低复杂度，非常适合大规模馈源阵列的紧凑化设计。大规模PAF信号速率高、通道多，对采样芯片的性能要求高，靠近望远镜焦点位置的数字化对电子器件的电磁辐射极其敏感，波束合成处理算法复杂、实时性要求高，对系统的处理能力挑战大。本项目将面向大口径射电望远镜，研究高速率、低功耗的射频信号直接采样技术，开发高性能数字混频和滤波模块，研究不同观测模式下的波束合成与校准算法，基于混合架构设计分布式信号处理与波束合成网络，开发高效并行计算与共享内存管理引擎，优化波束合成网络拓扑结构，为大规模、宽带PAF提供可行、可靠、经济、先进的数字化设计和实时信号处理方案。

Phased array feed (PAF) can form multiple synchronous beams through electronic scanning, which can increase the field of view(FoV) of the telescope and improve the survey efficiency. The large-scale PAF is densely arranged and numerous, if traditional superheterodyne structure is adopted, the design of analog front-end will be very complicated. Digital receiver design, close to the PAF feed digitization, can reduce analog links, reduce space and weight, degrade complexity, very suitable for the compact design of large-scale feed array. The large-scale PAF has high signal rate, numerous channels, which requires high performance of the sampling chip. The digitization near the focus of the telescope is extremely sensitive to the electromagnetic radiation of electronic devices. The PAF beam-forming algorithm is complex and high real-time requirements, which pose a great challenge to the signal processing.This project will research on the high sampling rate, low power consumption RF-direct sampling technology toward the large radio telescope, research on the beam-forming and calibration algorithm under different observation mode, develop high performance digital mixing and filtering module, design distributed signal processing and beam-forming network based on heterogeneous architecture, develop efficient engine for parallel computing and shared memory management, optimize the topology of beam-forming network. Finally, a feasible, reliable, economical, advanced full digital design and real-time signal processing solutions for large-scale, broadband PAF will be presented.