학술논문

FPGA Implementation of Efficient Beamformer for On-Board Processing in MEO Satellites
Document Type
Conference
Source
2023 IEEE 34th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) Personal, Indoor and Mobile Radio Communications (PIMRC), 2023 IEEE 34th Annual International Symposium on. :1-7 Sep, 2023
Subject
Communication, Networking and Broadcast Technologies
Computing and Processing
Robotics and Control Systems
Signal Processing and Analysis
Earth
Measurement
Power demand
Satellite broadcasting
Discrete Fourier transforms
Low earth orbit satellites
Planetary orbits
beamforming
fast Fourier transform (FFT)
FPGA implementation
MEO satellites
power estimation
Language
ISSN
2166-9589
Abstract
Medium Earth orbit (MEO) constellation is an appealing solution between geostationary equatorial orbit (GEO) and lower Earth orbit (LEO) in terms of latency and number of satellites required. On-board processing of digital beam-former in MEO satellites is an efficient solution for achieving wider bandwidth, increased flexibility, and lower latency. Power constraints, however, make it impractical to digitally create thousands of beams at once. In this paper, area-power efficient digital beamformer architectures are proposed considering key metrics of a typical MEO scenario. The proposed efficient digital beamformer is comprised of a sparse-matrix-based user selection, a 2D discrete Fourier transform (DFT)-based digital beam generation, which is implemented by a fast Fourier transform (FFT) algorithm, and a spatial windowing module for selecting the antenna pattern. Furthermore, architectures of digital beam-former using conventional 2D-FFT approach, fully unrolled 2D-FFT, and an area-power efficient twiddle factor (TF) quantized fully unrolled 2D-FFT are proposed. The spatial windowing architecture concerning 10 × 10 radio frequency chains and sparse matrix architecture for user selection is also proposed. The proposed architectures are implemented targeting Virtex ultrascale FPGA and the area-power utilization is reported. It is noticed that more than 50%-reduction in area and power is achieved with the beamformer incorporating the proposed TF quantized fully unrolled 2D-FFT.