An Efficient Hardware Generator for Massive Non-Stationary Fading Channels

Abstract

In this paper, a discrete non-stationary multipleinput multiple output (MIMO) channel model based on the sum of linear frequency modulation (SoLFM) method is proposed. The new model is suitable for generating non-stationary fading with continuous phase and accurate Doppler frequency. In order to implement the proposed model with large-scale antennas by fieldprogrammable gate array (FPGA) platform, an efficient coordinate rotation digital computer (CORDIC) method is proposed. By introducing a full parallel pipeline architecture, rotation factor state, and domain folding technique, the new generator can significantly reduce the hardware resource usage and meet the requirement of large-scale channel emulation. The measurement and analyze results show that the statistical properties, i.e., the probability density function (PDF) and autocorrelation function (ACF) of generated channels provide a good agreement to the theoretical ones.

Publication
IEEE Global Communications Conference
Shuangyi Yan
Shuangyi Yan
Senior Lecturer in High Performance Networking & Optical Network

My research focuses on AI-driven automatic dynamic optical networks with flexible network functions and fast network reconfigurations.

comments powered by Disqus