Modeling Dynamic Rural Clutter Channels With Generative Adversarial Networks

Time variant rural clutter channels have major effects on radar measurement results or communications, however their amplitude, Doppler and phase characteristics remain difficult to predict. To implement effects of rural bistatic radar channels in a simulated environment, the transfer functions collected from real measurement data in the X-band are analyzed and turned into a reproducible model. A generative adversarial network is proposed to learn the measured clutter transfer function characteristics of a fast fading channel and multiply the dataset of measured clutter channels.