GANWave: GAN-enabled OFDM High Capacity Waveform for Low Identification and Detection Probability with Energy Efficiency and Security

The Navy relies more on data than ever before for mission operations. However, the current tactical waveform cannot provide high-speed data transmission in support of Low Probability of Detection (LPD) and Low Probability of Identification (LPI). In this proposal, Infobeyond proposes GANWave (GAN-enabled OFDM High Capacity Waveform) as a waveform alternative to provide covert communications in the data-intensive network. Superior to other waveforms, GANWave achieves higher capability, lower LPD/LPI, and more energy-saving for Naval high-speed covert communications by leveraging AI-empowered promising technologies. At first, Complex-valued Generative Adversarial Networks (GANs) is employed to emulate hardware-generated noise, making signal transmission indistinguishable from natural variations or noise to unauthorized detectors to achieve high LPD/LPI. Secondly, OFDM allows multiple Naval users to access adjacent channels simultaneously, without inter-channel interference (ICI). This improves the waveform utilization to achieve high capacity of the network. Meanwhile, the transmitted power for each subcarrier is optimized to eliminate the PAPR (Peak-to-Average Power Ratio) problem, producing energy-efficient waveforms.  Thirdly, AI-based  Initial Access Learning dynamically learns the angle of arrival of the received Beyond Line-Of-Sight (BLOS) beams. This not only reduces the beam search time with energy-saving but also chooses the beam to avoid detection by enemy RF detectors. In addition to high waveform capability, GANWave supports communications in a multi-hop and high-mobile environment, benefiting from quick beam search. Further, it can be implemented as Software Defined Radios (SDR) in FPGA and interact with other systems without their modifications.