Sparse Waveform Design for Secure LPD/LPI Underwater Acoustic Communications

Harsh underwater acoustic (UW-A) channel conditions make reliable and secure underwater communications difficult to achieve. It is desirable to improve the physical layer security of the underwater acoustic channel by imparting the eavesdropper's interception efforts in the first place. In this work, we propose to design LPD/LPI waveforms that minimize the likelihood that a confidential message, transmitted between a pair of trusted underwater nodes, is intercepted by an eavesdropper. In particular, if the location of the eavesdropper is known or can be projected and prior knowledge of the channel channel state information (CSI) is available, we optimize the transmit energy and alphabet of all-spectrum sparse-binary waveforms to minimize the signal-to-interference-plus-noise ratio (SINR) at the output of the eavesdropper's maximum-SINR filter. In more practical applications, where the eavesdropper's CSI is not available, we design sparse-binary waveforms that maximize the amount of energy available for generating artificial noise (AN) to eavesdroppers. The proposed waveform designs offer enhanced LPD/LPI performance, while at the same time maintain the post-filtering SINR at the intended receiver above a pre-defined threshold level, thus providing quality-of-service performance guarantees. The effectiveness of the proposed security approach is demonstrated with simulation studies over multipath frequency-selective underwater acoustic channels and UW-A noise.