17 December 2024 in IEEE Journal of Oceanic Engineering
Recent advancements in acoustic and millimeter-wave technologies have enabled direct communication over water–air interfaces. This cross-medium communication involves three key steps: an underwater transducer emits acoustic waves that create surface vibrations, airborne radars detect these vibrations using millimeter waves, and finally, processed radar signals facilitate communication. However, this scheme suffers from short communication distances (approximately 1 m over water) and low resolution (millimetre level). In this article, a laser-acoustic cross-medium communication (LACMC) system is proposed and validated in anechoic tank experiments with the expectation of addressing the challenges of direct, high-precision, long-range communication across water--air interfaces for integrated air--sea networking. We analyze signal transmission through underwater, water surface, and air mediums and establish a robust LACMC channel model, which was verified by experiments. Experimental results prove that the channel is time-invariant and frequency-selective, with a single direct path and no multipath effects. We further enhance cross-medium communication quality by designing appropriate transmitted signals based on channel characteristics and employing laser Doppler vibrometry for high-precision, noncontact measurements. Our experimental results demonstrate the system's efficacy in long-distance applications, achieving an underwater communication distance of 3.5 m and an above-water distance of 10.8 m. The study is expected to provide a foundation for direct, high-precision, and long-distance communication between underwater nodes and airborne nodes.
1 January 2026 in IEEE Transactions on Instrumentation and Measurement
30 December 2025 in IEEE Communications Letters
This letter proposes a variational Bayesian sparse decision-feedback equalizer (VB-SDFE) based on the recursive least squares (RLS) algorithm for underwater acoustic communications. The method employs Bayesian inference to optimize the sparse pre-equalizer coefficients in overlapping subblocks and addresses the sensitivity of RLS-based decision-feedback equalizer (RLS-DFE) to channel variations. Key features include subblock processing with a forgetting factor, variational updates for sparsity control, and Doppler-resistant phase compensation. Simulations show that the proposed algorithm achieves superior and robust performance under different channel conditions. Sea trials demonstrate a 7.8-dB gain in SNR and error-free performance at 21.5 km.