First-hand information for everyone
This paper presents a practical way to measure and describe the sound that submerged electric vehicles (SEVs) make in coastal waters. SEVs include autonomous underwater vehicles (AUVs), remotely operated vehicles, and diver propulsion systems. The authors argue that these small electric platforms can produce narrowband tones, harmonics, and modulated sounds from their motors and electronics that are important for passive detection and diagnostics, but are hard to pick out in coastal settings because of background noise, shallow-water effects, and direction-dependent radiation.
The core result is an eight-step measurement and analysis workflow for underwater radiated noise. The steps cover measurement design, cavitation screening, choosing frequency bands, characterizing ambient noise, spectral and time–frequency analysis, subsystem-oriented interpretation (linking tones to propulsion or motor control), propagation-corrected source estimation, and angular and operational analysis. A key feature is combining calibrated pass-by acoustic recordings with synchronized vehicle metadata (for example speed, depth, and position) and with context about ambient noise so that narrowband and time-varying features can be separated from the background.
The authors demonstrate the workflow on an A18D AUV operating in a coastal environment. They report drive-related tonal groups near 5.56 kHz, 11.1 kHz, and 22.2 kHz, with harmonic structure visible up to 105 kHz. After a propagation correction to estimate source strength, the tonal power spectral density (PSD) levels they report range from about 77 to 120 dB re 1 µPa^2/Hz at 1 meter. These concrete results show that high-frequency components—often missed in earlier studies limited to roughly 5–16 kHz—can be important for SEV acoustic signatures.
Why this matters: passive acoustic signatures can help detect, classify, and monitor small underwater vehicles without active transmissions. They can also support engineering diagnostics by revealing drive, motor-control, or mechanical features. The paper fills a gap because existing standards for measuring ship noise were designed for surface vessels and do not cover the spectral, geometric, and operational complexities typical of submerged electric vehicles.
Important caveats noted by the authors include the challenges posed by coastal measurements. Ambient noise from vessel traffic, waves, and biological sources can obscure vehicle features. Shallow-water propagation and the vehicle’s aspect (which direction it faces relative to the receiver) can change the received spectrum and make source estimates uncertain. Close-range measurements improve signal-to-noise ratio but are more sensitive to geometry and near-field effects. The paper proposes procedures to reduce these uncertainties, but the authors emphasize that careful calibration, synchronized vehicle metadata, and explicit ambient-noise assessment are needed to make robust interpretations.