image: A survey vessel tows a 770 m hybrid fibre-optic cable in which an interferometric section (128 hydrophones over 104 m, blue) provides high-signal-to-noise acoustic detection while a distributed acoustic sensing section (32 elements at 5 m spacing over 160 m, red) also senses its own depth and the ocean wave field. Both sections use identical air-backed mandrels (42 mm long, 11.6 mm outer diameter) but differ in wound-fibre length and optical readout. In the open-sea trial the array was towed about 500 m astern, and a cooperative acoustic source 5.7 km away provided 1 kHz test pulses. Adapted from Figure 1A of the published article.
Credit: Li S., Chai Q., Jia M., et al., (2026). A smart fiber-optic towed hydrophone array for open-sea cruising observation. The Innovation 7(9), 101534. https://doi.org/10.1016/j.xinn.2026.101534
A vessel towing several hundred metres of hydrophone cable carries a blind spot: the cable bends under current and drag, and the software that turns its signals into bearings assumes a straight line. A team in China has built a towed array that measures its own depth and curvature while it listens, using the same optical fibre that carries the sound.
Towed arrays are prized because a long aperture sharpens bearing estimates, and because they are passive, cheap and easily launched. Beamforming, the step that converts arrival delays across elements into a direction, needs each element's true position. Conventional arrays carry a few depth sensors, so the shape between them is guessed; a curved array then smears the focus and biases the bearing.
The new cable, 770 m long and towed about 500 m astern, divides the work. An interferometric section of 128 elements over 104 m, each wound with 22 m of fibre, does the sensitive listening. A distributed acoustic sensing (DAS) section of 32 elements at 5 m spacing over 160 m reads optical phase along the fibre and doubles as a depth gauge, since hydrostatic pressure shifts that phase. Both ride on the same air-backed mandrel, 42 mm long and 11.6 mm across. The data are then split by frequency: below 0.02 Hz for depth and shape, 0.02–0.3 Hz for waves, 5 Hz–2 kHz for sound.
Depth came out right. Against a pressure sensor, DAS-inferred depth gave a coefficient of determination of 0.99 and average deviation of 0.80%; the reconstructed profile showed the cable bowing, not lying flat. In the wave band the measured spectral peaks, 0.102 and 0.098 Hz, sat near the 0.105 and 0.107 Hz expected once towing speed and heading were taken into account, differing by 2.86% and 8.41%. Significant wave height, retrieved from the same spectra and corrected for element depth, ranged from 1.41 to 2.66 m, within 4.31% and 11.03% of the Copernicus wave model. At 50 m tow depth the authors estimate a detection floor of 0.15 m and a dynamic range of 180 dB.
For sound, the DAS section ran 11.57 dB above its own land-quiet baseline across 5–2,000 Hz, while the interferometric elements stayed quieter above about 2 kHz. Beamforming tracked a GPS reference bearing in two bands: one carrying vessel noise at 120–150 Hz, the other 1 kHz pulses from the distant source. Feeding the measured depth profile into the array model lifted peak beamforming power by up to 4.10 dB.
The limits belong to open-water deployments in December 2025 and January 2026: one vessel, one speed, one sea state, with coastal and shallow water untested. Depth and wave retrieval lean on pressure, so accuracy will drift with tow depth and heavy weather. What the trial shows is a towed array that calibrates itself in place — of use for wave monitoring from vessels of opportunity, for fisheries and marine-mammal acoustics, and for any sonar that has to work without a straight line.
Journal
The Innovation
Method of Research
Experimental study
Subject of Research
Not applicable
Article Title
A smart fiber-optic towed hydrophone array for open-sea cruising observation
Article Publication Date
8-Sep-2026
COI Statement
The authors declare no conflicts of interest.